Next Article in Journal
A Comparative Study on the End-Bearing Capacity of Toe-Wing & Spiral Screw Piles in Cohesionless Soil
Previous Article in Journal
Overall Lifting Construction Control Method for Large-Segment Steel Arch Bridges Based on Unstressed State Control Theory
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Regional Color Study of Traditional Village Based on Random Forest Model: Taking the Minjiang River Basin as an Example

1
College of Landscape Architecture and Art, Fujian Agriculture and Forestry University, Fuzhou 350002, China
2
Faculty of Forestry, The University of British Columbia, Vancouver, BC V6T 1Z4, Canada
*
Author to whom correspondence should be addressed.
Buildings 2025, 15(4), 524; https://doi.org/10.3390/buildings15040524
Submission received: 6 January 2025 / Revised: 25 January 2025 / Accepted: 6 February 2025 / Published: 8 February 2025
(This article belongs to the Section Architectural Design, Urban Science, and Real Estate)

Abstract

From the color geography perspective, a field investigation was conducted in the Minjiang River Basin, constructing a color index system of traditional villages. In Python, a random forest model was constructed to screen out important color indexes for traditional village color classification and explore its influence mechanism. Among eight color indexes, the important indexes are wall form and building face form, accounting for 30.50% and 19.40%, respectively. Based on this, the basin was divided into four color zones presenting color characteristics and eight color subzones presenting architectural features. The influence mechanism concerns dialect divisions that have shaped traditional villages of different color types, and the interconnection of water systems has promoted the connections among them. The application of traditional village colors in the new urban and rural planning can enhance local characteristics. Integrating the color resources of traditional villages contributes to the regional protection of culture and economic development.

1. Introduction

China’s vast land has nurtured diverse traditional villages. For example, there are mobile settlements of nomadic civilizations, courtyard-style layouts of agricultural civilizations, and cave dwellings in the Loess Plateau region [1]. These traditional villages are not just living spaces. They represent the integration of regional cultures and the harmonious coexistence between humans and nature, embodying our ancestors’ survival wisdom. Colors, a key landscape representation, form a region’s first impression [2]. With new media and transportation development, inter-regional communication has become frequent. Globalization has also impacted the region, leading to profound domestic and foreign cultural collisions and integrations. As a result, regional construction homogenization is prominent. Architectural styles and colors converge and become chaotic, causing the disappearance of regional characteristics. Excavating traditional village colors that embody urban culture and regional connotations and integrating them into urban and rural planning is significant for reshaping urban and rural landscapes and inheriting regional cultures. The Minjiang River, Fujian Province’s mother river, has a basin that is the cultural birthplace of southeastern China. This region has many traditional villages with diverse types. Since 2012, 520 villages have been listed in the national traditional village directory. The traditional dwellings in the traditional villages of the Minjiang River Basin embody the essence of significant Fujian Province architectures [3]. Their unique colors, forms, and styles integrate north–south cultures and those of different ethnic and sub-ethnic groups, so they have great research value.
Color experience originates from the biological reaction to color stimuli [4]. Regional color characteristics are significant in shaping people’s perception of a region’s appearance. As the most direct dimension of visual communication, color is crucial for cultivating local identity and enriching the environmental experience [5]. Many studies have shown that color is an important factor in developing local awareness, enhancing the sense of belonging and constructing a city’s psychological image [1,4,6]. Swirnoff, in Urban Color: An International Perspective, pointed out that a deep understanding of the psychological and physiological reactions triggered by color can guide the construction of a color design method more in line with social needs. Using this method can effectively improve users’ satisfaction with the visual quality of the regional environment [7]. Western countries initiated regional color research earlier [8]. Among them, the most well known is the ‘Color Geography’ concept proposed by LencLos, a French color scientist. He argues that color, especially that of traditional dwellings, is closely related to the regional environment. The natural environment sets the color tone, and the humanistic environment adds rich layers. This makes colors show obvious differences in regional space, each with its unique style, becoming the region’s visual symbol. Meanwhile, as a symbol of regional culture, color carries cultural characteristics and values. In regional development, color can not only effectively shape the urban image but also drive the development of economic industries like tourism and cultural and creative industries [9]. The ‘basic investigation, extraction and quantification, planning and regulation’ color research approach proposed in color geography theory has been widely applied in the planning and design field [10]. Examples include the color concept planning in Oslo [11], the architectural color design in Savannah, Georgia, USA [12], and the architectural color planning in Essex, England, by Roger Evans Associates [13]. In addition, in 1970, Mahnke put forward the ‘Color Experience Pyramid’ concept in Color, Environment, and Human Response [14]. It consists of the physiological response to color stimuli, latent unconsciousness, conscious symbolism and associations, cultural influence and unique style, the influence of fashion, trends, and style, and personal experience. From this pyramid system, from primary to advanced, we can clearly understand the context of factors influencing people’s color psychology, from physiological reactions to psychological reactions, and then to cultural influence.
In 1999, Song introduced the research findings and practical achievements of color geography theory to China in Color Design in France, attracting researchers’ attention [15]. In recent years, under China’s rural revitalization strategy, applying color geography to the study of traditional villages has become a hot topic. From the cultural revitalization perspective, studying traditional villages based on color geography helps with cultural inheritance. In traditional Chinese Five-Element Color Theory, metal, wood, water, fire, and earth correspond to white, cyan, black, red, and yellow, respectively. Bian believes that the Nantou Ancient City in Shenzhen has inherited the Lingnan region’s preference for yellow (representing earth) and red (representing fire). This preference reflects local residents’ reliance on the land and their longing for vitality. The ancient city also adeptly uses complementary color contrasts, such as those between yellow and cyan, and red and black, which accord with the mutual generation and restriction logic within the Five-Element Color framework. Currently, improper color proportions and misplacement in the ancient city have disrupted the balance of the Five-Element colors, leading to a chaotic color landscape and the loss of its original cultural connotations and visual esthetic charm [16]. From the environment improvement perspective, studying the colors of traditional villages can provide references for creating a harmonious and comfortable living atmosphere. Traditional buildings with distinct regional features in these villages are often research objects. Xie et al. used HSL chromatographic analysis to analyze the colors of traditional architectures in central Anhui and discuss application strategies in modern antique commercial streets [17]. Zhu et al. quantitatively analyzed the color characteristics of oyster-shell houses and stone houses in Xunpu Village and qualitatively evaluated public perception, and then proposed color improvement suggestions for the whole village [18].
As the study of traditional villages deepens, integrating local characteristic resources and leveraging regional cultural advantages has become a key development direction for traditional villages. Chen et al. proposed a remote sensing extraction method for typical characteristics of traditional Tibetan Plateau residential buildings using technical means like computer vision, statistical analysis, and remote sensing imaging, and then carried out classification based on it [19]. Yin et al. took the residential plane forms of 64 traditional villages in the Xiangjiang River Basin as research objects, constructed a comprehensive analysis method of prototype gene-type regional systems, and explored the adaptive expression of prototype genes and the spatial differentiation of traditional villages [20]. These achievements have fared well in realizing the discrimination and classification of traditional villages based on landscape genes. However, research basic units are not unified, and there are numerous and complex factors for discriminating traditional villages. Also, classification characteristics are hard for the public to learn and understand, making it difficult to apply and promote these achievements. Color meets people’s ‘seeing is believing’ cognitive needs and can be an important basis for regional division [21]. Currently, there are relatively few studies on the color-based regional division of traditional villages, and research methods mainly focus on qualitative analysis, with few combinations with quantitative analysis [22]. With the development of artificial intelligence, machine learning algorithms have gradually become an important classification means. Scholars such as Li et al. [23], Luo et al. [24], and Qiu et al. [25] have successfully applied the random forest model to landscape classification research. These findings showed that the random forest model has great potential in the classification field. Because of its self-improving intelligent principle, it can calculate the weight coefficients of evaluation indicators through non-linear mapping, greatly reducing the subjectivity of manual weight assignment [26].
In summary, the current regional division of traditional villages has issues such as inconsistent evaluation criteria and blurred zoning boundaries. There is also relatively little research on the regional division of colors. Moreover, color collection methods and research approaches mainly rely on qualitative analysis, with limited integration of quantitative analysis. Therefore, this study combines the theory of color geography and random forest model techniques to attempt the zoning classification and characteristic summarization of the colors of traditional villages in the Minjiang River Basin. This aims to clearly present the overall color distribution characteristics of traditional villages in the basin, exploring the internal influencing mechanisms. Although this study demonstrates the positions of traditional villages and traditional dwellings within each zone of the Minjiang River Basin in a broader color regional context, it also integrates resources, providing a base map for the protection of the colors of traditional villages as well as for modern urban and rural planning and construction.

2. Methodology

2.1. Study Area

The Minjiang River Basin is located in the northern part of Fujian Province. It is 360 km long from north to south and 280 km wide from east to west, with a drainage area of 61,000 square kilometers. Its upstream consists of the Jianxi River, the Futunxi River, the Jinxi River, the Jiulongxi River, and the Shaxi River; its midstream is composed of the main stream, the Youxi River, and the Gutianxi River; and its downstream is formed by the main stream and the Dazhangxi River. There are 24 districts under the jurisdiction of the basin, among which nine major dialects are distributed, and it nourishes more than one-third of the residents in Fujian Province (Figure 1).
Liu et al. [27,28] analyzed the distance data between traditional villages and water systems in the Minjiang River Basin and confirmed that water systems have a pronounced impact on traditional villages within a 9 km range. Traditional villages ‘choose to live near water’ and rely on waterways for production and transportation, which directly impacts the colors of traditional villages [29]. In our study, the 150 national-level traditional villages located within a 9 km range of the Minjiang River water system were selected as case sites for investigation (Figure 2).

2.2. Research Method

2.2.1. Research Framework

Our study was completed in six steps. (1) Traditional village color elements were collected based on field investigations. (2) In accordance with aspects such as the human–land relationship [30], the relationship of color carriers [31], the relationship of the degree of human participation [32], and the color index systems of traditional villages in the Minjiang River Basin were constructed, and the color index data of 150 traditional villages were organized into a collection. (3) The collection of traditional village color data was imported into a random forest model to conduct an importance analysis of color indexes and screen out the key indexes that have a significant impact on color classification. (4) Based on the screened important color indexes, the color classification of traditional villages in the basin was carried out. (5) The random forest model was used to evaluate the accuracy of the color classification results. If the evaluation results showed that the classification accuracy met the standard, then the next stage would be entered; otherwise, it was necessary to return to the second step for reselection. (6) Finally, a color classification scheme for traditional villages in the Minjiang River Basin based on the random forest algorithm was formed (Figure 3).

2.2.2. Extraction of Traditional Village Color Elements

The Minjiang River Basin belongs to a ‘typical shaded area’ [33]. Our research team conducted on-site investigations during the periods from March to August in 2023 and 2024, respectively, specifically choosing the time when it was cloudy and the colors were relatively stable between 10:00 and 16:00. We collected the color elements of traditional villages using advanced equipment (Figure 4, Table 1). For traditional dwellings, which are the main components of the color elements of traditional villages, color collection was carried out using a spectrophotometer from the aspects of roof colors, main colors (that is, the colors of the walls), auxiliary colors (that is, the colors of the wall bases), and accent colors (that is, the colors of the entrance doors).

2.2.3. Construction of the Color Index System of Traditional Villages

Current research on the construction of color index systems for traditional villages is scarce, while numerous studies exist on the construction of traditional village landscape indexes [34,35]. Among them, the theory of the traditional village landscape gene information chain, led by Liu [36], is well known. Liu divided traditional village landscape indexes into material and non-material form indexes. The former mainly cover architectural style and street–alley patterns, with 7 major and 25 minor categories like overall layout, roof shape, gable shape, facade form, and plane shape, while the latter are divided based on folk culture and religious beliefs. Based on this, Lin et al. [37] took the dialect area as the sociocultural foundation, selected 6 major architectural landscape elements (roof forms, gable forms, color characteristics, the presence or absence of enclosing walls, facade forms, and plane shapes) and constructed the traditional village landscape index system in the Minjiang River Basin.
This study drew on previous studies and considered their practicality to avoid overly complex classifications and ensure index operability. From three dimensions—natural resource characteristics, social cultural characteristics, and traditional dwelling characteristics—we selected 8 traditional village color indexes: district (county), dialect, roof form, gable form, wall form, wall base material, entrance door form, and building form. These were used to construct the color index system of traditional village in the Minjiang River Basin (Table 2).

2.2.4. Parameter Optimization of the Random Forest Model

The scikit-learn library in the Python programming language was used to perform random forest classification on traditional village color data [38,39]. First, the 150 village samples were randomly divided into a test set and a training set at a ratio of 3:7 to ensure the effectiveness of model training and evaluation. In importing the RandomForestClassifier class from scikit-learn and using the color data of the training set as input, a random forest classification model was trained to identify the internal laws and patterns in the data. After training, the model could classify and predict traditional villages in the test set, thus achieving the classification of traditional village colors. This provides a machine learning-based quantitative classification method for the research and analysis of traditional village colors.
The K-fold cross-validation method was adopted to optimize the framework parameters and decision tree parameters of the random forest model. Based on previous studies, the value of K in our study was set to 5 [26]. Due to the small number of training samples, the maximum depth was not restricted. Through the cross-validation method, the optimal parameter combination of the model in this study was set as follows: the number of random forest trees was set to 150, the maximum number of features was set to ‘auto’, the minimum number of samples required for splitting was set to 2, and the minimum number of samples required at leaf nodes was set to 1.

2.2.5. Optimization of Classified Color Indexes and Accuracy Verification

The feature_importances function for calculating feature importance was employed to evaluate the importance of each color index variable. These variables were then sorted in descending order according to their importance scores. Subsequently, based on the optimized indexes, pairwise combinations were made to classify 150 traditional village samples. This approach can effectively improve the operating speed and accuracy of the model [40,41]. For accuracy verification, the following important indexes were introduced [42]:
Accuracy: It refers to the proportion of the number of correctly classified samples to the total number of samples. The calculation formula is
A c c u r a c y = T P + T N T P + T N + F P + F N
Here, TP represents the number of true positive cases; TN represents the true negative cases; FP represents the false positive cases; and FN represents the false negative cases. Accuracy is one of the most intuitive indexes.
Recall refers to the proportion of the number of samples that the model correctly predicts as positive cases to the number of actual positive cases. A high recall indicates that the model performs well. The calculation formula is
R e c a l l = T P T P + F N
Precision refers to the proportion of samples that the model correctly predicts as positive cases to the number of all samples predicted as positive cases. The calculation formula is
P r e c i s i o n = T P T P + F P
The F1-score is the harmonic mean of the precision and recall, with a range of [0, 1], and it can evaluate the performance of the model more comprehensively. The higher the F1-score, the higher the classification quality of the model.
F 1 - S c o r e   =   2 P r e c i s i o n × R e c a l l P r e c i s i o n + R e c a l l

3. Results

3.1. Screening Results of Important Color Indexes

The data of the eight color indexes of 150 traditional village samples were input into the random forest model for programming. The calculation results were output as shown in Figure 5. The importance indexes, ranked from largest to smallest, were as follows: wall form > building face form > geographical location > entrance door form > gable form > roof form > wall base material. It can be seen that the sum of the importance indexes of the two indexes, namely wall form (30.50%) and building face form (19.40%), is 49.90%, which is very important. Therefore, the two indexes of ‘wall form’ and ‘building face form’ were screened out as the key indexes for the colors of traditional villages. Their characteristics are shown in Table 3.

3.2. Classification Scheme and Accuracy Verification Results

3.2.1. Classification Scheme

To scientifically name the colors of traditional villages, our study adopted the method of combining indexes in pairs. During the naming process, special emphasis was placed on the ‘wall form’ as the dominant color index, while considering the important role of the ‘building face form’ in reflecting the abundance of regional landscape resources. Therefore, the ‘wall form’ was placed before the ‘building face form’ in the naming system. In this way, they were divided into 11 different categories (Table 4).

3.2.2. Accuracy Test

Figure 6 and Table 5 show that the classification accuracy of the 105-sample training set and the 45-sample test set both reached 100%. High accuracy in both sets indicates that the model has excellent predictive ability and generalization performance and has truly learned the inherent classification rules of the color index data. Thus, our study adopted the combination of ‘wall form’ and ‘building face form’ as the classification scheme to divide the 150 traditional villages in the Minjiang River Basin into 11 color classifications.

4. Analysis of Regional Division of Traditional Village Color in Minjiang River Basin

In the actual classification process, three color classifications of traditional village were found, namely ‘wooden terraced house’, ‘rammed earth terraced house’, and ‘terraced house with earth and wood composite walls’, widely distributed in the Minjiang River Basin. Their regional characteristics are indistinct, which is consistent with the research findings of scholars such as Wu et al. [22], Dai [43], and Lin et al. [37]. Therefore, to ensure concise and effective analysis, we decided to only carry out a comprehensive and in-depth spatial visualization analysis on the remaining eighth classifications with the help of GIS.
Guided by the above-mentioned color classification scheme and in considering practical factors like administrative division boundaries, natural boundaries, and color representativeness, the color zoning of the Minjiang River Basin was conducted. At the county level, if color classifications show no distinct differentiation, the county-level unit is generally not divided into different zones [22,44]. Since there was some overlap in the spatial distribution of the color classifications, when determining a county’s color zoning, the main reference was the color features with the highest conservation value and cultural significance [45]. For instance, traditional dwellings in Yongtai County have rammed-earth walls, showing a yellow color. Although there were also some white-plastered wall buildings, they were overall classified into the yellow color zone [46].
Finally, as Figure 7 and Table 6 show, the Minjiang River Basin was divided into four major color zones: gray, yellow, brown, and colorless. There were eighth minor color subzones, namely the skylight courtyard with black brick walls subzone, the hall-and-wing house with black brick walls subzone, the rammed earth skylight courtyard subzone, the earth fort subzone, the skylight courtyard with ‘boot-like’ and ‘hat-like’ walls subzone, the courtyard house with walls with tiles on the top subzone, the courtyard house with wooden structure and white-washed walls subzone, and the skylight courtyard with plastering walls subzone. The major zones represent regional color characteristics, and the minor subzones represent regional architectural form characteristics.

4.1. Gray Zone

The gray zone is in Fujian Province, which is the origin and a region with relatively complete development of black brick architecture [47]. In China, the history of brick making dates back 5000–5300 years. Bricks are made by firing clay. When fully oxidized during firing, they become red bricks. But when water is added for cooling during firing, incomplete oxidation occurs, producing black bricks. Black bricks have much greater hardness and strength than red bricks [48]. Black brick walls greatly improve wall performance in terms of thermal insulation, integrity, and earthquake resistance. This significantly enhances the quality of the living environment, allowing traditional buildings to remain intact without shaking even after hundreds of years [49]. The black brick walls of traditional dwellings are directly exposed, showing a cyan-gray color that gives a visual impression of tranquility and steadiness. According to differences in building face forms, the gray zone is divided into two subzones.

4.1.1. The Skylight Courtyard with Black Brick Walls Subzone

This subzone is mainly distributed along the Futunxi River. Here, the dialect of Fujian and Jiangxi is mainly spoken, and in a few areas, the Nanping dialect is used. Influenced by Jiangxi culture, the building face form is the skylight courtyard. The patios are generally small, which helps reduce sunlight exposure and hot-air convection. There are usually one to four patios. Thus, traditional dwellings here feature a square outline, a closed appearance, and a short-width, long-depth structure. All exterior walls are made of black bricks, and a few residences use white mortar to edge the brick joints [50]. The average color of black bricks in this subzone is Lab (48.23, 3.16, 12.33).

4.1.2. The Hall-and-Wing House with Black Brick Walls Subzone

This subzone is mainly distributed along the Jiulongxi River, where the dialect of Fujian province and Hakka is predominantly spoken. The hall-and-wing house layout is deeply influenced by Hakka culture, reflecting the strong centripetal cohesion of the Hakka people [51]. In the whole building, the hall has the most prominent volume and decoration and typically serves as the venue for family sacrificial and ceremonial activities. The wings are arranged facing the hall. Usually, elders reside in the hall, while juniors live in the wing houses, mirroring the Confucian concept of respecting the elders andemphasizing order. The wing houses are expandable; when the family population grows, more rows can be continuously added on both sides. As a result, the overall plan of the hall-and-wing house takes on a horizontally long and square-shaped flat form [52]. The exterior walls are relatively dark with a blackish hue, with an average color of Lab (45.11, 2.08, 11.05).

4.2. Yellow Zone

The yellow zone is mainly characterized by rammed earth walls. Mud is the main component of these building walls, and the buildings’ first impression is yellow. In the Minjiang River Basin, red and yellow soils are dominant and easily accessible. They are suitable for ramming into walls due to their solidity, durability, and ability to keep interiors warm in winter and cool in summer, making them a top choice for villagers building houses [53]. Rammed earth wall buildings are widespread in the basin. Based on building face form, they can be classified as skylight courtyards, courtyard houses, and earth forts. Since dry rammed earth walls have high strength but can easily fall apart when soaked, waterproof materials are used on the wall tops to prevent rain damage. According to different wall-top treatments, rammed earth wall forms include plain rammed earth walls, ‘boot-like’ and ‘hat-like’ walls, and walls with tiled tops.

4.2.1. The Rammed Earth Skylight Courtyard Subzone

This subzone is mainly distributed along the Gutianxi River, where the eastern Fujian dialect is predominantly spoken. The wall construction technique here is relatively simple. Yellow clay is pounded into identically sized earth blocks using molds and then stacked to form walls. Finally, tiles, thatch, stone flakes, or other materials are laid on the rammed earth top as the roof, creating a full rammed earth wall. The exterior walls have little decoration, displaying the original color of the rammed earth [16]. With an average color of Lab (54.69, 9.25, 23.13), it presents a simple earth-brown color, giving a sense of intimacy.

4.2.2. The Skylight Courtyard with ‘Boot-Like’ and ‘Hat-Like’ Walls Subzone

This subzone is mainly distributed along the Jianxi River, where the dialect of northern Fujian Province is mainly spoken. In a small number of northern areas, the boundary dialect of the western Zhejiang province is used. The wall with ‘boots and hats’ is the most representative three-section wall form in this subzone [1]. The brick and tile wall top is the ‘hat’, the rammed earth wall is the ‘body’, and the brick and stone wall foundation is the ‘boots’. The walls mainly display the color of rammed earth. Due to the addition of black brick elements, the overall tone tends to be yellowish-gray with relatively low saturation. The average color of the rammed earth is Lab (50.35, 7.40, 19.63).

4.2.3. The Earth Fort Subzone

This subzone is mainly distributed along the Shaxi River, where the middle Fujian dialect is predominantly spoken. Influenced by immigrant culture, for ancestors to safeguard their homeland and ensure ethnic group continuity, clan living became an inevitable choice. The earth fort emerged as the times demanded, with its large capacity and solid defensive features. The earth fort’s outer wall is tall and constructed by ramming raw earth. The wall base consists of super large stones, and the whole building exudes a unique style of simplicity, ruggedness, and solidity [54]. The earth fort’s overall appearance is yellowish-brown, and the average color of the rammed earth is Lab (54.69, 9.25, 23.13).

4.2.4. The Courtyard House with Walls with Tiles on the Top Subzone

This subzone is mainly distributed along the Dazhangxi River, where the eastern Fujian dialect prevails. Influenced by Jiangcheng culture, it features the architectural style of a quadrangle courtyard house. A two-courtyard five-bay layout is the most prevalent, forming the architectural characteristics of a short depth and a wide facade. The exterior walls have a tiled-top wall form. On the lofty saddle-shaped rammed earth fireproof walls, tiles are neatly arranged. From a distance, it resembles armor and is quite spectacular [55]. The architectural colors are mainly that of rammed earth, with an average color of Lab (54.22, 6.11, 19.40). With the addition of tiles, the overall wall appearance is yellowish-gray, giving a visual impression of steadiness and reserve.

4.3. Brown Zone

The Courtyard House with Wooden Structure and White-Washed Walls Subzone

This subzone is mainly distributed along the Youxi River, where the transitional dialect of Fujian prevails. The building face form follows the layout of a three-courtyard house, characterized by a short depth and a wide facade. The courtyard is relatively spacious, generally enclosed by rammed earth courtyard walls, with the main gate on one side of the wall. The wall form is ‘wooden structure and white-washed walls’, giving the whole building a lively color style, perfectly reflecting the local mountain forest culture and showing the natural beauty of ‘thousands of timbers standing on the ground’ [56]. The rammed earth courtyard walls are usually about 1.5 m high, functioning as half-walls. This design ensures better ventilation and light transmission in the courtyard, making it suitable for activities like grain drying. This traditional dwelling consists of materials with mixed colors. The average color of the wooden walls is Lab (42.00, 7.61, 5.78), that of the white-washed walls is Lab (85.50, 1.72, 5.94), and that of the rammed earth courtyard walls is (61.06, 6.84, 20.63). The colors of various building materials blend, and the overall appearance is brownish.

4.4. Colorless Zone

The Skylight Courtyard with Plastering Walls Subzone

This subzone is mainly distributed along the Lower Minjiang River, where the eastern Fujian dialect prevails. Inside the plastered walls are rubble-earth walls, a sustainable construction technique. Its components include broken bricks, tiles, small stones, and other construction waste. Rubble-earth walls solve the long-standing problem of dealing with construction waste. But they are prone to crumbling, alkali efflorescence, and weathering. Thus, moisture-proof and alkali isolation measures for these walls are crucial. The wall bases are made of relatively water-resistant materials like bricks and stones, and the wall surfaces are plastered [57]. This subzone uses regionally characteristic pot ash for wall plastering. The resulting walls are gray, with an average color of Lab (55.44, 0.20, 1.87). The plastered walls provide a cool visual sense. Also, the gray color absorbs sunlight, keeping the walls dry for long term use.

5. Analysis of Influence Mechanism of Color Classification

The traditional village color zones of the Minjiang River Basin were overlaid with the water system distribution and the dialect distribution. The analysis results are shown in Figure 8. There is a relatively high correlation among them. Figure 9 more intuitively shows the flow, splitting, combination, and affiliation of the traditional village color index data, including district divisions, dialect divisions, water system divisions, and traditional village color types.

5.1. Dialect Division Shapes Different Color Types of Traditional Villages

Over a thousand-year period, the Han people of the Central Plains migrated southward in large-scale waves multiple times, deeply influencing the Minjiang River Basin. Immigrants arrived via land and sea routes. Different routes brought distinct immigrant groups with diverse dialects. During migration, they were influenced by the languages of fellow travelers and the regions they traversed. After reaching the basin, they integrated with local dialects, forming unique languages, each with different cultural connotations [57]. In the Minjiang River Basin, clear dialect divisions reflect the Han people’s migratory patterns in different periods, as well as the diverse ethnic and sub-ethnic groups. These differences determine the color variations in traditional villages, especially prominent in traditional dwellings [58]. Dialect divisions can essentially represent cultural divisions. However, due to water systems and geographical locations, in some adjacent areas with different dialects, cultures are interconnected, and the cultural influences may overlap. Dai et al. [43,54] also had the same research findings. Specifically, there are the following five aspects.
The Jianxi River, Futunxi River, and Jinxi River areas are mainly dominated by the dialect of northern Fujian and the combination dialect of Fujian and Jiangxi, respectively. These areas are deeply influenced by the culture of Zhejiang Province and Jiangxi Province, with a developed academy culture. The landscapes of traditional villages also strictly adhere to the ritual system. The dwellings are mainly characterized by skylight courtyards, which are square and regular in shape and simple and elegant in color [59,60].
The dialects in the Jiulongxi River and Shaxi River areas are complex, mainly including the Fujian and Hakka dialect and the middle Fujian dialect. This area was developed relatively late, and its population was mainly composed of immigrants from other places. Influenced by immigrant culture, the unstable living conditions led the early settlers to exhibit a more conservative and stability-seeking cultural inclination [52,53]. They mostly lived in clans, with earth forts and hall-and-wing houses as their representative buildings.
The middle Minjiang River and the Gutianxi River area are located in the middle of the Minjiang River Basin, dominated by the Nanping dialect and the Eastern Fujian dialect, respectively. These two places are deep in the hinterland and have been influenced by various cultures, such as central plains culture, mountain forest culture, immigrant culture, and so on. Under the influence of multi-culture integration, traditional village colors also integrate the characteristics of multiple color zones. In terms of the traditional dwellings, they are consistent with the horse-head walls in the Jianxi River and Futunxi River areas, as well as consistent with the rammed earth buildings in the Shaxi area [53]. After integration, they have formed their own unique rammed earth skylight courtyards.
The Youxi River area is mainly dominated by the Fujian transitional dialect. Located in a forest area and influenced by the mountain forest culture, most of the residents there have the idea of a self-sufficient small-scale peasant economy, which is well reflected in the dwellings [56]. The three-courtyard houses with wooden structures and white-washed walls make use of local materials, such as wood and bamboo, complementing the surrounding mountain forest environment.
The lower Minjiang River and the Dazhangxi River area are dominated by the eastern Fujian dialect. As one of the earlier areas for economic development, the prosperity of the Maritime Silk Road led to a large number of merchants and noble clans settling here. The river city culture advocated for by the social elite groups had a profound impact on the traditional villages in this area. Based on the skylight courtyards, the saddle-shaped fireproof gable walls of traditional dwellings are more vivid and diverse in shape, and the building scale is also larger. Moreover, the detailed decorations such as carvings, colored paintings, and lime sculptures are more exquisite, reflecting the pursuit of esthetics and craftsmanship [55,61].

5.2. The Interconnection of Water Systems Facilitates the Connection of Traditional Villages Among Different Color Types

Basin ancestors migrated along the Minjiang River’s creeks and rivers. They moved downstream, lived by the water, traveled via waterways, and settled in the mountains. Longer migration routes and more traversed regions led to more complex cultural integration and dissemination paths. The networking and richness of Minjiang River Basin traditional village colors reflect the diachronic layers of long-distance migrations [29]. When looking beyond modern administrative divisions to physical geography’s original features, the basin’s traditional village colors show obvious zoning. However, the Minjiang River Basin has a wide geographical reach and many tributaries. Since the Western Han Dynasty, it has been an important transportation waterway, promoting external connections and strengthening internal bonds. Leveraging its advantages, the Minjiang River links Jianan, Fuzhou, and Chongan, three key early development nodal cities in Fujian Province, significantly impacting regional coordinated development [51].
The traditional villages colors also exhibit the characteristics of homogeneity and centripetal development in the coordinated development linked by the Min River. In particular, there are several common features in traditional dwellings, such as the following: (1) Whether it is the skylight courtyard, the hall-and-wing house, or so on, the essence of the building’s human living environment is an enclosed and introverted architectural space [37]. (2) The roofs are all paved with blackish-gray tiles, presenting a calm and reserved tone and without excessive decorations, complementing the architectural style [22]. (3) The wall bases are mostly made of stone blocks, and the materials for the building’s exterior enclosure are mostly black bricks or rammed earth [27,53]. (4) The entrance door is the focus of architectural decoration [60]. (5) Most of the planar layout of the buildings is symmetrical along the main axis, with a clear distinction between the principal and the subordinate parts. The layout in mountainous areas is a bit more flexible than that in plain areas [59,61].
Water systems are one of the crucial factors for human settlement. Areas along the Minjiang River and its tributaries are often the main distribution regions of settlements. A well-developed water system, on the one hand, facilitates the transportation of construction materials. On the other hand, it enables residents from different regions to conduct trade, cultural exchanges, and other activities through waterway transportation [27]. This leads to the mutual dissemination and integration of architectural techniques and landscape cultures in various places, thereby forming some common characteristics in the colors of traditional villages and traditional dwellings. In particular, the color characteristics of areas adjacent to the water system are more similar [28,43].

6. Conclusions

The main conclusions in our paper are as follows:
(1) From the three dimensions of natural resource characteristics, social and cultural characteristics, and traditional dwellings characteristics, eight color indexes of traditional villages, namely the geographical location, dialect area, roof form, gable form, wall form, wall base material, entrance door form, and building form, were selected to construct the color system of traditional villages in the Minjiang River Basin.
(2) The number and categories of color index variables involved in the classification affect the final color classification effect. In our study, ‘wall form’ and ‘building face form’ were preferably selected. Among them, the wall form (30.50%) had the greatest impact on the colors of traditional villages, followed by the building face form (19.40%).
(3) Based on the ‘wall form’ and ‘building face form’, the Minjiang River Basin was divided into 11 categories using the random forest algorithm. The classification accuracy of the test set and the training set reached 100%. This model truly learned the inherent classification rules of the color index data.
(4) Based on the random forest classification scheme and in comprehensively considering practical factors, the Minjiang River Basin was divided into four major color zones and eight minor color subzones. Respectively, ① the gray zone included the skylight courtyard with black brick walls and the hall-and-wing house with black brick walls. ② The yellow zone included the rammed earth skylight courtyard, the earth fort, the skylight courtyard with ‘boot-like’ and ‘hat-like’ walls, and the courtyard house with walls with tiles on the top. ③ The brown zone included the courtyard house with a wooden structure and white-washed walls. ④ The colorless zone included the skylight courtyard with plastering walls.
(5) The influencing mechanism of the color classification is that the dialect divisions shape the different color types of traditional villages, and the interconnection of water systems promotes the connection among traditional villages of different color types.
In the Minjiang River Basin, traditional villages are diverse in type. To a large extent, their colors reflect various aspects such as local geographical climate, ethnic culture, customs, and so on. Conducting a regional color study of traditional villages is indeed an effective way to protect regional traditional culture.
In traditional culture, regional comparative studies show that the colors of traditional villages and dwellings are mainly influenced by natural conditions (water systems) and sociocultural factors (dialects). Natural conditions mainly affect the choice of building materials for traditional dwellings [21]. Areas with adjacent water systems have similar natural conditions, leading to similar material selection. Thus, the colors of traditional dwellings are also similar. Cultural customs mainly influence the facade forms of traditional dwellings [10]. Dialect areas represent ethnic or sub-ethnic divisions, which largely stand for different settlement cultures. This directly causes differences in the external appearances of traditional dwellings in the Minjiang River Basin [59]. The colors of traditional villages in the basin are deeply influenced by the Five-Phase concept, emphasizing the unity of man and nature and the pursuit of environmental harmony. So, building materials are mostly from nature and show their original colors [16]. In contrast, Western traditional dwellings often have painted walls with different colors. However, both are essentially influenced by natural conditions and cultural concepts [9]. For instance, traditional dwellings in Italy are often painted yellow. In the Mediterranean climate, yellow building facades are brighter and warmer, matching the surrounding environment. Also, influenced by Renaissance color culture, yellow usually symbolizes holiness and dignity, wisdom and light, joy and vitality, making it popular among locals [7]. Our regional color research fills a gap in color geography research for specific regions, offering abundant cases and empirical evidence for the discipline’s development. The in-depth analysis of Minjiang River Basin’s traditional village colors contributes to improving the theoretical system of the relationships among color, geography, culture, society, and other factors. It promotes interdisciplinary research and provides a solid basis for subsequent related studies.
At the level of modern development, regional color research offers valuable inspiration and references for modern urban and rural planning. Incorporating the color elements of traditional villages in the Minjiang River Basin into new urban and rural construction can create an urban landscape with a strong local flavor. This not only avoids the homogenization phenomenon of ‘one size fits all’ cities but also makes urban spaces more recognizable and endows people with a sense of belonging [10]. This coincides with Mahnke’s view of the ‘Color Experience Pyramid’ [4] and also aligns with the critical regionalism theory proposed by Frampton [62]. In addition, regional research on the colors of traditional villages contributes to the development of cultural tourism. On the one hand, the uniqueness of regional color landscapes can enhance the local brand value [37]. When a region becomes famous worldwide for its unique color landscape, a brand effect is formed, attracting more high-end industries to settle in. On the other hand, regional color research also helps to promote the in-depth integration of related industries, such as the handicraft industry and the cultural and creative industries. Thus, an industrial chain is formed, promoting the optimization and upgrading of the industrial structure [30].
It is crucial to arouse people’s attention to regional colors and apply the colors representing the culture and spirit of a region to new urban and rural planning, endowing modern urban and rural construction with local ‘characteristics’. This can ensure the preservation and continuation of the colors of traditional villages [52]. Moreover, it should be noted that the social and economic conditions determine the scale and sophistication of traditional dwellings. Any change in these conditions will lead to different morphological characteristics of the dwellings [46]. Important historical events serve as critical catalysts in this transformation process, either accelerating or decelerating the entire process. In the future, research can take significant historical events within the river basin that have influenced human activities as time nodes to analyze the development and evolution laws of architecture in different periods. Further in-depth analysis and exploration of this conclusion are needed.

Author Contributions

Conceptualization, D.K. and X.F.; methodology, D.K.; software, D.K. and X.F.; investigation, D.K., X.F., Z.L., X.L. and M.C.; writing—original draft preparation D.K.; writing—review and editing, Z.C.; funding acquisition, Z.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was funded by the Industry-University Cooperation Project of Fujian Province ‘Construction and Application of Color Map of Traditional Dwellings in the Minjiang River Basin Based on Color Preference of Indigenous Residents’, grant number 2023Y4003.

Data Availability Statement

Research data are available upon request from the corresponding author.

Conflicts of Interest

The authors declare no conflicts of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.

References

  1. Kong, D.; Lin, X.; Lu, Z.; Fei, X.; Xie, Y.; Chen, Z. Quantitative study on the color of traditional dwellings in jianxi river basin, fujian province, based on the lab color space model. J. Asian Archit. Build. Eng. 2024, 27, 1–27. [Google Scholar] [CrossRef]
  2. Najar, K.; Nylander, O.; Woxnerud, W. Public’s Visual Preferences Survey Facilitates Community-Based Design and Color Standards Creation. Buildings 2024, 14, 2929. [Google Scholar] [CrossRef]
  3. Qu, L.; Xiao, C. Fangliang’s Dream: Fujian Classic Ancient Residence; Haichao Photographic Art Publishing House: Fuzhou, China, 2008. [Google Scholar]
  4. Mahnke, F.H. Color, Environment, and Human Response an Interdisciplinary Understanding of Color and Its Use as a Beneficial Element in the Design of the Architectural Environment; John Wiley & Sons: Hoboken, NJ, USA, 1996; pp. 1–248. [Google Scholar]
  5. Yang, L.; Liu, P.; Lin, L. The Effect of Traditional Village Landscape Genes in Construction Place Identity: Taking the Dong Minority Village as An Example. Sci. Geogr. Sin. 2015, 35, 593–598. [Google Scholar]
  6. Yoshida, S. Environmental Color Design Techniques: Neighborhood Color Creation; China Architecture and Building Press: Beijing, China, 2011. [Google Scholar]
  7. Svenover, L. Unban Color: An International Perspective; China Water Resources and Hydropower Press: Beijing, China, 2007. [Google Scholar]
  8. Anter, K.F.; Billger, M. Colour research with architectural relevance: How can different approaches gain from each other? Color Res. Appl. 2010, 35, 145–152. [Google Scholar] [CrossRef]
  9. Lenclos, J.P.; Lenclos, D. Colors of the World A Norton Book for Architects And Designers; WW Norton: New York, NY, USA, 2004. [Google Scholar]
  10. Cheng, G.; Li, Z.; Xia, S.; Gao, X. Study on the color of traditional villages in Huizhou based on regional perspective. Archit. J. 2023, 142–149. [Google Scholar]
  11. Porter, T. Environmental color mapping. J. Urban Des. Int. 1997, 2, 23–31. [Google Scholar] [CrossRef]
  12. Porter, T.; Mikellides, B. Color for Architecture Today; New York Taylor & Francis: New York, NY, USA, 2009; pp. 128–134. [Google Scholar]
  13. Linton, H. Color in Architecture: Design Methods for Buildings, Interiors, and Urban Spaces; McGraw Hill: Hong Kong, China, 1999; pp. 151–157. [Google Scholar]
  14. Mahnke Frank, H. Color, Environmental and Human Response; Wiley: Hoboken, NJ, USA, 1996. [Google Scholar]
  15. Song, J. Color Design in France; Shanghai People’s Fine Arts Publishing House: Shanghai, China, 1999. [Google Scholar]
  16. Bian, W. Study on the color evolution and renewal strategy of Old urban buildings in the context of urban restoration: A case study of Nantou Ancient City in Shenzhen. Mod. Urban Res. 2021, 73–79. [Google Scholar] [CrossRef]
  17. Xie, L.; Fu, Y.; Sun, C.; Guo, Y. Research on the application of traditional residential architecture colors in modern antique street in Central Anhui. Furnit. Inter. Decor. 2021, 34–38. [Google Scholar] [CrossRef]
  18. Zhu, X.; Zhang, H.; Lan, S.; Huang, Q. Research on color landscape optimization of traditional villages based on quantitative Analysis and public perception: A case study of Xun pu Village in southern Fujian Province. J. Cent. China Norm. Univ. (Nat. Sci. Ed.) 2020, 54, 50–59. (In Chinese) [Google Scholar] [CrossRef]
  19. Chen, S.; Li, Z.; Wang, Z.; Liu, N.; He, Y. Remote Sensing Extraction Method of Typical Features of Traditional Residential Buildings in Tibet. Comput. Eng. Appl. 2025, 61, 349–358. Available online: http://kns.cnki.net/kcms/detail/11.2127.tp.20240509.1129.009.html (accessed on 22 November 2024).
  20. Yin, L.; Liu, P. Type expression and spatial differentiation of planar prototype genes in traditional dwellings in Xiangjiang River Basin. Geogr. Res. 2023, 42, 2191–2210. [Google Scholar]
  21. Liu, H.; Yang, C.; Su, Y.; Qiang, M.; Zhou, X.; Yuan, Z. A Study on the Atlas and Influencing Factors of Architectural Color Paintings in Tibetan Timber Dwellings in Yunnan. Buildings 2024, 14, 3971. [Google Scholar] [CrossRef]
  22. Wu, S.; Chen, Z.; Wang, B. Regional study on the color of traditional houses in Fujian. Archit. J. 2017, 105–109. [Google Scholar] [CrossRef]
  23. Li, L.; Chen, X.; Zhang, H.; Sun, D.; Wen, H.; Chen, S. Rural residential land suitability evaluation based on Random forest algorithm: A case study of Hechuan District, Chongqing. J. Chongqing Norm. Univ. (Nat. Sci. Ed.) 2022, 39, 93–104. [Google Scholar]
  24. Luo, Y.; Zhang, R.; Xue, X.; Liu, M.; Wang, J.; Lou, J. Research on karst landform classification based on positive and negative landforms: A case study of Guizhou Karst Area. J. Guizhou Norm. Univ. (Nat. Sci. Ed.) 2024, 42, 9–19. [Google Scholar] [CrossRef]
  25. Qiu, H.; Chen, C. Habitat suitability evaluation of ancient Ginkgo trees in Changsha City based on random forest and MaxEnt model. J. Cent. South For. Univ. Sci. Technol. 2024, 44, 87–97. [Google Scholar] [CrossRef]
  26. Guo, H.; Boonprong, S.; Wang, S.; Zhang, Z.; Liang, W.; Xu, M.; Yang, X.; Wang, K.; Li, J.; Gao, X.; et al. Dominant Tree Species Mapping Using Machine Learning Based on Multi-Temporal and Multi-Source Data. Remote Sens. 2024, 16, 4674. [Google Scholar] [CrossRef]
  27. Liu, S.; Huang, L.; Huang, J.; Xu, G. Spatial Pattern Analysis of historical and cultural protection and inheritance in Minjiang River Basin: From the perspective of cultural landscape. Mod. Urban Res. 2024, 1–8. [Google Scholar] [CrossRef]
  28. Liu, S.; Fan, H.; Wang, Y.; Gao, X. Spatial characteristics and correlation analysis of traditional villages in Minjiang River Basin. Mod. Urban Res. 2019, 9, 17–25. [Google Scholar]
  29. Zhao, Z.; Kong, Y.; Guan, Z.; Liu, Y.; Wang, J. Spatial morphological characteristics and formation mechanism of villages in river basin: A case study of the Yellow River Delta. City Plan. Rev. 2023, 47, 69–81. [Google Scholar]
  30. Li, Z.; Zhao, J.; Wu, X. Recognition and protection of landscape features of traditional villages from the perspective of human-land relationship: A case study of Southwest Hubei Province. Land Nat. Resour. Res. 2024, 52–58. [Google Scholar] [CrossRef]
  31. Liu, Y. Study on the Color System of Classical Gardens in Suzhou; Beijing Forestry University: Beijing, China, 2014. [Google Scholar]
  32. Zhang, C.; Zi, M.; Luo, J.; Chen, T. Evolution and reconstruction of forest and rural landscape driven by tourism. J. Cent. China Norm. Univ. (Nat. Sci. Ed.) 2022, 56, 189–200. [Google Scholar]
  33. Weber, R. Monochromatic design in a polychrome world. Why our cities have become increasingly gray: A dichotomy between production and reception in architectural color design. Color Res. Appl. 2023, 48, 543–556. [Google Scholar] [CrossRef]
  34. Liu, X.; Tu, Z. Quantitative analysis of the evolution of traditional village features. J. Huaqiao Univ. (Nat. Sci.) 2017, 38, 811–817. [Google Scholar]
  35. Xu, H.; Zhang, T.; Ge, B.; Song, Y. Developing of rural settlement landscape gene research system based on content analysis. J. Asian Archit. Build. Eng. 2023, 22, 2839–2850. [Google Scholar] [CrossRef]
  36. Liu, P. Construction and Application of Landscape Gene Map in Traditional Chinese Settlements; Peking University: Beijing, China, 2011. [Google Scholar]
  37. Lin, R.; Yang, F.; Zhang, D.; Zou, C.; Zeng, Z.; Li, X. Landscape feature extraction and floristic division of traditional villages in Minjiang River Basin. South. Archit. 2022, 1, 54–60. [Google Scholar]
  38. Van Strien, M.J.; Keller, D.; Holderegger, R.; Ghazoul, J.; Kienast, F.; Bolliger, J. Landscape genetics as a tool for conservation planning: Predicting the effects of landscape change on gene flow. Ecol. Appl. 2014, 24, 327–339. [Google Scholar] [CrossRef] [PubMed]
  39. Fang, Q.; Wu, J.; Zhu, J.; Xie, B. A review of stochastic forest methods. Stat. Inf. Forum 2011, 26, 32–38. [Google Scholar]
  40. Liu, M.; Liu, J.; Ma, M.; Jiang, Y.; Zeng, J. Planting monitoring of pepper tree in Linxia, Gansu Province based on GF-2 PMS image and random forest. Remote Sens. Nat. Resour. 2022, 34, 218–229. [Google Scholar]
  41. Li, L.; Huang, C.; Zhang, Y.; Liu, L.; Wang, Z.; Zhang, H.; Ding, M.; Zhang, H. Spatial-temporal pattern simulation of grazing intensity in Tibetan Plateau based on geographically weighted random forest. Sci. Geogr. Sin. 2023, 43, 398–410. [Google Scholar]
  42. Chen, H. Building Scale Population Estimation Based on Multi-Source Data and Random Forest Algorithm. Master’s Thesis, East China Normal University, Shanghai, China, 2021. [Google Scholar]
  43. Dai, Z. Fujian culture and its influence on Fujian traditional houses. South. Archit. 2011, 1, 24–28. [Google Scholar]
  44. Jin, F. Analysis on the geographical distribution and causes of color regions of ancient residential buildings in Fujian. J. Fuzhou Univ. (Philos. Soc. Sci.) 2015, 29, 87–91. [Google Scholar]
  45. Shi, D. Xiangxi Traditional Village Township Building Color Design Research. Ph.D. Thesis, Hunan University, Changsha, China, 2023. [Google Scholar] [CrossRef]
  46. Qiu, Z.; Hu, X.; Qian, H.; Liu, Y. Distribution characteristics and influencing factors of Different types of dwellings in traditional villages in Fujian. Furnit. Inter. Decor. 2021, 54–59. [Google Scholar] [CrossRef]
  47. Han, F.; Leng, C.; Chen, J.; Zhan, Y. Identification of apatite rock types by trace elements based on machine learning. Bull. Mineral. Petrol. Geochem. 2024, 43, 607–620. [Google Scholar] [CrossRef]
  48. Hao, Y.; Wu, R.; Bao, Y.; A, S.; Wang, L. Deterioration mechanism and residual life prediction of ancient blue bricks in Longshengzhuang under complex environment. J. Xi’an Univ. Archit. Technol. (Nat. Sci. Ed.) 2024, 56, 690–700. [Google Scholar] [CrossRef]
  49. Liu, X. Research on Clean Brick Wall of Traditional Architecture in Fuhe River Basin. Ph.D. Thesis, Wuhan University, Wuhan, China, 2020. [Google Scholar]
  50. Shi, J.; Chun, Q.; Mi, Z.; Feng, S.; Liu, C.; Liu, Z.; Wang, D.; Zhang, Y. Comparative study on material properties of ancient fired clay bricks of China. Case Stud. Constr. Mater. 2023, 19, 12. [Google Scholar] [CrossRef]
  51. Lu, J. Analysis on the architectural characteristics of ancient folk houses in northern Fujian. Fujian Cult. Mus. 2019, 56–59. [Google Scholar]
  52. Zhuo, X.; Xiao, D. Study on the development law and geographical spatial distribution characteristics of Hakka Traditional dwellings in Jiangxi, Fujian and Guangdong Province based on global survey. Archit. J. 2020, 16–22. [Google Scholar] [CrossRef]
  53. Huang, J.; Huang, Z.; Mo, X.; Liu, Y. A Preliminary study on the replication-expression mechanism of traditional Hakka huts adapted to the region: A case study of the Tang Hengwu of Nanhua Ancient Village in Western Guangdong in Qing Dynasty. Ind. Archit. Constr. 2024, 54, 94–105. [Google Scholar]
  54. Lin, C. Interpretation of the forms and characteristics of traditional dwellings in Wuyi. J. WuyiColl. 2019, 38, 7–11. [Google Scholar] [CrossRef]
  55. Zhu, G. Study on defense mechanism of folk dwellings in Fujian—A case study of Minzhong Fort. Urban Archit. 2022, 19, 105–107+113. [Google Scholar]
  56. Zhang, H. Research on Landscape Planning and Design of “Beautiful Countryside” in Banshan Village, Youxi County. Master’s Thesis, Fujian Agriculture and Forestry University, Fuzhou, China, 2016. [Google Scholar]
  57. Wang, X. Generation and Variation: A Study of Saddle Wall in Fuzhou; Chinese National Academy of Arts: Beijing, China, 2013. [Google Scholar]
  58. Zhang, Z. Economic Development and Cultural Change in the Pearl River Delta: A Study on language. J. Sun Yat-Sen Univ. (Soc. Sci. Ed.) 1997, 56–60. [Google Scholar]
  59. Dai, Z.; Chen, Q. Fujian Ancient Architecture; China Architecture Industry Press: Beijing, China, 2015. [Google Scholar]
  60. Luo, H.; Leng, T. Analysis on the cooperation Principle of soil, stone and wood in the environmental strain construction of “Earthwork”. Constr. J. 2020, 1–8. [Google Scholar] [CrossRef]
  61. Wang, Y. Overview of Fujian Culture; Fujian Education Press: Fuzhou, China, 1994. [Google Scholar]
  62. Kenneth, F. Towards a Critical Regionalism: Six Points for an Architecture of Resistance; The Anti-Aesthetic. Bay Press: Port Townsend, WT, USA, 1983; pp. 16–30. [Google Scholar]
Figure 1. Study area.
Figure 1. Study area.
Buildings 15 00524 g001
Figure 2. Distribution map of traditional villages with pronounced hydrotaxis (within 9 km) in the Minjiang River Basin.
Figure 2. Distribution map of traditional villages with pronounced hydrotaxis (within 9 km) in the Minjiang River Basin.
Buildings 15 00524 g002
Figure 3. Color classification steps of traditional villages based on random forest model.
Figure 3. Color classification steps of traditional villages based on random forest model.
Buildings 15 00524 g003
Figure 4. The research team extracted color elements in the Minjiang River basin.
Figure 4. The research team extracted color elements in the Minjiang River basin.
Buildings 15 00524 g004
Figure 5. Evaluation results of the importance of traditional village color indexes.
Figure 5. Evaluation results of the importance of traditional village color indexes.
Buildings 15 00524 g005
Figure 6. Heatmaps of the confusion matrix for the test set and the training set.
Figure 6. Heatmaps of the confusion matrix for the test set and the training set.
Buildings 15 00524 g006
Figure 7. Regional division of traditional village colors in Minjiang River Basin.
Figure 7. Regional division of traditional village colors in Minjiang River Basin.
Buildings 15 00524 g007
Figure 8. Water system distribution and dialect distribution together affect the color classification of traditional villages.
Figure 8. Water system distribution and dialect distribution together affect the color classification of traditional villages.
Buildings 15 00524 g008
Figure 9. Sankey diagram of districts, dialects, water systems, and traditional village color types.
Figure 9. Sankey diagram of districts, dialects, water systems, and traditional village color types.
Buildings 15 00524 g009
Table 1. Advanced research equipment.
Table 1. Advanced research equipment.
Device NameFunctional RolePhysical Picture
Konica Minolta CM-700d spectrophotometerThis handheld color meter makes direct contact with the building surface to obtain accurate color values.Buildings 15 00524 i001
SONY A7RM4 micro single digital cameraRecords real photos of traditional villages, records the color of traditional villages, and reviews the data.Buildings 15 00524 i002
Dji innovation royal MAVIC2 droneCollects full-picture information of traditional villages and photos that are difficult for human eyes to observe.Buildings 15 00524 i003
Table 2. Color index system of traditional villages in Minjiang River Basin.
Table 2. Color index system of traditional villages in Minjiang River Basin.
Index
Dimensions
Secondary IndexIndex Attribute
The characteristics of natural resourcesDistrict (county)Pucheng, Wuyishan, Pucheng, Songxi, Jianyang, Jianou, Zhenghe, Guangze, Shaowu, Jianning, Taining, Jiangle, Shunchang, Ninghua, Qingliu, Yongan, Sanyuan, Shaxian, Yanping, Youxi, Gutian, Minqing, Minhou, Fuzhou, Yongtai
The characteristics of social cultureDialectThe boundary dialect of Zhejiang; the dialect of northern Fujian; the dialect of Fujian and Jiangxi; the transitional dialect of Fujian, Hakka, and Jiangxi; the dialect of Fujian and Hakka; the dialect of middle Fujian; the Nanping dialect; Fujian province’s transitional dialect; the dialect of eastern Fujian
The characteristics of traditional dwellingsRoof formFlush gable roof,
overhanging gable roof
Gable formThe peak of the overhanging gable roof
The fire-proof gable wall with horse-head-shaped coping
The fire-proof gable wall with five-element-shaped coping
The fire-proof gable wall with saddle-shaped coping
Wall formAll wooden wall
Black brick wall
Rammed earth wall
Earth and wood composite wall
‘Boot-like’ and ‘hat-like’ wall
Wall with tiles on the top
Wooden structure and white-washed wall
Plastering wall
Wall base materialStone wall base
Black brick wall base
Sanhe earth wall base
Entrance door formDoor-cover-style gate (brick door-cover-style gate and wooden door-cover-style gate),
Gate-tower style door
Stone-built fort–gate
Flat-faced door
Building face formTerraced house
Courtyard house
Skylight courtyard
Earthen fort
Hall-and-wing house
Table 3. Characteristics of important color indexes of traditional villages in Minjiang River Basin.
Table 3. Characteristics of important color indexes of traditional villages in Minjiang River Basin.
Important Color IndexIndex AttributeExampleIndex Description
Wall formWooden wallBuildings 15 00524 i004The building walls are all made of wood.
Earth and wood composite wallBuildings 15 00524 i005Rammed earth walls are built on the left and right sides of the wooden walls for reinforcement and enclosure.
Black brick wallBuildings 15 00524 i006The building walls are all made of black bricks.
Rammed earth wallBuildings 15 00524 i007The building walls are all made of rammed earth.
‘Boot-like’ and ‘hat-like’ wallBuildings 15 00524 i008The structure of this kind of wall is as follows: The wall foundation is stones. Two layers of blue bricks are laid on it for leveling. Then, 60 cm of rammed earth walls are built about one story high. The top of the wall is covered with bricks and tiles. This forms the most representative three-section rammed earth wall form of a ‘boot-like’ and ‘hat-like’ wall. The tile and brick coping is the ‘hat’, the rammed earth wall is the ‘body’, and the stone wall foundation is the ‘boots’.
Wall with tiles on the topBuildings 15 00524 i009This kind of wall involves fixing perforated tiles on the top of the rammed earth wall to achieve the effect of rain protection. Generally, earth slurry is used to apply a base coat on the outside of the rammed earth wall, and then the tiles are embedded at the top. Bamboo nails are inserted to fix them. The heads of the bamboo nails are sealed with white lime, and a beautiful square shape is scraped out with a trowel.
Wooden structure and white-washed wallBuildings 15 00524 i010This kind of wall uses wood as the structural framework of the building. Between the wooden frames, wall panels made of reed stalks or bamboo slices are used as filling materials. White lime is then applied to the wall panels to form the white-washed wall.
Plastering wallBuildings 15 00524 i011The raw materials for the plastered walls come from the soot that accumulates on the bottoms of cooking pots and in chimneys and flues when people cook. Walls painted in this way can prevent moisture and alkali and effectively protect the internal structure of the walls.
Building face formTerraced houseBuildings 15 00524 i012The floor plan of a terraced house is mostly long and narrow. This is the earliest and most widely distributed architectural type, referring to the expansion of a basic componen in a one-way dimension.
Courtyard houseBuildings 15 00524 i013A courtyard house is centered around a courtyard and surrounded by buildings on all sides. A three-sided courtyard house is composed of houses on three sides and a surrounding wall on one side, while a four-sided courtyard house, based on the three-sided courtyard house, has a gatehouse added in the front to form a closed courtyard.
Skylight courtyardBuildings 15 00524 i014The core feature of a skylight courtyard is that there is an open-air space—a patio—in the middle. The patio is usually relatively small, located in the central part of the building.
Hall-and-wing houseBuildings 15 00524 i015A hall-and-wing house mainly consists of a main hall and several wing rooms. It is an architectural layout with distinct characteristics of family residence. The main hall is the core part of the building, located on the central axis, while the wing rooms are distributed on both sides of the main hall and are perpendicular to the main hall.
Earthen fortBuildings 15 00524 i016An earthen fort is a large-scale dwelling architecture with defensive functions. It is an architectural form where a clan lives together. It can effectively protect the safety of the clan or village residents during turbulent times.
Table 4. Results of random forest classification of traditional village colors.
Table 4. Results of random forest classification of traditional village colors.
Classification StandardClassification Naming
Wall FormBuilding Face Form
1Wall with tiles on the topCourtyard houseCourtyard house with walls with tiles on the top
2Wooden structure and white-washed wallCourtyard houseCourtyard house with wooden structure and white-washed walls
3Rammed earth wallSkylight courtyardRammed earth skylight courtyard
4Rammed earth wallEarthen fortEarthen fort
5Plastering wallSkylight courtyardSkylight courtyard with plastering walls
6Rammed earth wallTerraced houseRammed earth terraced house
7Wooden wallTerraced houseWooden terraced house
8Earth and wood composite wallTerraced houseTerraced house with earth and wood composite walls
9Black brick wallSkylight courtyardSkylight courtyard with black brick walls
10Black brick wallHall-and-wing houseHall-and-wing house with black brick walls
11‘Boot-like’ and ‘hat-like’ wallSkylight courtyardSkylight courtyard with ‘boot-like’ and ‘hat-like’ walls
Table 5. Results of classification accuracy evaluation of random forest model.
Table 5. Results of classification accuracy evaluation of random forest model.
AccuracyRecallPrecisionF1
training set1111
cross-validation set1111
test set1111
Table 6. Regional division of traditional village colors in Minjiang River Basin.
Table 6. Regional division of traditional village colors in Minjiang River Basin.
Color Zone Color SubzoneExamples
Gray ZoneThe Skylight Courtyard with Black Brick Walls SubzoneBuildings 15 00524 i017
The Hall-and-Wing House with Black Brick Walls SubzoneBuildings 15 00524 i018
Yellow ZoneThe Rammed Earth Skylight Courtyard SubzoneBuildings 15 00524 i019
The Skylight Courtyard with ‘Boot-Like’ and ‘Hat-Like’ Walls SubzoneBuildings 15 00524 i020
The Earth Fort SubzoneBuildings 15 00524 i021
The Courtyard House with Walls with Tiles on the Top SubzoneBuildings 15 00524 i022
Brown ZoneThe Courtyard House with Wooden Structure and White-Washed Walls SubzoneBuildings 15 00524 i023
Colorless ZoneThe Skylight Courtyard with Plastering Walls SubzoneBuildings 15 00524 i024
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Kong, D.; Fei, X.; Lu, Z.; Lin, X.; Cai, M.; Chen, Z. Regional Color Study of Traditional Village Based on Random Forest Model: Taking the Minjiang River Basin as an Example. Buildings 2025, 15, 524. https://doi.org/10.3390/buildings15040524

AMA Style

Kong D, Fei X, Lu Z, Lin X, Cai M, Chen Z. Regional Color Study of Traditional Village Based on Random Forest Model: Taking the Minjiang River Basin as an Example. Buildings. 2025; 15(4):524. https://doi.org/10.3390/buildings15040524

Chicago/Turabian Style

Kong, Deyi, Xinhui Fei, Zexuan Lu, Xinyue Lin, Mengqing Cai, and Zujian Chen. 2025. "Regional Color Study of Traditional Village Based on Random Forest Model: Taking the Minjiang River Basin as an Example" Buildings 15, no. 4: 524. https://doi.org/10.3390/buildings15040524

APA Style

Kong, D., Fei, X., Lu, Z., Lin, X., Cai, M., & Chen, Z. (2025). Regional Color Study of Traditional Village Based on Random Forest Model: Taking the Minjiang River Basin as an Example. Buildings, 15(4), 524. https://doi.org/10.3390/buildings15040524

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop