Visual Quality Evaluation of Historic and Cultural City Landscapes: A Case Study of the Tai’erzhuang Ancient City

Abstract

As a World Cultural Heritage site, the Beijing–Hangzhou Grand Canal is lined with historic and cultural cities that bear abundant historical and cultural connotations. It is of vital importance to address the issues of landscape homogenization, the disconnect between conservation measures and public needs, and other related challenges in the development of these cities. This study adopts a mixed-methods approach combining qualitative and quantitative research. By integrating subjective landscape evaluation with eye tracking analysis, the ancient city of Taierzhuang along the Beijing–Hangzhou Grand Canal was selected as the research subject to conduct an in-depth investigation into the visual experience and preferences for different types of landscapes in this area. The study yielded the following findings: There are significant differences in visual experiences among various types of landscapes in Taierzhuang Ancient City. Among them, participants exhibited the highest level of attention toward canal landscapes and the lowest toward heritage site landscapes. In terms of visual distribution differences, eye tracking heatmaps revealed that attention was primarily concentrated on architectural structures and water surface reflections. In the subjective evaluation analysis, canal cultural landscapes received the highest comprehensive score (4.39 points), followed by historical architectural landscapes (4.02 points), historical street landscapes (3.93 points), modern commemorative landscapes (3.72 points), and heritage site landscapes (3.69 points). Additionally, a significant correlation was found between eye tracking data and subjective evaluation results, validating the scientifically effective relationship between subjective assessments of historical cultural landscapes and eye tracking analysis. The findings of this study not only provide a scientific basis for landscape improvement and optimization in other canal-related historical and cultural cities but also offer new research methods and perspectives for the protection and development of other heritage landscapes.

1. Introduction

Under the backdrop of globalization and urbanization, historic and cultural cities are confronted with the dual pressures of conservation and development. As a current trend in development, tourism exerts significant pressure on heritage conservation. Achieving a balance between the conservation of historic and cultural cities and tourism development is an urgent issue that needs to be addressed. Historic and cultural cities serve as significant bearers of urban multicultural landscapes, sustaining the collective memory of residents while also shaping the unique character of the city [1]. As one of the longest and largest man-made waterways in the world, the Beijing–Hangzhou Grand Canal is endowed with a wealth of natural, engineering, and cultural heritage resources [2], and its entire basin encompasses numerous historic and cultural cities. In 2019, China began the development of cultural parks along the Beijing–Hangzhou Grand Canal. These parks differ in size and features but all prioritize key tasks such as heritage protection and transmission, integration of culture and tourism, and environmental support. Their goal is to comprehensively safeguard and pass on the cultural heritage of the Grand Canal while boosting regional economic and social development. Significant achievements have been made in the protection and inheritance of cultural heritage [3]. The renewal of ancient cities guided by cultural principles necessitates the selection of appropriate renewal methods based on the unique characteristics of each city. This approach aims to avoid excessive commercialization and homogenization, thereby preserving the vitality and uniqueness of the city’s culture [4]. Additionally, the construction of historic and cultural cities is also an intrinsic requirement for enhancing national cultural soft power and achieving the goal of a strong cultural nation [5]. Zaozhuang City is located at the center of the Beijing–Hangzhou Grand Canal, while the ancient city of Taierzhuang lies in Taierzhuang District of Zaozhuang City. As a significant historical and cultural city along the Beijing–Hangzhou Grand Canal, it is characterized by its unique canal cultural landscape and historical relics, making it a representative case in terms of conservation and development. Therefore, this study selects the ancient city of Taierzhuang as the research subject, aiming to explore its landscape visual experience and conservation strategies, thereby providing references for other similar historical and cultural cities.

1.1. Methods for Landscape Visual Quality Assessment

Landscape visual quality assessment often employs methods such as questionnaire surveys or image rating studies, based on which corresponding landscape visual quality evaluation systems are constructed [6]. In the 1970s and 1980s, Professors Kaplan S. and Kaplan R. [7] in the field of psychology conducted in-depth analyses from the perspective of landscape visual perception psychology. They proposed concepts such as environmental preference and scenic information, and introduced a landscape aesthetic evaluation model. In 1972, Kaplan R. [8] proposed a dimensional analysis for landscape assessment, emphasizing the spatial significance of scenery. Simultaneously, they pioneered research on landscape visual preference theory in response to the damage caused by development to important landscape environments and heritage [8]. Their work holds significant importance for the protection of landscape heritage. Their research promoted reforms in the World Heritage Convention and ICOMOS regarding the identification and protection of landscape heritage, drawing attention to the visual quality and cultural value of landscape heritage. It underscored the importance of preserving the visual integrity and cultural connotations of landscape heritage, providing new perspectives and methods for the protection and management of landscape heritage. The psychophysical school employs the “stimulus–response” relationship to explain the “landscape–aesthetics” relationship in landscape evaluation. This approach encompasses several commonly used methods [9,10], including the Scenic Beauty Estimation (SBE) [11] method proposed by Daniel et al., the Landscape Comparison Judgment (LCJ) [12] method proposed by Buhyoff et al., the Semantic Differential (SD) [13] method proposed by Osgood, and the Balanced Incomplete Block–Landscape Comparison Judgment (BIB-LCJ) [14] method proposed by Yu, which integrates SBE and LCJ [15]. The study by Zhang [16] quantitatively confirmed the interactive effects of visual features on landscape preferences, establishing openness and orderliness as core influencing factors, thereby providing a foundational framework for landscape evaluation. Building on this, Aghabozorgi et al. [17] employed the PPGIS method to focus on micro-landscapes in campus settings, discovering that the combination of green plant diversity, blue spaces, and recreational facilities not only enhances spatial attractiveness but also directly influences user behavioral choices by promoting psychological restoration. This conclusion resonates with the perspective of Xie et al. [18] that “cultural scene perception requires the integration of functional and experiential dimensions”—the latter emphasized in their study of historic districts that design should be guided by visitor experience preferences, integrating cultural narratives into spatial perception elements. At the deeper construction level of cultural landscapes, related research has further expanded cognitive dimensions. Some scholars proposed [19] that planning should strengthen the implantation of cultural symbols in different landscape units and enhance experiential authenticity through interactive design. This complements the study by Zhang et al. [20] on group perception differences in historic cities—their findings revealed that tourists pay more attention to the geographical imagery of ancient cities, while residents are more sensitive to the orderliness of architectural colors and layouts, highlighting the cognitive divide between “insider” and “outsider” perspectives. Yang et al. [21] delved into the emotional experience dimension, demonstrating the partial mediating effect of positive emotions between tourists’ cognitive evaluation and cultural identity, providing a theoretical pathway for quantifying the emotional value of landscapes. In the latest research, Tung et al. [22] compared human judgments with the performance of GPT-4 and LLaVA models using natural language processing techniques. They found that large language models align closely with human preferences in dimensions such as complexity and coherence but exhibit significant deviations in “readability” judgments. This highlights both the potential of technological assistance and the irreplaceable subjectivity of landscape evaluation. Notably, the theory of sense of place [23] further points out that the synergistic construction of material environments, cultural symbols, and behavioral activities is a key mechanism for enhancing visitors’ emotional pleasure. The aforementioned studies all converge on a central point: whether for natural landscapes or cultural heritage spaces, public perception needs must be thoroughly explored through qualitative methods such as questionnaires and interviews [24]. Particularly in the evaluation of historic cities, the value dimension of landscapes as aesthetic objects consistently dominates.

1.2. Application of Eye Tracking Technology in Landscape Evaluation

Currently, landscape visual quality assessment methods predominantly originate from the subjective perspectives of evaluators, relying on public perceptions and often involving complex indicator systems (Figure 1). As human behavior research technologies keep evolving and innovating, researchers have increasingly turned to methods like eye tracking, electroencephalography (EEG), and magnetic resonance imaging (MRI) to explore the psychological activities of individuals during information processing. Landscape is regarded as a vision-centered cognitive process and serves as a medium for action [25,26]. Analyzing eye movements can help to uncover the audience’s preferences for landscapes [27,28], as well as how different audiences vary in their perception of landscapes [29,30,31]. Eye tracking technology has also demonstrated unique value in the visual perception research of landscape and architectural heritage. The team led by Deng [32] focused on traditional architecture in Nan’an, revealing through eye movement data the differences in visual attention to cultural value characteristics of heritage among groups with diverse cultural backgrounds, thereby providing an empirical foundation for cross-cultural landscape cognition studies. Hao et al. [33] further examined elements of historical landscape evaluation, finding that architectural components such as doors, windows, and walls accounted for over 40% of fixation duration and frequency, confirming the central role of physical interfaces in the visual cognition of traditional architecture. Building on this, Zheng et al. [34] introduced the concept of “information density” to quantify visual environmental appeal, with dynamic eye movement analysis showing that observers’ fixation targets within blocks exhibited exploratory shifts over time, revealing the temporal characteristics of environmental cognition. In urban and tourism contexts, Zhang et al. [35] combined eye tracking devices with emotional scales to confirm the significant impact of interactive elements on museum facades and artificial environmental components on visual quality and emotional experience. Meanwhile, Ye et al. [36] extended this method to rural tourism environments, discovering that different landscape types (e.g., farmland, water systems) directly influenced tourists’ emotional fluctuations through variations in visual stimuli, collectively highlighting the applicability of eye tracking technology in multi-scenario emotional prediction. Xu et al. [37] innovatively integrated the SBE-SD method with eye movement analysis to complete a visual experience evaluation of different areas and landscape types in Yi’an Ancient Fortress, providing a composite methodological example for the quantitative assessment of historical spaces. Li et al. [38] combined eye tracking experiments with the Semantic Differential method, finding that visual elements such as wall styles and window-to-wall ratios in Macau’s historic center, along with their coupling effects with positive emotions like “warmth”, significantly influenced tourists’ behavioral intentions, deepening the understanding of the linkage mechanism between heritage visual perception and behavioral decision-making. Xing et al. [39] employed mixed analysis of eye movement data and semi-structured interviews to reveal correlations between environmental element types and visual behavior characteristics in historic districts, further validating the advantages of eye tracking technology in analyzing unconscious visual preferences. In summary, existing research has combined eye tracking technology with subjective evaluation methods such as SBE and SD [40], achieving both the reliability verification of traditional landscape assessment approaches and capturing physiological instinctive responses in visual experiences, significantly enhancing the objectivity and scientific rigor of landscape evaluation. However, current studies remain concentrated on natural landscapes such as forests and waterfronts, as well as individual historic districts, with systematic assessments of visual quality in canal-based historic and cultural cities still insufficient, providing an important direction for future research expansion.

Therefore, the study primarily explores the following issues:

Q1. 

Are there significant differences in visual behavior when people view different types of landscapes in the Tai’erzhuang Ancient City?

Q2. 

Are there significant differences in subjective evaluations when people view different types of landscapes in the Tai’erzhuang Ancient City?

Q3. 

Is there a correlation between visual behavior and subjective evaluation?

2. Materials and Methods

2.1. Study Area

Tai’erzhuang Ancient City, a key node of the Beijing–Hangzhou Grand Canal (included on the World Heritage List in 2014), holds significance as one of the world’s two cities rebuilt as World Cultural Heritage sites after the destruction of the Second World War. Spanning 2 km², it integrates ancient waterways, historical relics (e.g., Tai’erzhuang Battle Memorial Hall), and cultural exchange functions. Designated as China’s first National Cultural Heritage Park in 2011, its unique post-war reconstruction and canal-side heritage make it a critical case for studying historic urban landscapes [41]. Therefore, selecting Tai’erzhuang Ancient City for the study of historic and cultural city landscapes holds great significance. Scientifically and objectively revealing tourists’ preferences and landscape evaluations during their visits is essential for the protection, planning, and enhancement of visitor experiences of cultural heritage. It also helps to better preserve and showcase its unique charm (Figure 2).

The principles for classifying historical and cultural landscapes encompass spatial differences, functional relevance, dominant factors, scale principles, and practicality [42]. Historic and cultural landscapes are constructed from five core categories: architectural landscape, human figure landscape, artistic landscape, living landscape, and public landscape [43]. Their connotations include both material and intangible elements, such as spatial patterns, natural environmental landscapes, historical districts and architectural complexes, cultural heritage protection units, local folklore, and folk crafts [44]. In terms of landscape cultural characteristics, they can be categorized into “cultural landscapes”, “historic urban landscapes”, “historic districts”, and “historic and cultural blocks [45]”. Tai’erzhuang Ancient City, as a National Cultural Heritage Park, integrates different cultural attributes of history, ruins, and modernity in its landscape. In landscape design and conservation, it is essential to comprehensively consider the complexity, integrity, layout logic, cultural depth, accessibility, and comfort of the landscape, as these influencing factors may interfere with or limit visitors’ viewing experience. Therefore, full consideration and optimization must be given in the design and classification of landscape types, based on previous research. This study categorizes the landscape of Taierzhuang Ancient Town into five types: historical building landscape, ruins landscape, canal cultural landscape, modern commemorative landscape, and historical street landscape. The historical building landscape denotes the intricate material entity shaped by the interaction of primary structures., characteristic building complexes, archways, and the surrounding environment. The ruins landscape consists of architectural ruins such as walls, houses, and bunkers damaged by war. The canal cultural landscape is the canal water town landscape formed along the Beijing–Hangzhou Grand Canal, composed of different waterways and buildings. The modern commemorative landscape is mainly composed of monuments, memorial halls, former residences of celebrities, and commemorative sculptures. The historical street landscape is the cultural block landscape with streets as the main element.

2.2. Research Methods

This study primarily employed eye tracking technology, questionnaires, and statistical analysis methods: (1) Eye tracking devices captured participants’ ocular physiological responses as they viewed landscape images, and the collected data were analyzed to identify visual response patterns to scenery. (2) Questionnaires were administered to collect participants’ subjective evaluations and aesthetic ratings of landscape visual quality. (3) Subjective evaluation data were combined with eye tracking data, and correlation analysis was conducted to explore the relationship between eye tracking metrics and subjective evaluations during image observation.

2.3. Experiment Design

2.3.1. Experiment Image Selection

Considering the limitations of manpower and material resources in on-site landscape evaluation, images were selected as the medium for this study. Studies have indicated that the assessment outcomes derived from images closely align with those gathered through on-site evaluations [46,47]. The images were sourced from the Dianping website [48]. The images were chosen based on the main tourist nodes of Tai’erzhuang Ancient City, using a human eye perspective (about 1.60 m above the ground) to exclude non-landscape elements such as tourists or vehicles. The images were captured during daytime hours, specifically between 10 a.m. and 5 p.m. To ensure the representativeness of the images, those that appeared more frequently were selected. Initially, 150 images were chosen, all with a consistent resolution of 1600 × 1440 pixels and an aspect ratio of 16:9. A panel of experts was invited to select the most representative images for each category. In the end, 25 images (5 for each landscape type) were chosen from the initial pool of 156 for the experiment. Additionally, one extra image was selected from the remaining pool to serve as a “test” (warm-up) image, with the goal of mitigating the primacy effect and the beginning set effect [49].

2.3.2. Participant Selection

A total of 153 participants took part in this study, resulting in 150 valid datasets. Participant screening criteria included the following: All participants had normal or corrected-to-normal vision, no color blindness or color vision deficiency, normal cognition, and could accurately identify the landscape types presented in the images. The main body of participants consisted of university students, along with a small number of teachers and staff members. Their ages ranged from 19 to 39 years old, with no specific requirements regarding occupation, gender, or educational background. Before the experiment, we provided some sample landscape images and asked participants to evaluate them aesthetically. This ensured that participants possessed the capability to assess landscape aesthetics. The diverse backgrounds of the participants, including varying ages, genders, and education levels, further demonstrate the representativeness of their preferences. These participants had some travel experience and the ability to evaluate landscape aesthetics, with preferences that aligned with those of the general public. The choice of university students as the primary research subjects was both feasible and representative [50,51]. The participants included 75 males and 75 females, ranging in age from 19 to 39 years (with an average age of 25 years). They had a wide range of academic backgrounds, including history, architecture, geography, landscape architecture, computer science, chemistry, agriculture, and management, among others. This diversity ensured a broad spectrum of preferences within the test population. According to psychological research standards, a sample size of over 30 participants is generally regarded as sufficient to ensure the reliability and credibility of the experimental results [52]. Therefore, selecting 150 participants is deemed reasonable.

2.3.3. Experimental Equipment

During the experiment, a Dikablis Glasses 3.0 eye tracking device (Ergoneers Company, Geretsried, Germany) was employed to capture eye movements. This device comprises an eye tracker, a dongle, and an adapter. A Dell Alienware m16 R2 (Alienware, a subsidiary of Dell Inc., Miami, FL, USA) laptop with a 16-inch screen was used to display the experimental images. Additionally, a tablet computer was prepared for administering questionnaires and viewing images (Figure 3).

2.3.4. Experiment Procedure

The eye tracking experiments were conducted indoors to minimize external interferences such as lighting and weather conditions, ensuring that participants were tested under consistent experimental conditions and maintaining the calibration reliability of the eye tracking devices [53]. 1. Participants were positioned about 50 cm away from the computer screen, which was equipped with eye tracking devices. They were also given an explanation of the experiment’s purpose, procedure, and requirements. 2. Once the participants were settled, the calibration process was carried out. 3. After the calibration was successfully completed, 2 warm-up images were shown to the participants without their prior knowledge, followed by the 25 experimental images. Eye movement data collection commenced simultaneously with the presentation of visual stimuli and persisted until the entire image sequence was displayed. Each image remained on screen for a duration of 10 s, followed by a 3 s intermission of black screen, allowing participants to reset their visual attention. Studies have indicated that the display time for eye tracking experimental images generally falls between 7 and 20 s. This duration ensures that participants have adequate time to observe the images, while avoiding invalid fixations or the creation of eye movement data caused by overly long viewing times [54]. 4. Following the experiment, participants were led to a brightly lit room to fill out the subjective evaluation questionnaire. Participants evaluated the landscapes in the same order as the images were shown. Eye tracking metrics and subjective evaluation preferences exert a certain degree of mutual influence and constraint on each other [55] (Figure 4).

2.3.5. Indicator Selection

Eye movement indicators: The fundamental visual processes of the human eye consist of fixations, saccades, and blinks. Prior research has shown that fixations are a relatively stable state of visual movement necessary for visual information processing and reflect cognitive processing. Psychological changes are indicated by fixation duration and frequency [56]. Saccades refer to the quick, abrupt eye movements that enable the gaze to swiftly move from one fixation point to another. They are primarily used to rapidly locate targets or information of interest within a visual scene. Fixations and saccades represent the processes of information processing and search, respectively. Consequently, this research chose four eye movement metrics: average fixation duration, fixation count, average saccade duration, and saccade count. The precise meanings and definitions of these metrics are outlined in

Table 1. The basic significance of eye movement metrics.

Eye Movement Metrics	Saccade Count	Basic Significance
Average fixation duration/ms	AFD	The average duration of each fixation when observing a specific area or object reflects the individual’s level of attention to that area and the length of time required for information processing.
Number of views/n	NV	The total number of fixations made when observing a specific area or object reflects the frequency of attention and the degree of interest that an individual has toward that area.
Average scan duration/ms	ASD	The average duration of each saccade, which is the rapid movement of the eyes from one fixation point to another during the observation process, reflects the speed of visual information processing and the efficiency of eye movement.
Scan count/n	SC	The total number of saccades, which are the rapid movements of the eyes from one fixation point to another during the observation process, reflects the frequency and range of visual exploration.

[57].

Subjective evaluation scale: The study scale employs a combination of the SBE (Scenic Beauty Estimation) method and the SD (Semantic Differential) method to construct a subjective evaluation system for the visual quality of historic and cultural cities. The Scenic Beauty Estimation (SBE) method is a commonly utilized approach in environmental psychology for assessing landscape aesthetic quality. This technique involves evaluating the overall landscape based on participants’ visual perceptions in conjunction with predefined evaluation standards. The Semantic Differential (SD) method employs linguistic scales to measure individuals’ intuitive psychological responses, thereby enabling a quantitative assessment of subjective mental experiences [58]. In this experiment, the SBE method used a 5-point scale to directly rate the aesthetic quality of the landscape, with 1 indicating “very unattractive” and 5 indicating “very attractive”. The SD method typically employs 5 to 7 evaluation scales [58]. Taking into account the accuracy and complexity of the evaluation, this study adopted a five-segment scale with scores ranging from 1 to 5, where 1 represents the lowest score and 5 the highest score. This decision was based on the landscape features of Tai’erzhuang Ancient City and relevant research [59]. Five pairs of adjectives were selected as evaluation indicators for this study, including relaxation, curiosity, liking, cultural significance, and uniqueness.

2.3.6. Data Analysis

After the experiment, four eye tracking metrics related to the landscape characteristics of Taierzhuang Ancient City and the SBE-SD survey data were extracted for processing. The process flow is shown in Figure 5.

3. Results

3.1. Differences in Eye Movement Indicators

A one-way analysis of variance (ANOVA) was performed on the eye tracking data for the five types of landscapes, and the results are shown in

Table 2. Analysis of differences in eye movement indicators among different landscape types.

Eye Movement Metrics	Average Fixation Duration/ms	Number of Views/n	Average Scan Duration/ms	Scan Count/n
Overall Differences	0.001 **	0.00 **	0.128	0.022 *
Differences Among Landscape Types
Historical Building Landscape (a)	835.14	59.85 b	139.92	67.42 e
Ruins Landscape (b)	755.34 c	52.57 acde	146.02	73.96
Canal Cultural Landscape (c)	873.7 bd	62.67 bde	133.3	72.96
Modern Commemorative Landscape (d)	759.86 c	56.96 bc	143.99	71.76
Historical Street Landscape (e)	852.57	57.2 bc	139.73	75.19 a

Note: * indicates a significant difference at the 0.05 level; ** indicates a significant difference at the 0.01 level; in the site landscape, 52.57 ^acde, where 52.57 represents the mean fixation count for the site landscape, and acde indicates significant differences between the site landscape and historical architectural landscapes, canal cultural landscapes, modern commemorative landscapes, and historical street landscapes. The interpretation of other values follows the same approach.

(Figure 6). Significant differences were found among the five types of landscapes in terms of fixation count, average fixation duration, and saccade count (with significance levels of 0.001, 0.000, and 0.022, respectively), while no significant differences were observed in average saccade duration. The degree of difference varied between different types of landscapes. The ruins landscape exhibited the greatest differences in eye movement characteristics compared to the other landscapes. Specifically, the fixation count for the ruins landscape was significantly different from the other four types of landscapes, and the average fixation duration was significantly different from that of the canal landscape. Secondly, the canal cultural landscape exhibited significant differences in average fixation duration compared to the ruins landscape and the modern commemorative landscape, and in average fixation count compared to the ruins landscape, the modern commemorative landscape, and the historical street landscape. Regarding saccade count, the historical building landscape showed a significant difference compared to the historical street landscape. The ruins landscape had the lowest average fixation duration and fixation count, but a relatively high saccade count. This indicates that participants quickly scanned this landscape without in-depth attention, suggesting lower attention levels, possibly due to insufficient attractiveness or inconspicuous presentation methods. In contrast, the canal cultural landscape had the highest average fixation duration and fixation count, and a relatively low saccade count. This indicates that the landscape’s distinct features attracted significant attention from participants, who were willing to spend more time carefully observing it and did not need to frequently shift their gaze. This reflects its strong appeal and clear visual focus, allowing participants to perceive the cultural value and charm embodied in the landscape. The historical building landscape exhibited the lowest saccade count, with a relatively higher number of fixation points and an average fixation duration at a moderate level, indicating that the information is easily captured and the visual focus is clear. Certain key elements within this landscape are highly attractive to participants, but since these elements have a moderate amount of information or are easily understood, participants can achieve satisfaction without prolonged examination, achieving a good balance between information acquisition efficiency and visual comfort. The modern commemorative landscape had lower saccade counts, fixation counts, and average fixation durations, indicating that this landscape failed to effectively attract participants’ attention and interest. The historical street landscape had a significantly higher saccade count compared to other areas, with a relatively high average fixation duration and a moderate fixation count, suggesting that the landscape provides rich visual information and diverse elements. It reflects that observers actively search for and identify different elements within the landscape during the viewing process, while also conducting detailed observations of specific points of interest.

3.2. Eye Tracking Heatmaps

Eye movement indicators are strongly correlated with environmental attributes and visual stimuli [60]. While eye tracking data can capture variations in visual perception among individuals when viewing images, they do not directly identify the specific landscape components that are most visually engaging. Fixation heatmaps, serving as a straightforward and visual analytical tool, can assist in determining which landscape elements are more likely to draw observers’ attention [61]. In this study, D-LAB 3.72 software was used to visualize the location and duration of fixations among all participants to generate heatmaps of the landscapes (Figure 7), which intuitively display the regions of interest for the subjects.

It can be observed that the visual focus of the historical building landscape is relatively concentrated, with fixations mainly concentrated on the main body of the buildings, including gate towers, archways (text), and architectural roof designs. For the ruins landscape and the historical street landscape, fixations are not concentrated on any specific element. Although the ruins landscape has a relatively clear main body, participants are not very familiar with this landscape, resulting in no distinct point of attraction and fixations scattered across the walls and surrounding landscape elements. The historical street landscape, which is primarily centered on the district, mainly focuses on the midpoints of the roads, surrounding buildings, shops, and so on, and is also relatively dispersed. In the canal cultural landscape, fixations are more concentrated on the buildings and more scattered on the water surface, with the main points of fixation being on the buildings, vegetation, and reflections on the water. For the modern commemorative landscape, fixations are more concentrated on single commemorative sculptures, while in landscapes with more architectural elements such as memorial halls, fixations are more dispersed.

3.3. Analysis of Landscape Preference Questionnaire Evaluation Results

Through the SBE–SD questionnaire, the landscape experiences of different types of landscapes were analyzed, with the results presented in

Table 3. Analysis of differences in SBE–SD evaluation indicators among different types of landscapes.

	Relax	Curiosity	Attractiveness	Aesthetics	Characteristics	Cultural
Overall Differences	0.000 **	0.000 **	0.000 **	0.000 **	0.186	0.000 **
Differences Among Landscape Types
Historical Building Landscape (a)	3.53 c	3.82 bcde	3.96 bcd	4.28 bcd	4.17	4.35 bde
Ruins Landscape (b)	3.35 ce	3.37 ac	3.44 ace	3.85 ace	4.05	4.1 a
Canal Cultural Landscape (c)	4.56 abde	4.25 abde	4.6 abde	4.72 abde	4.06	4.17
Modern Commemorative Landscape (d)	3.44 c	3.42 ac	3.52 ace	4 ace	3.95	3.99 a
Historical Street Landscape (e)	3.61 bc	3.54 ac	3.99 bcd	4.43 bcd	4.05	3.99 a

Note: ** indicates a significant difference at the 0.01 level. For example, in the canal cultural landscape, 4.56 ^abde indicates that the mean value for relaxation in the canal cultural landscape is 4.56, and abde signifies that there are significant differences between the canal cultural landscape and the historical building landscape, ruins landscape, modern commemorative landscape, and historical street landscape. The interpretation of other values follows the same method. Among them, ‘aesthetics’ primarily reflects the artistic and cultural value of the landscape, while ‘attractiveness’ reflects the degree of visual and emotional appeal the landscape holds for viewers.

(Figure 8). The historical building landscape, canal cultural landscape, and historical street landscape received higher scores, with SBE average ratings ranging from 4.28 to 4.72. In contrast, the modern commemorative landscape and ruins landscape received relatively lower scores, with SBE average ratings of 4 and 3.85, respectively. Overall, all ratings were above 3.8, indicating that participants generally held a positive evaluation of the landscapes in Tai’erzhuang Ancient City. Among them, the canal cultural landscape received the highest score (4.72), while the ruins landscape received the lowest score (3.85). Specifically, participants had the most satisfactory viewing experience with the canal cultural landscape, whereas the ruins landscape provided a less satisfactory experience. Combining the SBE–SD evaluation further revealed significant differences among the five types of landscapes through one-way ANOVA, except for the uniqueness aspect. The canal cultural landscape received positive evaluations in terms of relaxation, curiosity, and attractiveness, thus yielding an overall satisfactory viewing experience. The historical building landscape and historical street landscape also had relatively good evaluations in these three aspects, while the historical street landscape was somewhat lacking in cultural aspects. The historical building landscape and ruins landscape had higher ratings in cultural aspects, but compared to the ruins landscape, which lacked a distinct point of attraction, the viewing experience was less favored, resulting in deficiencies in overall relaxation, curiosity, and attractiveness. The modern commemorative landscape received a rather mediocre overall evaluation, indicating that participants did not have significant aesthetic fluctuations for modern landscapes. In summary, each type of landscape has its unique characteristics, such as cultural aspects, attractiveness, and aesthetic appeal, and differences in participants’ interests and viewing experiences were observed through SBE–SD questionnaire analysis. Eye tracking analysis provided insights into the visual behavior characteristics and points of attention of participants.

3.4. Correlation Analysis Results Between Eye Movement Indicators and Subjective Evaluations

In terms of eye movement indicators, average fixation duration was significantly negatively correlated with fixation count (r = −0.581, p < 0.01), average saccade duration (r = −0.651, p < 0.01), and saccade count (r = −0.789, p < 0.01). This indicates that when participants dwell longer on a particular location, their fixation count, saccade duration, and saccade count tend to decrease, possibly because they have already acquired and deeply processed sufficient information during the fixation process. Conversely, fixation count was significantly positively correlated with average saccade duration (r = 0.237, p < 0.01) and saccade count (r = 0.347, p < 0.01), as well as average saccade duration with saccade count (r = 0.591, p < 0.01). This suggests that frequent fixation behavior is often accompanied by longer saccade durations and more saccades, likely because participants need to spend more time during the saccade process to find new points of interest and to process and integrate information while exploring more information. All correlations were significant at levels below 0.001, indicating that these relationships are highly reliable statistically (Figure 9).

In the landscape evaluation relationships, relaxation was significantly positively correlated with curiosity (r = 0.684, p < 0.01) and attractiveness (r = 0.795, p < 0.01), indicating that individuals in a relaxed state not only have stronger curiosity but also a more acute perception of the attractiveness of things. However, there was no significant correlation between relaxation and uniqueness, and the relationship with cultural significance (r = 0.081, p < 0.05) was relatively weak. Curiosity also showed a strong positive correlation with attractiveness (r = 0.769, p < 0.01), but the relationships with uniqueness (r = 0.073, p < 0.05) and cultural significance (r = 0.087, p < 0.05) were relatively weak. There was no significant correlation between attractiveness and uniqueness, but there was a weak positive correlation with cultural significance (r = 0.092, p = 0.01). Notably, there was a significant positive correlation between uniqueness and cultural significance (r = 0.621, p < 0.01), indicating that things with unique characteristics often contain rich cultural connotations. These relationships between variables reflect to some extent the complexity of human psychology and behavioral diversity (Figure 9).

In the relationship between aesthetic quality and eye movement indicators, aesthetic quality was positively correlated with average fixation duration and fixation count (r = 0.160, p < 0.01; r = 0.118, p < 0.01), significant at the 0.01 level. It was negatively correlated with saccade count and average saccade duration (r = −0.101, p < 0.01; r = −0.125, p < 0.01), yet significant at the 0.01 level. This indicates that landscape types with higher aesthetic quality can attract visual attention, guide visual focus, and influence eye movement behavior, increasing average fixation duration and fixation count while reducing saccade count and average saccade duration (Figure 9).

In the relationship between aesthetic quality and subjective landscape evaluation, aesthetic quality was significantly positively correlated with relaxation, curiosity, attractiveness, uniqueness, and cultural significance (r = 0.318, p < 0.01; r = 0.265, p < 0.01; r = 0.396, p < 0.01; r = 0.119, p < 0.05; r = 0.153, p < 0.01), although the strength of these correlations varied (Figure 9). Among them, the correlation between aesthetic quality and attractiveness was the strongest, indicating a high degree of association between the two in visual aesthetics. Aesthetic quality also had strong correlations with relaxation and curiosity, indicating that individuals’ psychological states and curiosity significantly influence aesthetic evaluation. The correlations between aesthetic quality and uniqueness as well as cultural significance were relatively weaker but still significant, indicating that these factors also influence individuals’ aesthetic preferences to some extent.

In the relationship between subjective landscape evaluation and eye movement indicators, relaxation, curiosity, and attractiveness were positively correlated with average fixation duration (r = 0.076, p < 0.05; r = 0.072, p < 0.05; r = 0.120, p < 0.01) and fixation count (r = 0.164, p < 0.01; r = 0.147, p < 0.01; r = 0.128, p < 0.01), showing significance at the 0.05 and 0.01 levels, respectively. This suggests that these characteristics may prompt individuals to fixate more frequently and for longer durations on the landscape to acquire more information. However, these characteristics were not significantly correlated with average saccade duration and saccade count, indicating that they do not significantly influence individuals’ saccadic behavior. In contrast, uniqueness was positively correlated with average saccade duration (r = 0.130, p < 0.01) and saccade count (r = 0.098, p < 0.01), suggesting that landscapes with strong uniqueness may arouse individuals’ desire to explore, prompting more saccadic behavior. Additionally, cultural significance was positively correlated with saccade count (r = 0.117, p < 0.01), indicating that landscapes with strong cultural significance may arouse individuals’ interest and curiosity, prompting more saccadic behavior to gain a deeper understanding of their cultural connotations, although this effect is not significant for fixation behavior (Figure 9).

4. Discussion

4.1. Differences in Visual Attention Among Different Types of Landscapes in Historic and Cultural Cities

This study utilized eye tracking technology to comprehensively examine the visual behavior characteristics of participants when they encountered five different types of landscapes in historic and cultural cities. In a study on the landscapes of Lijiang Ancient Town and Yulong Snow Mountain in China, it was found that landscapes with cultural characteristics scored higher in value assessments, while those that attracted more direct visual attention also received higher scores in aesthetic and cultural value assessments [62]. In a study conducted at Wanshou Palace in Nanchang, China, it was found that participants’ fixation durations on different district elements varied, with longer fixation times on architectural elements and shorter fixation times on commercial elements [34]. In the landscape study of Tai’erzhuang Ancient City, it was found that different types of landscapes significantly differ in their impact on participants’ visual attention. Specifically, the ruins landscape had lower fixation counts and average fixation durations but a higher saccade count, indicating that participants quickly scanned this landscape without in-depth attention. This may be related to the inconspicuous presentation or insufficient attractiveness of the ruins landscape [37]. In contrast, the canal cultural landscape had the highest average fixation duration and fixation count, with a relatively low saccade count, indicating that the landscape’s distinct features attracted significant attention from participants [57]. These differences not only uncover participants’ visual preferences for various landscape types but also highlight the significance of efficient information transmission and strategic visual focus in landscape design to capture viewers’ attention. The historical building landscape and the historical street landscape also exhibited unique visual behavior patterns. The historical building landscape had a significantly higher fixation count than other types of landscapes, with an average fixation duration in the mid-range, while its saccade frequency was significantly lower than other landscape types. This indicates that its information is easily captured and the visual focus is clear. The historical street landscape, on the other hand, provided rich visual information and diverse elements. Participants actively searched for and identified different elements during the viewing process, while also conducting detailed observations of specific points of interest. These findings further support the importance of element diversity and clear visual focus in landscape design for enhancing viewers’ visual experience (Figure 10).

4.2. Differences and Similarities in the Attractive Features of Landscape Types in Historic and Cultural Cities

In the research on landscape types in historic and cultural cities, it was discovered that fixation duration is a crucial visual behavior influencing preference. Studies have indicated that artificial elements tend to attract more attention than natural elements. Moreover, increasing the proportion of water bodies and reflective elements can result in higher evaluation scores [57]. Some scholars have conducted more detailed research on historical buildings and found that participants mainly focus on windows and doors, followed by walls, and then floors and roofs. This order of attention allocation reflects the varying degrees of attractiveness of different elements to participants. In scenes of modern, neutral, and traditional styles, the attention allocation in modern style scenes is more dispersed, while that in traditional style scenes is more balanced [33]. This study employed eye tracking heatmaps in conjunction with the SBE–SD subjective landscape evaluation method to uncover both the differences and similarities in the attractiveness of landscape elements in historic and cultural cities. The visual range of the historical building landscape is relatively concentrated, mainly focusing on the main body of the buildings, such as gate towers, archways, and building roofs, indicating that these elements have significant attractiveness. The fixation points of the canal cultural landscape are more dispersed but mainly concentrated on buildings, vegetation, and water reflections, reflecting the attractiveness of the landscape across multiple elements. In contrast, the fixation points of the ruins landscape and the historical street landscape are more scattered, without a clear point of attraction, which may be related to the layout of the landscape elements and the way information is conveyed.

In terms of landscape evaluation, the canal cultural landscape received high ratings for relaxation, curiosity, and attractiveness, while the ruins landscape scored lower in these aspects. This further confirms the significant impact of landscape element layout, information transmission efficiency, and visual focus settings on viewers’ experiences and perceptions. In terms of eye tracking, positive emotional experiences were positively correlated with fixation duration (TDF) and negatively correlated with saccade count (ANS). This indicates that during the core phase, positive emotional experiences are associated with more stable visual search behavior and longer fixation durations [37]. It is worth noting that, despite differences in attractive features among different landscape types, they all exhibit a certain degree of attractiveness in terms of cultural significance, reflecting the shared cultural value of landscapes in historic and cultural cities. This suggests that in the development of historic and cultural cities, the protection and inheritance of local elements are crucial for enhancing visitor experiences and promoting sustainable tourism development [23] (Figure 7).

4.3. Landscape Enhancement, Optimization, and Conservation Strategies

Tai’erzhuang Ancient City is rich in genuine historical content and is widely acknowledged for its scenic value. Nevertheless, it currently lacks an integrated development strategy that fully exploits the distinctive characteristics of its landscapes while ensuring comprehensive preservation. The study suggests proposing enhancement strategies for the landscapes of historic and cultural cities. The specific content of the proposed landscape enhancement and conservation strategies is as follows:

• Enhance Visual Foci of Landscape Elements: For landscape types with insufficient attractiveness, such as ruins landscapes, efforts should be made to optimize the layout and enhance visual foci to capture more attention from the audience. This can be achieved by setting up conspicuous signs and adding interactive elements to guide the viewers’ gaze and arouse their curiosity.
• Improve Information Transmission Efficiency: In landscape design, emphasis should be placed on the efficiency and clarity of information transmission. While the arrangement of landscape elements and provision of visual guidance can potentially enhance viewers’ experience, it is crucial to ensure that such practices are in harmony with the protection and preservation of cultural heritage landscapes. In the context of international heritage policies, it is recommended to focus on enhancing the landscape experience through educational and interpretive approaches that respect the authenticity and integrity of the heritage site. This can be achieved by providing informative signage, guided tours, and interactive digital content that deepen visitors’ understanding and appreciation of the landscape’s cultural and historical significance, thereby enriching their overall experience without compromising the landscape’s heritage values.
• Protect and Highlight Landscape Features: Different landscape types possess unique characteristics and cultural connotations, which are crucial for attracting audiences. During the enhancement process, it is essential to focus on protecting and highlighting these distinctive elements to boost the landscapes’ appeal and uniqueness. Furthermore, effort should be made to preserve the original natural landscapes and cultural relics, enhance the quality of local tourism performances, and reinforce the attachment to film effects and celebrity effects. Local elements should be integrated into transportation, service facilities, accommodation, and shopping to strengthen the unique brand [20].
• Enhance Landscape Diversity and Hierarchy: Landscape types such as historical street landscapes, which provide rich visual information, demonstrate that diversity and hierarchy are significant for enhancing the audience experience. In landscape design, emphasis should be placed on the diversity and hierarchy of landscape elements to offer a richer and more interesting viewing experience. Culture-oriented urban space production in ancient cities helps to overcome the dilemma of “constructive destruction” and “protective decay”, achieving the revitalization of ancient cities.
• Enhance Landscape Interactivity with Technological Approaches: With the development of technology, advanced techniques such as Virtual Reality (VR) and Augmented Reality (AR) can be utilized to enhance the interactivity and entertainment of landscapes, thereby attracting more attention and participation from audiences.

Although the ancient city of Taierzhuang itself possesses certain unique characteristics, this hybrid research method combining eye tracking with Semantic Differential analysis demonstrates potential adaptability to diverse heritage scenarios. The visual preference evaluation framework applied in this study can be extended to other canal cities by adjusting cultural landscape indicators. The modular design of this experimental protocol—including standardized image sampling and quantitative fixation analysis—enables its replication in different geographical contexts, provided that local heritage value elements are appropriately incorporated.

4.4. Limitations

• Participant Selection: This study assessed the experiences of various landscape types in Tai’erzhuang Ancient City but did not examine demographic factors like gender and age among participants. The preferences of visitors who have previously been to Tai’erzhuang Ancient City might vary, particularly in their appreciation of architectural and cultural landscapes. Future research could offer a more comprehensive evaluation of experiences by incorporating participants with diverse demographic backgrounds.
• Observation Indicator Limitations: This study mainly utilized eye tracking devices to record eye movements across different landscape types, but was unable to assess other indicators. To enhance the precision of evaluative outcomes, future studies may consider combining eye tracking data with other physiological measurement techniques, such as electroencephalography (EEG), to enable a more in-depth and holistic analysis of diverse landscape types and their influence on audience experience.

5. Conclusions

This study employs a combination of eye tracking technology and subjective landscape evaluation methods to conduct an in-depth investigation of five types of historic and cultural city landscapes within the Tai’erzhuang Ancient City. It reveals the differences in visual attention elicited by different landscape types among participants, as well as the variations and similarities in the attractiveness of landscape elements.

The findings are as follows: 1. The eye tracking data indicate significant differences among different landscape types in terms of the number of fixations, average fixation duration, and number of saccades. 2. The subjective landscape evaluation method using the SBE–SD scale further corroborates the conclusions drawn from the eye tracking data, showing that landscape types with clear visual focal points and rich cultural connotations are more favored by participants. 3. The fixation points vary across different types of landscapes. 4. A significant correlation exists between eye tracking data and subjective evaluation results, further validating the scientific validity and effectiveness of the connection between subjective landscape evaluation methods and eye tracking analysis in the context of historic and cultural landscapes. 5. Strategies for the enhancement and optimized protection of historic and cultural city landscapes are proposed. These strategies aim to optimize landscape design while preserving the cultural heritage and visual integrity of the landscape, in alignment with international policies on heritage landscapes. This dual focus ensures the sustainable development of historic and cultural cities without compromising the landscape’s heritage values.

Based on the research findings, it is recommended that in the future development and conservation of historical and cultural city landscapes, the cultural connotations of these landscapes should be thoroughly explored and integrated into the design. This can be achieved by installing cultural display boards, interpretive systems, and other methods to convey the historical stories and cultural values behind the landscapes to visitors. At the same time, protective measures should be strengthened by formulating comprehensive conservation plans for historical and cultural city landscapes, clearly defining the scope, objectives, and methods of protection. The enhanced monitoring and evaluation of landscape cultural heritage should be implemented to promptly identify and address potential conservation issues. The principle of conservation priority should be adhered to, ensuring that new development projects are harmoniously integrated with the landscape features.