Next Article in Journal
An Analytical Method Evaluating the Evolution of Group Effect for Vertically Loaded Pile Groups Subjected to Tunnel Excavation
Previous Article in Journal
Analysis of Breath-Holding Capacity for Improving Efficiency of COPD Severity-Detection Using Deep Transfer Learning
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Applied Sciences
Volume 13
Issue 1
10.3390/app13010515
applsci-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Vegetation Classification in Urban Areas by Combining UAV-Based NDVI and Thermal Infrared Image

by
Geunsang Lee
1,
Gyeonggyu Kim
2,
Gyeongjo Min
2,
Minju Kim
3,
Seunghyun Jung
3,
Jeewook Hwang
4,* and
Sangho Cho
2,5,*
1
Department of Cadastre & Civil Engineering, Vision College of Jeonju, Jeonju 54896, Republic of Korea
2
Department of Mineral Resources and Energy Engineering, Jeonbuk National University, Jeonju 54896, Republic of Korea
3
Korea Institute of Civil Engineering and Building Technology, Gyeonggi-do 10223, Republic of Korea
4
Department of Urban Engineering, Jeonbuk National Unviersity, Jeonju 54896, Republic of Korea
5
Department of Mineral Resources and Energy Engineering, Energy Storage and Conversion Engineering of Graduate School, Jeonuk National University, Jeonju 54896, Republic of Korea
*
Authors to whom correspondence should be addressed.
Appl. Sci. 2023, 13(1), 515; https://doi.org/10.3390/app13010515
Submission received: 10 November 2022 / Revised: 19 December 2022 / Accepted: 22 December 2022 / Published: 30 December 2022
Browse Figures
Review Reports Versions Notes

Abstract

Vegetation has become very important decision-making information in promoting tasks such as urban regeneration, urban planning, environment, and landscaping. In the past, the vegetation index was calculated by combining images of various wavelength regions mainly acquired from the Landsat satellite’s TM or ETM+ sensor. Recently, a technology using UAV-based multispectral images has been developed to obtain more rapid and precise vegetation information. NDVI is a method of calculating the vegetation index by combining the red and near-infrared bands, and is currently the most widely used. In this study, NDVI was calculated using UAV-based multispectral images to classify vegetation. However, among the areas analyzed using NDVI, there was a problem that areas coated with urethane, such as basketball courts and waterproof coating roofs, were classified as vegetation areas. In order to examine these problems, the reflectance of each land cover was investigated using the ASD FieldSpec4 spectrometer. As a result of analyzing the spectrometer measurements, the NDVI values of basketball courts and waterproof coating roofs were similar to those of grass with slightly lower vegetation. To solve this problem, the temperature characteristics of the target site were analyzed using UAV-based thermal infrared images, and vegetation area was analyzed by combining the temperature information with NDVI. To evaluate the accuracy of the vegetation classification technology, 4409 verification points were selected, and kappa coefficients were analyzed for the method using only NDVI and the method using NDVI and thermal infrared images. Compared to the kappa coefficient of 0.830, which was analyzed by applying only NDVI, the kappa coefficient, which was analyzed by combining NDVI and thermal infrared images, was 0.934, which was higher. Therefore, it is very effective to apply a technology that classifies vegetation by combining NDVI and thermal infrared images in urban areas with many urethane-coated land cover such as basketball courts or waterproof coating roofs.

1. Introduction

The occurrence of natural disasters due to climate change is increasing worldwide. In particular, as the global temperature rises, heat waves’ frequency, intensity, and duration increase causing casualties [1,2,3,4,5]. The frequency and intensity of heat waves are influenced not only by topographical characteristics but also by urban development [6]. Recently, the phenomenon of “Urban Heat Islands” (UHIs) on a regional scale is increasing due to the increase in artificial coverings such as concrete and asphalt, the decrease in vegetation, and the increase in high-rise buildings due to urbanization. The urban heat island (UHIs) effect is caused by changes in surface thermal properties due to artificial covering and cooling reduction due to lack of vegetation. It also accounts for hotter urban areas than nearby rural areas due to anthropogenic heat dissipation and air pollutants, which can increase significantly [7,8,9,10].
Unlike concrete and asphalt, vegetation is a permeable surface and can control the temperature because rainwater cools it [11,12,13]. However, over the past few decades, urbanization has destroyed nearly 88% of the world’s primary vegetation-covered land in urban areas and replaced it with artificial surfaces [14,15]. Therefore, to create a pleasant residential environment in an urban area, it is necessary to understand the vegetation index accurately.
The vegetation index quantifies vegetation by analyzing the difference between near-infrared light, which is strongly reflected by vegetation, and red light, which is strongly absorbed by vegetation [16,17,18,19]. Additionally, it is mainly used in the field of remote sensing. The study of vegetation index using remote sensing technology uses the spectral characteristics of satellite images among the analysis methods using satellite images. The satellite data mainly used are Landsat satellites [20,21,22,23,24], and recently Sentinel satellites [25,26,27,28] and MODIS satellites [16,29,30,31] and images from SPOT satellites [32,33] are also being used.
The most commonly used vegetation index is the NDVI (Normalized Difference Vegetation Index), which was reported by Rouse et al. [34]. In addition, GNDVI (Green Normalized Difference Vegetation Index), BNDVI (Blue Normalized Difference Vegetation Index), RGBVI (Red Green Blue Vegetation Index), GRVI (Green Red Vegetation Index), NDRE (Normalized Difference Red Edge), EVI (Enhanced Vegetation Index), SAVI (Soil Adjusted Vegetation Index), NDMI (Normalized Difference Moisture Index), and other vegetation index have been reported.
In particular, GNDVI is an index that responds sensitively to changes in chlorophyll in plants as one of the modified forms of NDVI. In SAVI, the soil adjustment factor L is used to compensate for the NDVI that may be affected by soil brightness in an area where vegetation density is not high.
Recently, with the development of unmanned aerial vehicles (UAV) and high-resolution cameras, the application of environmental monitoring is increasing as an important supplement to the existing satellite remote sensing [35,36,37,38,39].
The vegetation index study using UAV is as follows. Candiago et al. [40] applied a UAV equipped with a multispectral camera for NDVI, GNDVI, and SAVI evaluation, and Zhou et al. [41] applied a UAV to predict rice grain yield and determine the optimal period and optimal VI. Mazzia et al. [42] acquired high-resolution images through UAV to solve the problem of limited monitoring due to the low resolution of existing satellite images in specific agricultural applications. Song and Park et al. [43] determined the effective vegetation index for detecting aquatic plants in small reservoirs through multispectral UAV. Kerkech et al. [44] determined the vegetation index detect disease in vine leaves using high-resolution UAV images. Yu et al. [45] proposed a Soil Salinity Retrieval Index (SSRI) for monitoring the salinity of saline soils using UAVs. Qiao et al. [46] proposed a method for accurately monitoring corn growth and estimating the Leaf Area Index (LAI) based on a UAV multispectral sensor. Zhang et al. [47] proposed a novel complex vegetation index including critical physiological parameters of plants and chlorophyll fluorescence caused by the sun.
Recently, interest in urban regeneration projects for disasters such as heat waves and the urban heat island phenomenon has been increasing in Korea. Overseas, there is an eco-friendly urban regeneration plan at South East Wedge in Edinburgh and Castle Value in Birmingham has an urban regeneration project with a pleasant and safe environment [48]. In order to plan and promote urban regeneration projects like this, data to support decision-making is needed. Vegetation provides a comfortable living environment from heat waves and plays an important role in environmentally friendly urban regeneration plans [49,50].
The idea for this study was discovered by accident. NDVI was calculated using a UAV to obtain vegetation information of an apartment complex necessary for the urban regeneration project. However, it was confirmed that the NDVI value appeared higher in the basketball court and the waterproof coating roof coated with the urethane. Therefore, a method was sought to effectively remove areas that were incorrectly classified as vegetation. As a result, considering the fact that the surface temperature of the basketball court and the waterproof coating roof is very high compared to the vegetation areas such as trees and grass land, the temperature was analyzed from a UAV equipped with a thermal infrared camera. Additionally, by combining the analyzed temperature information with NDVI, a study was conducted to create a highly accurate vegetation map.
This study aims to develop a technology to effectively extract vegetation information by analyzing UAV images captured with multispectral and thermal infrared cameras for urban areas including basketball courts and the waterproof coating roofs covered with urethane. Finally, vegetation information analyzed by combining NDVI and thermal infrared images was compared with vegetation information to which only NDVI was applied to evaluate accuracy.

2. Materials and Methods

2.1. Research Process

Figure 1 shows the process of conducting this study.
First, the study site was selected, and the coordinates of the ground control point (GCP) were acquired by performing a VRS (Virtual References Station) survey to secure the location accuracy of the UAV image.
Additionally, Canon S110 NIR camera (Jeonju, South Korea) and thermoMAP camera (Jeonju, South Korea) were mounted on UAV to obtain multispectral and thermal infrared images, respectively. In addition, a flight plan was established using eMotion S/W to set the UAV flight altitude and speed, and the longitudinal and lateral overlap of images.
By merging images using Pix4D Mapper S/W, orthomosaic images for each blue, green, red, near infrared, and thermal infrared band were created. Additionally, by combining the images for each band using ArcGIS S/W, NDVI value was calculated. Field surveys were conducted to verify the NDVI results analyzed through UAV images, and a method to solve the problem of misclassification as vegetation such as basketball courts and waterproof coating roofs was studied.
Finally, a new technology was developed to analyze vegetation in urban areas by combining NDVI and thermal infrared images.

2.2. Study Area

Figure 2 shows the part of the apartment complex of Jeonju-si located in Jeollabuk-do, Korea that was selected in this study as the target site. The area is composed of vegetation areas, such as trees and grass lands, as well as non-vegetation areas, such as basketball court, building, road, and parking lot.

2.3. UAV Image Processing

In order to obtain an image of the study site, a Canon S110 NIR and ThermoMAP camera were mounted on the eBee model manufactured by SenseFly in Switzerland as shown in Figure 3.
The fixed-wing UAV, called eBee, can capture an area of about 12 km2 with a flight of more than 40 min, and has the advantage of securing a resolution of 1.5 cm when the flight altitude is low [51]. The Canon S110 NIR camera is used to calculate various vegetation indices because it can obtain RGB and near-infrared (NIR) bands, and the thermoMAP camera was used to create temperature information.
In order to establish a UAV flight plan such as flight altitude, resolution, longitudinal overlap, and lateral overlap, eMotion S/W was used as shown in Figure 4.
Since the Canon S110 NIR camera (Table 1) has a high resolution, images with a resolution of 4 cm were taken by designing 80% and 70% longitudinal and lateral overlap, respectively, at a flight altitude of 115 m as shown in Figure 4a. In addition, since the thermoMAP thermal infrared camera has a relatively low resolution, images with a resolution of 15 cm were taken by designing 95% and 70% longitudinal and lateral overlap, respectively, at a flight altitude of 80 m as shown in Figure 4b.
Table 2 shows the details of designing a flight plan by mounting Canon S110 NIR and thermoMAP thermal infrared cameras, respectively [51].
To secure the positional accuracy of the UAV image, six ground control points were selected as in Table 3, and Transverse Mercator (TM) coordinates of the GRS80 ellipsoid were acquired through VRS survey. The captured image has Universal Transverse Mercator (UTM) coordinates of the WGS84 ellipsoid through a Global Positioning System (GPS) receiver equipped with a UAV. Therefore, in order to convert into TM coordinates of the GRS80 ellipsoid used in Korea, the survey results were matched to ground control points. Figure 5 shows the VRS surveying for ground control point (No.1).
As for the video taken by the UAV, as shown in Figure 6, 70 red, green, blue, and near-infrared (NIR) bands pictures were taken with a Canon s110 NIR camera and 3131 thermal-infrared bands pictures used a thermoMAP camera. The image was merged using Pix4D Mapper S/W, and the ‘GCP Manger’ function was used to match the surveying coordinates of the ground control point to increase the positional accuracy of the UAV image. Through this process, orthomosaic images of red, green, blue, near-infrared (Nir), and thermal-infrared bands required for vegetation index analysis were constructed (Figure 7).

2.4. Spectroradiometer

This study used the equipment of ASD FieldSpec 4, and the performance of the equipment is shown in Table 4. The reflectance of each land cover was investigated using the related equipment. When measuring reflectance using a spectrometer, errors may occur due to the influ-ence of weather, such as the amount of light and humidity at the time of measurement. In this study, the white standard was used, as shown in Figure 8, and the correction was made just before the measurement of the land cover. The reflectance of each land cover was investigated using a calibrated spectrometer.
Figure 9 is a field photograph measuring reflectance by wavelength using an ASD FieldSpec4 spectrometer for tree, grass, basketball court, waterproof coating roof, and road.

3. Result and Discussion

3.1. NDVI Analyzed by UAV-Based Multispectral Image

Using the RGB and near-infrared images taken by UAV, an NDVI map was created as shown in Figure 10, and the statistical characteristics of the NDVI vegetation index are shown in Table 5. In order to classify vegetation areas from the NDVI map, the threshold for NDVI must be determined.
In this study, vegetation areas were analyzed while changing the NDVI threshold value from 0.2 to 0.5 as shown in Figure 11, and the vegetation area according to the threshold value of NDVI is shown in Table 6. Suppose the NDVI threshold is set to 0.2. In that case, the vegetation area is 49,344 m2, and the NDVI threshold is 0.3, the vegetation area is 37,201 m2, and the NDVI threshold is 0.4, the vegetation area is 25,586 m2, and if the NDVI threshold is 0.5, the vegetation area is 14,320 m2 classified. When the threshold value of NDVI is set high, it can be seen that only areas with good plant growth are classified as vegetation areas.
In this study, NDVI was analyzed using UAV images taken with a Canon S110 NIR camera mounted on 21 July 2020, and vegetation areas were classified by applying various threshold values of NDVI. The weather conditions at the time of UAV image capture were average cloudiness of 8.1, a maximum temperature of 30 °C, a minimum temperature of 21.5 °C, and an average temperature of 24.9 °C.
July, when UAV photographing was conducted, is the season when the growth vitality of vegetation is very good, and as a result of confirming the vegetation area through field survey, the vegetation map when the threshold value of NDVI is 0.2 or more shows the vegetation area of the target area most accurately.
Figure 12 shows the location of grass and trees in the vegetation map analyzed with the threshold value set to 0.2 or higher, and it was confirmed that the vegetation condition of the region is consistent with the vegetation map analyzed using UAV images.
However, during the field investigation, it was confirmed that the basketball court and the waterproof coating roof were classified as vegetation as shown in Figure 13.
The NDVI analyzes the vitality of vegetation using the wavelength characteristics of near-infrared and red band, so green artificial structures should not be classified as vegetation. The problem of vegetation misclassification to the basketball court and the waterproof coating roof covered with urethane coatings could not be found in previous studies.

3.2. Spectral Comparison and Vegetation Misclassification Analysis Using a Spectrometer

As the basketball court and waterproof coated roof described above were classified as vegetation, the reflectance of each land cover was investigated using a spectrometer. Figure 14 shows the reflectance at each wavelength for each land cover. In particular, the measured results confirmed that vegetation was significantly different in the wavelength range of 700–900 nm compared to a basketball court and a waterproof coated roof. Table 7 shows the reflectance and calculated NDVI values for specific wavelengths of each land cover, with the blue wavelength at 475 nm, the green wavelength at 560 nm, the red wavelength at 668 nm, and the NIR wavelength at 840 nm. NDVI uses specific values of the red wavelength of 668 nm and NIR wavelength of 840 nm. The NDVI value is a calculation formula of R and NIR, and if the calculated result is similar to vegetation, it can be classified as vegetation. Therefore, since the spectroscopic results can confirm a wide range of wavelength bands, differences in each land cover can be identified. However, if vegetation is classified only through NDVI, only R and NIR values are used, so if the calculated results are similar, it may be misclassified as vegetation, as in this study.

3.3. Vegetation Classification by Combining UAV-Based NDVI and Thermal Infrared Image

As a result of trying to send ideas to solve the problem of basketball courts and waterproof-coated roofs misclassified as vegetation, images were taken using a thermal infrared ThermoMAP camera by paying attention to the temperature distribution characteristics of the urethane-coated area.
Thermal infrared images taken with a thermoMAP camera were merged using Pix4D Mapper S/W, and then a temperature map was created using ArcGIS S/W, as shown in Figure 15. In order to remove basketball courts and waterproof coating roofs that are incorrectly classified as vegetation based on the temperature information analyzed by UAV, it is necessary to set the temperature threshold between artificial structures and vegetation.
The study by Lee et al. (2021) [52] analyzed the temperature distribution characteristics according to the roof material and the type of roof coating. Lee et al. (2021) [52] analyzed the temperature characteristics of the urethane coating area when the average cloud cover was 8.5, and the maximum temperature was 31 °C. The average temperature of the urethane coating area showed a difference of about 5 °C from the day’s highest temperature. In this study, when the thermal image was taken, the average cloudiness was 8.1, the maximum temperature was 30 °C, and the basketball court and waterproof coating roof were coated with urethane. Accordingly, the highest temperature on the measurement day was set to 30 °C as the temperature threshold for classification with vegetation by referring to the literature.
Figure 16 shows the vegetation map classified by combining NDVI and thermal infrared image. It was confirmed that the basketball court and the waterproof coating roof coated with urethane were well classified as non-vegetation areas.
Table 8 shows the results of analyzing the vegetation and non-vegetation area for each apartment complex from the vegetation map finally generated using NDVI and thermal infrared images.
To evaluate the accuracy of the vegetation map analyzed using only the NDVI index and the vegetation map analyzed by combining the NDVI vegetation index with the thermal image, a validation point was selected as shown in Figure 17. For the target area, 4409 verification points were selected at 5 m intervals, and classified into vegetation and non-vegetation through field surveys.
Figure 18 shows the vegetation area classified using only NDVI for the basketball court and the waterproof coating roof, along with the verification points. When applying only NDVI, it was confirmed that the basketball court and the waterproof coating roof coated with urethane were misclassified as vegetation.
Figure 19 shows the vegetation area classified by combining NDVI and thermal infrared images for the basketball court and the waterproof coating roof, along with the verification points. When vegetation was classified by combining NDVI and thermal infrared images, it was confirmed that basketball courts and waterproof coating roofs were well classified as non-vegetation.
In order to compare the vegetation map classified by using only NDVI and the vegetation map classified by combining the NDVI and thermal infrared images, kappa coefficients were analyzed as shown in Table 9. As a result of the analysis, the kappa coefficient analyzed by applying only the NDVI was 0.830, and the kappa coefficient analyzed by combining the NDVI and the thermal image was 0.934. Therefore, it was found that it was more effective to classify vegetation by combining NDVI and thermal image compared to classify vegetation by applying only NDVI in urban areas where urethane coated coverings are distributed.

4. Conclusions

Vegetation is considered a very important factor in urban planning and landscaping, including urban regeneration projects, because vegetation plays a role in providing a comfortable living space for humans.
Satellite images have been mainly used to construct vegetation information on a wide area. However, since the satellite moves only in a fixed orbit, it is difficult to obtain an image at a desired point of view. In particular, there is a limit to obtaining a good quality image due to the influence of clouds. Recently, these problems have been greatly improved by utilizing the UAV system equipped with a multispectral camera.
In this study, NDVI was analyzed using UAV multispectral images to construct vegetation information in urban areas, and in the process, it was accidentally found that NDVI values appeared high in basketball courts and waterproof coating roofs coated with urethane. In order to examine these problems, the reflectance of each land cover such as trees, grass, basketball court, waterproof coating roof, and road was investigated using the ASD FieldSpec4 spectrometer. As a result of the analysis, the NDVI values of basketball courts and waterproof coating roofs were similar to those of grass with slightly lower vegetation.
To solve this problem, the temperature of the target area was analyzed using a UAV system equipped with a thermal infrared camera called themoMAP, and the final vegetation map was created by combining the temperature information and NDVI.
In order to compare the vegetation map classified by using only NDVI and the vegetation map classified by combining the NDVI and thermal infrared images, 4409 validation points were selected and the kappa coefficients were calculated. The kappa coefficient analyzed by applying only the NDVI was 0.830, and the kappa coefficient analyzed by combining the NDVI and the thermal image was 0.934. Therefore, it was found that it was more accurate to classify vegetation by combining NDVI and thermal images than the result of classifying vegetation by applying only NDVI in urban areas where urethane coating was distributed, such as basketball courts and waterproof coating roofs.
Recently, in Korea, waterproof coating roofs are coated with urethane paint having a waterproof effect, and basketball courts coated with urethane are also being constructed in order to provide living sports facilities to residents. Therefore, when analyzing the vegetation information of urban areas necessary for the promotion of work such as urban regeneration, landscaping, and the environment, it is necessary to first investigate the land cover characteristics of the target area. Additionally, if there is a lot of urethane-coated land cover in urban areas, it is suggested that the method of classifying vegetation by combining NDVI and thermal infrared images will be effective.

Author Contributions

Conceptualization, G.L., J.H. and S.C.; Investigation, G.L., G.K. and G.M.; software. G.L.; supervision, J.H. and S.C.; writing—original draft, G.L., G.K. and S.C.; writing—review and editing, G.K., M.K. and S.J. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by a grant (22TSRD-B151228-04) from Urban Declining Area Regenerative Capacity-Enhancing Technology Research Program funded by Ministry of Land, Infrastructure and Transport of Korean government.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Fahad, M.G.R.; Nazari, R.; Bhavsar, P.; Jalayer, M.; Karimi, M. A Decision-Support Framework for Emergency Evacuation Planning during Extreme Storm Events. Transp. Res. D Transp. Env. 2019, 77, 589–605. [Google Scholar] [CrossRef]
  2. Harrington, L.J.; Ebi, K.L.; Frame, D.J.; Otto, F.E.L. Integrating Attribution with Adaptation for Unprecedented Future Heatwaves. Clim. Change 2022, 172, 2. [Google Scholar] [CrossRef]
  3. Rousi, E.; Kornhuber, K.; Beobide-Arsuaga, G.; Luo, F.; Coumou, D. Accelerated Western European Heatwave Trends Linked to More-Persistent Double Jets over Eurasia. Nat. Commun. 2022, 13, 3851. [Google Scholar] [CrossRef] [PubMed]
  4. Sabrin, S.; Karimi, M.; Nazari, R. Developing Vulnerability Index to Quantify Urban Heat Islands Effects Coupled with Air Pollution: A Case Study of Camden, NJ. ISPRS Int. J. Geo-Inf. 2020, 9, 349. [Google Scholar] [CrossRef]
  5. Uddin, M.N.; Saiful Islam, A.K.M.; Bala, S.K.; Islam, G.M.T.; Adhikary, S.; Saha, D.; Haque, S.; Fahad, M.G.R.; Akter, R. Mapping of Climate Vulnerability of the Coastal Region of Bangladesh Using Principal Component Analysis. Appl. Geogr. 2019, 102, 47–57. [Google Scholar] [CrossRef]
  6. Park, K. Analysis on the Effects of Land Cover Types and Topographic Features on Heat Wave Days. J. Korean Assoc. Geogr. Inf. Stud. 2016, 19, 76–91. [Google Scholar] [CrossRef]
  7. Peng, J.; Hu, Y.; Dong, J.; Liu, Q.; Liu, Y. Quantifying Spatial Morphology and Connectivity of Urban Heat Islands in a Megacity: A Radius Approach. Sci. Total Environ. 2020, 714, 136792. [Google Scholar] [CrossRef]
  8. Sabrin, S.; Karimi, M.; Nazari, R.; Pratt, J.; Bryk, J. Effects of Different Urban-Vegetation Morphology on the Canopy-Level Thermal Comfort and the Cooling Benefits of Shade Trees: Case-Study in Philadelphia. Sustain. Cities Soc. 2021, 66, 102684. [Google Scholar] [CrossRef]
  9. Voogt, J.A.; Oke, T.R. Thermal Remote Sensing of Urban Climates. Remote Sens. Environ. 2003, 86, 370–384. [Google Scholar] [CrossRef]
  10. Ziter, C.D.; Pedersen, E.J.; Kucharik, C.J.; Turner, M.G. Scale-Dependent Interactions between Tree Canopy Cover and Impervious Surfaces Reduce Daytime Urban Heat during Summer. Proc. Natl. Acad. Sci. USA 2019, 116, 7575–7580. [Google Scholar] [CrossRef]
  11. Ellison, D.; Morris, C.E.; Locatelli, B.; Sheil, D.; Cohen, J.; Murdiyarso, D.; Gutierrez, V.; van Noordwijk, M.; Creed, I.F.; Pokorny, J.; et al. Trees, Forests and Water: Cool Insights for a Hot World. Glob. Environ. Change 2017, 43, 51–61. [Google Scholar] [CrossRef]
  12. Shamsudeen, M.; Padmanaban, R.; Cabral, P.; Morgado, P. Spatio-Temporal Analysis of the Impact of Landscape Changes on Vegetation and Land Surface Temperature over Tamil Nadu. Earth 2022, 3, 614–638. [Google Scholar] [CrossRef]
  13. Su, Y.; Liu, L.; Wu, J.; Chen, X.; Shang, J.; Ciais, P.; Zhou, G.; Lafortezza, R.; Wang, Y.; Yuan, W.; et al. Quantifying the Biophysical Effects of Forests on Local Air Temperature Using a Novel Three-Layered Land Surface Energy Balance Model. Environ. Int. 2019, 132, 105080. [Google Scholar] [CrossRef] [PubMed]
  14. Seto, K.C.; Güneralp, B.; Hutyra, L.R. Global Forecasts of Urban Expansion to 2030 and Direct Impacts on Biodiversity and Carbon Pools. Proc. Natl. Acad. Sci. USA 2012, 109, 16083–16088. [Google Scholar] [CrossRef]
  15. D’Amour, C.B.; Reitsma, F.; Baiocchi, G.; Barthel, S.; Güneralp, B.; Erb, K.H.; Haberl, H.; Creutzig, F.; Seto, K.C. Future Urban Land Expansion and Implications for Global Croplands. Proc. Natl. Acad. Sci. USA 2017, 114, 8939–8944. [Google Scholar] [CrossRef]
  16. Bojanowski, J.S.; Sikora, S.; Musiał, J.P.; Woźniak, E.; Dąbrowska-Zielińska, K.; Slesiński, P.; Milewski, T.; Łączyński, A. Integration of Sentinel-3 and MODIS Vegetation Indices with ERA-5 Agro-Meteorological Indicators for Operational Crop Yield Forecasting. Remote Sens. 2022, 14, 1238. [Google Scholar] [CrossRef]
  17. Thomas, J.R.; Gausman, H.W. Leaf Reflectance vs. Leaf Chlorophyll and Carotenoid Concentrations for Eight Crops 1. Agron. J. 1977, 69, 799–802. [Google Scholar] [CrossRef]
  18. Zhang, F.; Zhang, L.W.; Shi, J.J.; Huang, J.F. Soil Moisture Monitoring Based on Land Surface Temperature-Vegetation Index Space Derived from MODIS Data. Pedosphere 2014, 24, 450–460. [Google Scholar] [CrossRef]
  19. Jiang, Z.; Huete, A.R.; Chen, J.; Chen, Y.; Li, J.; Yan, G.; Zhang, X. Analysis of NDVI and Scaled Difference Vegetation Index Retrievals of Vegetation Fraction. Remote Sens. Environ. 2006, 101, 366–378. [Google Scholar] [CrossRef]
  20. Turner, D.P.; Cohen, W.B.; Kennedy, R.E.; Fassnacht, K.S.; Briggs, J.M. Relationships between Leaf Area Index and Landsat TM Spectral Vegetation Indices across Three Temperate Zone Sites. Remote Sens. Environ. 1999, 70, 52–68. [Google Scholar] [CrossRef]
  21. Schultz, M.; Clevers, J.G.P.W.; Carter, S.; Verbesselt, J.; Avitabile, V.; Quang, H.V.; Herold, M. Performance of Vegetation Indices from Landsat Time Series in Deforestation Monitoring. Int. J. Appl. Earth Obs. Geoinf. 2016, 52, 318–327. [Google Scholar] [CrossRef]
  22. Roy, D.P.; Kovalskyy, V.; Zhang, H.K.; Vermote, E.F.; Yan, L.; Kumar, S.S.; Egorov, A. Characterization of Landsat-7 to Landsat-8 Reflective Wavelength and Normalized Difference Vegetation Index Continuity. Remote Sens. Environ. 2016, 185, 57–70. [Google Scholar] [CrossRef] [PubMed]
  23. Liu, J.; Pattey, E.; Jégo, G. Assessment of Vegetation Indices for Regional Crop Green LAI Estimation from Landsat Images over Multiple Growing Seasons. Remote Sens. Environ. 2012, 123, 347–358. [Google Scholar] [CrossRef]
  24. Jiménez-Jiménez, S.I.; de Marcial-Pablo, M.J.; Ojeda-Bustamante, W.; Sifuentes-Ibarra, E.; Inzunza-Ibarra, M.A.; Sánchez-Cohen, I. VICAL: Global Calculator to Estimate Vegetation Indices for Agricultural Areas with Landsat and Sentinel-2 Data. Agronomy 2022, 12, 1518. [Google Scholar] [CrossRef]
  25. Hawryło, P.; Bednarz, B.; Wężyk, P.; Szostak, M. Estimating Defoliation of Scots Pine Stands Using Machine Learning Methods and Vegetation Indices of Sentinel-2. Eur. J. Remote Sens. 2018, 51, 194–204. [Google Scholar] [CrossRef]
  26. Mandal, D.; Kumar, V.; Ratha, D.; Dey, S.; Bhattacharya, A.; Lopez-Sanchez, J.M.; McNairn, H.; Rao, Y.S. Dual Polarimetric Radar Vegetation Index for Crop Growth Monitoring Using Sentinel-1 SAR Data. Remote Sens. Environ. 2020, 247, 111954. [Google Scholar] [CrossRef]
  27. Hill, M.J. Vegetation Index Suites as Indicators of Vegetation State in Grassland and Savanna: An Analysis with Simulated SENTINEL 2 Data for a North American Transect. Remote Sens. Environ. 2013, 137, 94–111. [Google Scholar] [CrossRef]
  28. Jia, M.; Wang, Z.; Wang, C.; Mao, D.; Zhang, Y. A New Vegetation Index to Detect Periodically Submerged Mangrove Forest Using Single-Tide Sentinel-2 Imagery. Remote Sens. 2019, 11, 2043. [Google Scholar] [CrossRef]
  29. Hashemzadeh Ghalhari, M.; Vafakhah, M.; Damavandi, A.A. Agricultural Drought Assessment Using Vegetation Indices Derived from MODIS Time Series in Tehran Province. Arab. J. Geosci. 2022, 15, 412. [Google Scholar] [CrossRef]
  30. Zeng, Y.; Hao, D.; Huete, A.; Dechant, B.; Berry, J.; Chen, J.M.; Joiner, J.; Frankenberg, C.; Bond-Lamberty, B.; Ryu, Y.; et al. Optical Vegetation Indices for Monitoring Terrestrial Ecosystems Globally. Nat. Rev. Earth Environ. 2022, 3, 477–493. [Google Scholar] [CrossRef]
  31. Shamloo, N.; Sattari, M.T.; Apaydin, H. Agricultural Drought Survey Using MODIS-Based Image Indices at the Regional Scale: Case Study of the Urmia Lake Basin, Iran. Theor. Appl. Climatol. 2022, 149, 39–51. [Google Scholar] [CrossRef]
  32. Fensholt, R.; Sandholt, I.; Stisen, S. Evaluating MODIS, MERIS, and VEGETATION Vegetation Indices Using in Situ Measurements in a Semiarid Environment. IEEE Trans. Geosci. Remote Sens. 2006, 44, 1774–1786. [Google Scholar] [CrossRef]
  33. Huong, N.T.T.; Phuong, L.T. Quantify Forest Stand Volume Using SPOT 5 Satellite Image. In Global Changes and Sustainable Development in Asian Emerging Market Economies; Springer Nature: Berlin/Heidelberg, Germany, 2022; Volume 2. [Google Scholar]
  34. Rouse, J.W.; Haas, R.H.; Schell, J.A.; Deering, D.W. Monitoring Vegetation Systems in the Great Plains with ERTS. NASA Special Publication; NASA: Washington, DC, USA, 1974; p. 24.
  35. Guo, Z.; Wang, T.; Liu, S.; Kang, W.; Chen, X.; Feng, K.; Zhang, X.; Zhi, Y. Biomass and Vegetation Coverage Survey in the Mu Us Sandy Land–Based on Unmanned Aerial Vehicle RGB Images. Int. J. Appl. Earth Obs. Geoinf. 2021, 94, 102239. [Google Scholar] [CrossRef]
  36. Yang, Z.; Yu, X.; Dedman, S.; Rosso, M.; Zhu, J.; Yang, J.; Xia, Y.; Tian, Y.; Zhang, G.; Wang, J. UAV Remote Sensing Applications in Marine Monitoring: Knowledge Visualization and Review. Sci. Total Environ. 2022, 838, 155939. [Google Scholar] [CrossRef] [PubMed]
  37. Asadzadeh, S.; de Oliveira, W.J.; de Souza Filho, C.R. UAV-Based Remote Sensing for the Petroleum Industry and Environmental Monitoring: State-of-the-Art and Perspectives. J. Pet. Sci. Eng. 2022, 208, 109633. [Google Scholar] [CrossRef]
  38. Alvarez-Vanhard, E.; Corpetti, T.; Houet, T. UAV & Satellite Synergies for Optical Remote Sensing Applications: A Literature Review. Sci. Remote Sens. 2021, 3, 100019. [Google Scholar] [CrossRef]
  39. Feroz, S.; Dabous, S.A. Uav-Based Remote Sensing Applications for Bridge Condition Assessment. Remote Sens. 2021, 13, 1809. [Google Scholar] [CrossRef]
  40. Candiago, S.; Remondino, F.; de Giglio, M.; Dubbini, M.; Gattelli, M. Evaluating Multispectral Images and Vegetation Indices for Precision Farming Applications from UAV Images. Remote Sens. 2015, 7, 4026–4047. [Google Scholar] [CrossRef]
  41. Zhou, X.; Zheng, H.B.; Xu, X.Q.; He, J.Y.; Ge, X.K.; Yao, X.; Cheng, T.; Zhu, Y.; Cao, W.X.; Tian, Y.C. Predicting Grain Yield in Rice Using Multi-Temporal Vegetation Indices from UAV-Based Multispectral and Digital Imagery. ISPRS J. Photogramm. Remote Sens. 2017, 130, 246–255. [Google Scholar] [CrossRef]
  42. Mazzia, V.; Comba, L.; Khaliq, A.; Chiaberge, M.; Gay, P. UAV and Machine Learning Based Refinement of a Satellite-Driven Vegetation Index for Precision Agriculture. Sensors 2020, 20, 2530. [Google Scholar] [CrossRef]
  43. Song, B.; Park, K. Detection of Aquatic Plants Using Multispectral UAV Imagery and Vegetation Index. Remote Sens. 2020, 12, 387. [Google Scholar] [CrossRef]
  44. Kerkech, M.; Hafiane, A.; Canals, R. Deep Leaning Approach with Colorimetric Spaces and Vegetation Indices for Vine Diseases Detection in UAV Images. Comput. Electron. Agric. 2018, 155, 237–243. [Google Scholar] [CrossRef]
  45. Yu, X.; Chang, C.; Song, J.; Zhuge, Y.; Wang, A. Precise Monitoring of Soil Salinity in China’s Yellow River Delta Using UAV-Borne Multispectral Imagery and a Soil Salinity Retrieval Index. Sensors 2022, 22, 546. [Google Scholar] [CrossRef] [PubMed]
  46. Qiao, L.; Gao, D.; Zhao, R.; Tang, W.; An, L.; Li, M.; Sun, H. Improving Estimation of LAI Dynamic by Fusion of Morphological and Vegetation Indices Based on UAV Imagery. Comput. Electron. Agric. 2022, 192, 106603. [Google Scholar] [CrossRef]
  47. Zhang, J.; Sun, B.; Yang, C.; Wang, C.; You, Y.; Zhou, G.; Liu, B.; Wang, C.; Kuai, J.; Xie, J. A Novel Composite Vegetation Index Including Solar-Induced Chlorophyll Fluorescence for Seedling Rapeseed Net Photosynthesis Rate Retrieval. Comput. Electron. Agric. 2022, 198, 107031. [Google Scholar] [CrossRef]
  48. Deakin, M.; Allwinkle, S. Urban Regeneration and Sustainable Communities: The Role of Networks, Innovation, and Creativity in Building Successful Partnerships. J. Urban Technol. 2007, 14, 77–91. [Google Scholar] [CrossRef]
  49. Templeton, L.K.; Neel, M.C.; Groffman, P.M.; Cadenasso, M.L.; Sullivan, J.H. Changes in Vegetation Structure and Composition of Urban and Rural Forest Patches in Baltimore from 1998 to 2015. For. Ecol. Manage. 2019, 454, 117665. [Google Scholar] [CrossRef]
  50. Trentanovi, G.; Campagnaro, T.; Kowarik, I.; Munafò, M.; Semenzato, P.; Sitzia, T. Integrating Spontaneous Urban Woodlands into the Green Infrastructure: Unexploited Opportunities for Urban Regeneration. Land Use Policy 2021, 102, 105221. [Google Scholar] [CrossRef]
  51. Lee, G.; Choi, M.; Yu, W.; Jung, K. Creation of River Terrain Data Using Region Growing Method Based on Point Cloud Data from UAV Photography. Quat. Int. 2019, 519, 255–262. [Google Scholar] [CrossRef]
  52. Lee, G.S.; Kim, G.G.; Cho, S.H. Temperaure analysis by roof material using UAV-based thermal infrared image. J. Korean Cadastre Inf. Assoc. 2021, 23, 57–72. [Google Scholar] [CrossRef]
Applsci 13 00515 g001
Figure 1. The research process.
Figure 1. The research process.
Applsci 13 00515 g001
Applsci 13 00515 g002
Figure 2. The study area.
Figure 2. The study area.
Applsci 13 00515 g002
Applsci 13 00515 g003
Figure 3. The UAV system.
Figure 3. The UAV system.
Applsci 13 00515 g003
Applsci 13 00515 g004
Figure 4. Set up of UAV flight planning using eMotion S/W.
Figure 4. Set up of UAV flight planning using eMotion S/W.
Applsci 13 00515 g004
Applsci 13 00515 g005
Figure 5. VRS surveying for ground control point (No.1).
Figure 5. VRS surveying for ground control point (No.1).
Applsci 13 00515 g005
Applsci 13 00515 g006
Figure 6. Number of images taken by each camera.
Figure 6. Number of images taken by each camera.
Applsci 13 00515 g006
Applsci 13 00515 g007
Figure 7. Orthoimage results for each camera image.
Figure 7. Orthoimage results for each camera image.
Applsci 13 00515 g007
Applsci 13 00515 g008
Figure 8. Calibration of ASD FieldSpec4 spectrometer.
Figure 8. Calibration of ASD FieldSpec4 spectrometer.
Applsci 13 00515 g008
Applsci 13 00515 g009
Figure 9. Investigation of reflectance by wavelength using ASD FieldSpec4 spectrometer.
Figure 9. Investigation of reflectance by wavelength using ASD FieldSpec4 spectrometer.
Applsci 13 00515 g009
Applsci 13 00515 g010
Figure 10. NDVI map analyzed by UAV-based multispectral image.
Figure 10. NDVI map analyzed by UAV-based multispectral image.
Applsci 13 00515 g010
Applsci 13 00515 g011aApplsci 13 00515 g011b
Figure 11. Vegetation map according to the threshold value of NDVI.
Figure 11. Vegetation map according to the threshold value of NDVI.
Applsci 13 00515 g011aApplsci 13 00515 g011b
Applsci 13 00515 g012
Figure 12. Vegetation map classified by applying NDVI around trees and grass lands.
Figure 12. Vegetation map classified by applying NDVI around trees and grass lands.
Applsci 13 00515 g012
Applsci 13 00515 g013
Figure 13. Vegetation map classified by applying NDVI around basketball court and waterproof coating roof.
Figure 13. Vegetation map classified by applying NDVI around basketball court and waterproof coating roof.
Applsci 13 00515 g013
Applsci 13 00515 g014
Figure 14. Reflectance by wavelength using ASD FieldSpec4 spectrometer.
Figure 14. Reflectance by wavelength using ASD FieldSpec4 spectrometer.
Applsci 13 00515 g014
Applsci 13 00515 g015
Figure 15. Temperature map analyzed from UAV-based thermal infrared image.
Figure 15. Temperature map analyzed from UAV-based thermal infrared image.
Applsci 13 00515 g015
Applsci 13 00515 g016
Figure 16. Vegetation map classified by combining NDVI ad thermal infrared image.
Figure 16. Vegetation map classified by combining NDVI ad thermal infrared image.
Applsci 13 00515 g016
Applsci 13 00515 g017
Figure 17. Validation points selected at 5 m intervals.
Figure 17. Validation points selected at 5 m intervals.
Applsci 13 00515 g017
Applsci 13 00515 g018
Figure 18. Vegetation map classified by applying only NDVI around the basketball courts and the waterproof coating roofs.
Figure 18. Vegetation map classified by applying only NDVI around the basketball courts and the waterproof coating roofs.
Applsci 13 00515 g018
Applsci 13 00515 g019
Figure 19. Vegetation map classified by applying NDVI and thermal infrared image around the basketball courts and the waterproof coating roofs.
Figure 19. Vegetation map classified by applying NDVI and thermal infrared image around the basketball courts and the waterproof coating roofs.
Applsci 13 00515 g019
Table 1. Specifications of Canon S110 NIR camera.
Table 1. Specifications of Canon S110 NIR camera.
CategorySpecification
BandBlue, Green, Red, NIR
Band wavelength(nm)Blue: 450
Green: 550
Red: 625
NIR: 850
Resolution12 Mp
Ground resolution at 100 m3.5 cm/pix
Sensor size7.44 ∗ 5.58 mm
Pixel pitch (width ∗ height)1.33 um
Image formatJPEG/CR2
Table 2. Specifications of multispectral and thermal infrared camera.
Table 2. Specifications of multispectral and thermal infrared camera.
ItemsCanon S110 NIRThermoMAP
Flight altitude114.5 m79.4 m
Ground resolution4.0 cm/px15.0 cm/px
Longitudinal overlap80%95%
Lateral overlap70%70%
Flight time18.5 min20.5 min
Table 3. Information of the ground control points.
Table 3. Information of the ground control points.
No.E (m)N (m)Z (EL m)
1210,054.29358,696.7325.063
2210,001.42358,902.9226.465
3209,871.17358,991.7925.373
4209,576.31359,061.5223.952
5209,596.15358,658.6124.415
6209,893.41358,669.3024.036
Table 4. Specification of ASD FieldSpec4 spectrometer.
Table 4. Specification of ASD FieldSpec4 spectrometer.
ContentsSpecification
Spectral range350–2500 nm
Spectral resolution700 nm (3 nm VNIR)
1400/2100 nm (8 nm SWIR)
Spectral sampling (bandwidth)1.4 nm (350–1000 nm)
1.1 nm (1001–2500 nm)
Scanning time100 milli seconds
Stray light specificationVNIR 0.02%, SWIR 1 & 2 0.01%
Wavelength reproducibility0.1 nm
Weight5.44 kg
Channels2151
Table 5. Statistical value of NDVI map.
Table 5. Statistical value of NDVI map.
Vegetation IndexMinimumMaximum
NDVI−0.2500.863
Table 6. Vegetation area according to the threshold value of NDVI.
Table 6. Vegetation area according to the threshold value of NDVI.
Threshold Value of NDVIVegetation (m2)Non-Vegetation (m2)Sum (m2)
>0.249,34460,669110,013
>0.337,20172,812
>0.425,58684,427
>0.514,32095,693
Table 7. Reflectance by bands (R, G, B and Nir) and NDVI for each land cover using ASD FieldSpec4 spectrometer.
Table 7. Reflectance by bands (R, G, B and Nir) and NDVI for each land cover using ASD FieldSpec4 spectrometer.
Land CoverReflectanceNDVI
BlueGreenRedNir
Tree
(vegetation: high)		0.00980.05230.01330.40390.9362
Grass
(vegetation: medium)		0.03110.07780.04890.28250.7048
Grass
(vegetation: slightly low)		0.02910.06610.09030.23170.4391
Basketball court0.13630.20970.10930.27330.4286
Waterproof
Coating roof		0.07760.10170.05190.12760.4217
Road0.37660.43490.47220.4429−0.0320
Table 8. Vegetation area by apartment complex analyzed by combining NDVI and thermal infrared image.
Table 8. Vegetation area by apartment complex analyzed by combining NDVI and thermal infrared image.
Apartment ComplexVegetation (m2)Non-vegetation (m2)Sum (m2)
A12,98330,161110,013
B15,26730,785
C621314,604
Table 9. Analysis of the kappa coefficient for vegetation classification.
Table 9. Analysis of the kappa coefficient for vegetation classification.
MethodsField
Survey		Vegetation Analysis by UAV ImagesKappa
Coefficient
VegetationNon-
Vegetation		Sum
NDVIVegetation239727726740.830
Non-
vegetation		216471735
Sum248519244409
NDVI
+
Thermal
infrared image		Vegetation26433126740.934
Non-
vegetation		10616291735
Sum274916604409
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

Lee, G.; Kim, G.; Min, G.; Kim, M.; Jung, S.; Hwang, J.; Cho, S. Vegetation Classification in Urban Areas by Combining UAV-Based NDVI and Thermal Infrared Image. Appl. Sci. 2023, 13, 515. https://doi.org/10.3390/app13010515

AMA Style

Lee G, Kim G, Min G, Kim M, Jung S, Hwang J, Cho S. Vegetation Classification in Urban Areas by Combining UAV-Based NDVI and Thermal Infrared Image. Applied Sciences. 2023; 13(1):515. https://doi.org/10.3390/app13010515

Chicago/Turabian Style

Lee, Geunsang, Gyeonggyu Kim, Gyeongjo Min, Minju Kim, Seunghyun Jung, Jeewook Hwang, and Sangho Cho. 2023. "Vegetation Classification in Urban Areas by Combining UAV-Based NDVI and Thermal Infrared Image" Applied Sciences 13, no. 1: 515. https://doi.org/10.3390/app13010515

APA Style

Lee, G., Kim, G., Min, G., Kim, M., Jung, S., Hwang, J., & Cho, S. (2023). Vegetation Classification in Urban Areas by Combining UAV-Based NDVI and Thermal Infrared Image. Applied Sciences, 13(1), 515. https://doi.org/10.3390/app13010515

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