Vegetation ecological benefits index (VEBI): a 3D spatial model for evaluating the ecological benefits of vegetation

ABSTRACT Urban population explosion may increase ecological environment discomfort, thereby affecting negatively humans’ mental and physical performance. Therefore, it is important to detect and monitor vegetation and predict its ecological benefits. The complex composition of urban environment ground objects, such as steel roofs, plastic courts, and building shadows, significantly interferes with vegetation detection and monitoring. The optimized hyperspectral image-based vegetation index (OHSVI) constructed in this study effectively solves this problem. However, it is difficult to accurately predict the ecological benefits of vegetation based on the two-dimensional vegetation information extracted based on remote sensing images; this is related to the three-dimensional (3D) structure of vegetation and the 3D pattern of buildings. Therefore, we first proposed the vegetation ecological benefits index (VEBI) based on the 3D structure of vegetation to reveal how vegetation acts on its 3D surroundings. The method was tested in a playground, an academic building, and a parking space. The results showed that the vegetation extraction accuracy of the OHSVI exceeded 93%, which is better than that of the existing indices. Our findings suggest that VEBI may be efficient in predicting 3D vegetation ecological benefits combined with remote sensing and lidar datasets.


Introduction
In recent years, the world has experienced rapid urbanization, with the proportion of the global population rising from 10% in 1900-68.4% in 2050 (Grimm et al. 2008;Li et al. 2018;Bai, Chen, and Shi 2012;Zhou et al. 2018).With the rapid increase in the urban population, environmental problems brought about by urbanization, such as urban heat islands, urban heat waves, and chronic respiratory diseases, have received increasing attention (Liu et al. 2020;Lai et al. 2020;Liang, Wang, and Li 2019;Zhang, Fan, and Jiao 2023).As a significant component of the urban environment, vegetation generates multiple ecological benefits of cooling, humidifying, and purifying the air, thus improving the urban ecological environment (Dimoudi and Nikolopoulou 2003;Richards et al. 2020;Jiao et al. 2021).
A theme that has become increasingly distinct in recent years is urban vegetation and how it is perceived, valued, and engaged with, particularly in urbanized societies (Li et al. 2021;Wu et al. 2021;Chen, Jin, and Du 2020).Interventions prompted by urban vegetation, such as urban parks, have long remained in place, presumably because the public has enjoyed them and believed in their salutary values (Hartig et al. 2014;Liu et al. 2019;Wang et al. 2020).The accurate quantification of the ecological benefits of urban vegetation tests such beliefs and encourages the adoption of a more nuanced theoretical and practical consideration of nature-health relations (Peng et al. 2021;Wang et al. 2019;Van de Voorde 2017).
Currently, research on the ecological benefits of urban vegetation has mainly focused on the twodimensional (2D) spatial scale.Using GIS technology, Qiu et al. (Qiu et al. 2020) investigated the responses of vegetation to urbanization and climate change.Zhang et al. (Zhang et al. 2017) observed the microclimate effects of different types of urban vegetation.Bagheri et al. estimated the microclimate effect in different urban vegetation structures by calculating the amount of green growth based on the theory of plant transpiration (Bagheri et al. 2021).However, research on natural ecosystems shows that the plant community, as an open biological system, can influence the three-dimensional (3D) space environment around urban vegetation through system exchange (Chi et al. 2020).Urban vegetation ecosystems are typical artificial ecological communities, and their ecological processes are much more complex than those of natural ecosystems.The existing 2D urban vegetation ecological benefit evaluation theory and system cannot cope with the influence of the 3D properties of vegetation.
To explore the ecological benefits of vegetation more accurately, scholars have gradually turned to the 3D space scale.Schoepfer et al. combined the relative proportion of multistory houses and the distances between vegetation to propose a green index (GI), which linked human viewpoints with ecological aspects of urban green (Schoepfer, Lang, and Blaschke 2005).Although GI considers the 3D spatial structure, vegetation research remains on the 2D plane.Liu et al. used LiDAR data and multispectral remotely sensed imagery with the information of urban buildings and vegetation to propose the building's proximity to green spaces index (BPGI), which can evaluate the proximity of residents to green spaces at the building level (Liu et al. 2016).BPGI considering 2D vegetation only evaluates the impact of vegetation from a distance, which is not enough to describe the ecological benefits of vegetation.Yu discovered that 3D landscape pattern metrics could better describe the undulation and heterogeneity of urban surfaces and were essential when explaining the variation in LST compared with conventional 2D landscape pattern metrics (Yu et al. 2020).However, the ecological benefits of vegetation are not only controlled by the 3D structure of vegetation but are also inevitably affected by the surrounding 3D architectural landscape (Rebane et al. 2020).The relationship between vegetation and the surrounding 3D architectural landscape is an obstacle that must be broken through the accurate characterization of the ecological benefits of vegetation.
Moreover, the composition of the city is complex, especially the emergence of new materials such as steel roofs and plastic courts, which bring great challenges to the identification of urban vegetation (Sun et al. 2021).Simultaneously, with the increasing number of high-rise buildings in cities, the shadows generated by them gradually cover many vegetation types, resulting in the loss of vegetation information (Jiao et al. 2021;Jiang et al. 2019).In addition, as withered vegetation no longer exchanges energy with the external environment, it no longer produces ecological benefits, so this should also be considered when identifying vegetation.The accurate identification of vegetation from complex urban environments is important for the quantification its ecological benefits.
The aim of this study was to calculate the ecological benefits of urban vegetation.To this end, we first solved the problem of vegetation identification.After an in-depth analysis of the spectral features of steel roofs, plastic courts, shadows, and withered vegetation, an optimized hyperspectral image-based vegetation index (OHSVI) was constructed to effectively suppress the influence of interference factors.Then, we combined the OHSVI and 3D properties of vegetation to present a three-dimensional vegetation ecological benefits radiation model (3D-VEBRM), which breaks through the obstacle between vegetation and the surrounding 3D architectural landscape.To accurately quantify the ecological benefits of vegetation, a vegetation ecological benefits index (VEBI) was proposed based on the 3D-VEBRM; this allowed for the calculation of the ecological benefit of vegetation acceptable to humans in a 3D hemispherical space.The VEBI is more in line with the people-oriented concept and can visualize intangible ecological benefits; it is more in line with the 3D spatial perception of the ecological benefits of vegetation.The VEBI has successfully provided new evaluation indicators and directions suited to the spatial layout of urban vegetation and urban green construction.
The remainder of this paper is organized as follows.Section 2 introduces the data used in the study.The proposed method is described in Section 3. The experimental results and analysis as well as the discussion of VEBI are presented in Sections 4 and 5, respectively.Finally, concluding remarks are presented in Section 6.

Study data
We selected hyperspectral image (HSI) datasets and lidar point cloud data to test the experimental results (Figure 1).To highlight the ecological benefits of vegetation, three scenarios were selected to test the effectiveness of the algorithm.As shown in Figure 1(a), the playground includes a lawn with a relatively high density of vegetation, which is one of the scenarios.As a large open space with little vegetation cover, the parking lot scene has a very different vegetation structure from that of the park and academic buildings, as shown in Figure 1(b).Figure 1(c) shows that many wooded parks are located near academic buildings, in contrast to the playground.
The HSI datasets were acquired by the National Center for Airborne Laser Mapping (NCALM) over the University of Houston campus and its neighborhood between 16:31 and 18:18 GMT on February 16, 2017.The sensor used in this dataset was an airborne ITRES CASI 1500 (a hyperspectral imager).The image size was 594 × 599 pixels and the geometric resolution was 1 m per pixel.This dataset covers a 380-1,050 nm spectral range with 48 bands and includes complex urban surface features such as shadows and synthetic courts.This can be used to test the efficacy of vegetation indices in extracting vegetation from complex environments.
We also selected lidar point cloud data from the same location.The sensors used in this campaign included an Optech Titan MW (14SEN/CON340) with an integrated camera (a multispectral LIDAR sensor operating at three different laser wavelengths).We included multispectral LIDAR point cloud data at 1,550 nm, 1,064, and 532 nm as well as intensity rasters from the first return per channel and DSMs at a 50-cm GSD.From this dataset, we obtained the 3D structure of an urban surface.

Data preprocessing
Lidar point cloud data must be smoothed by mean filtering (Lim 1989) and then input into the model.This process can ensure good collaboration between lidar point cloud data and HSI.The calculation method is as follows: where (x,y) represents the spatial position of the pixel, K represents the kernel of the mean filter, I represents the lidar point cloud data, and H represents the final digital surface model (DSM).

Materials and methods
In this section, we focus on the construction methods of the OHSVI and 3D-VEBRM.On this basis, a quantitative expression of the VEBI is provided.

The construction of OHSVI
To effectively extract vegetation, a suitable method must be selected based on the analysis of the spectral characteristics of vegetation and non-vegetation areas.To accomplish this objective, we selected 1,000 sample points for each land cover type.These samples were separated into two categories: vegetation and non-vegetation objects.Specifically, vegetation included vegetation under   All these factors have a serious impact on vegetation extraction.To overcome these problems caused by shadows, we calculated their spectral curves, as shown in Figure 3.
The vegetation index cannot only effectively extract vegetation, but also characterize the growth state of vegetation; the normalized difference vegetation index (NDVI) is the most commonly used (Schnell 1974).However, as shown in Figure 3, the near-infrared band (800 nm) and red band (680 nm) used for NDVI are no longer suitable for vegetation extraction.Plastic courts and steel roofs have spectral characteristics similar to withered vegetation and shaded vegetation at 800 and 680 nm, respectively, which makes it easy for NDVI to identify plastic courts and steel roofs as vegetation.Herein, we further analyzed the spectral curves shown in Figure 3 and found that at 760 nm and 689 nm, the spectral characteristics of vegetation are obvious, with obvious differences from those of plastic courts and steel roofs.
In addition, we found that the index structure of the normalized difference can lead to the saturation of the NDVI, and we deformed the NDVI into Eq.3.
where ρ 800 and ρ 680 are the spectral bands of 800 and 680 nm, respectively.Figure 3 shows that the value of the vegetation reflectance is significantly different at 800 and 680 nm, which causes the structure of r 680 r 800 in Equation 3 to tend to 0. With the gradual increase in vegetation density, the structure of r 680 r 800 always maintains a very small value close to 0, which leads to NDVI always being in a saturated state, and the value remains close to 1.In summary, we consider 760 and 689 nm as the main spectral characteristic bands, and simultaneously change the index structure of the normalized difference by introducing a blue band (474 nm) that can weaken the influence of the atmosphere (Kaufman and Tanré 1992)

The formation of 3D-VEBRM
According to ecological field theory, there is a close interaction between the artificial landscape vegetation ecosystem and the outside world (Wu et al. 1985).The ecological benefits of vegetation, such as cooling, humidifying, and improving air quality, radiate to surrounding areas (Richards et al. 2020).Considering the height, area, and growth state of vegetation, a 3D model of vegetation ecological benefit attenuation was proposed, as shown in Figure 4.
In Figure 4, the red area is the outside world, where the ecological benefits of vegetation cannot be radiated.The blue area represents the range that can be affected by the ecological benefits of vegetation.The green area represents the artificial landscape vegetation ecosystem.The aggregation characteristics of vegetation lead to a cumulative efficacy of ecological benefits in the vertical direction.Therefore, to better describe the phenomenon of the 'ecological benefit island' in the vertical direction, we constructed a vertical extension height (hrv) to characterize it.This is specifically expressed as the maximum vertical extension height of the ecological benefits produced by plants owing to the aggregation mode (vertical direction): where OHSVI represents the growth state of vegetation, hrv represents the vertical extension height of vegetation ecological benefits, and rm represents the maximum radius from the center of the vegetation-gathering area to the edge.
The ecological benefit radiation of vegetation gradually changes in the horizontal direction; therefore, the radiation intensity is different in space.With a gradual increase in the distance from the vegetation, the ecological benefit radiation in the horizontal direction tends to attenuate.The attenuation intensity (a) of the vegetation ecological benefits was determined by considering the height and location of the vegetation.Its specific meaning is the attenuation of ecological benefits produced by plants per 1 m in the horizontal direction: where a represents the maximum radiation radius of the vegetation ecological benefits and hv represents the height of the vegetation.

The construction of VEBI
The ecological benefits of vegetation include a certain impact range, but this is not the mean value.The index should be able to monitor dynamic spatial changes in ecological benefits.Based on 3D-VEBAM, we obtained the distribution of ecological benefits around the spatial vegetation, as shown in Figure 5.
As shown in Figure 5, the observer height was set to 1.70 m.The distribution of objects is complex in the observer's 3D perception space.Owing to the obstruction of the building to the west of the observer, the ecological benefits generated by the vegetation on the south side of the building are blocked and cannot be extended to the location of the observer.As the observer is blocked by the building to the west, the ecological benefits generated by the vegetation on the south side of the building are blocked and cannot be extended to the observer's location.Although there is a large amount of vegetation in the southern part of the observer, the observer can only perceive the ecological benefits of vegetation in the circular space shown in Figure 5, which is due to the limited range of perception.
Owing to the differences in the spatial distribution of artificial buildings, the distribution of ecological benefits that people can feel in space is different.To quantify the spatial distribution of vegetation ecological benefits from a human perspective, a VEBI was proposed based on 3D-VEBRM to break through obstacles between vegetation and the 3D surrounding landscape; its schematic diagram is shown in Figure 6.
As shown in Figure 6, we used a single plant as an example to introduce the method of quantifying the vegetation ecological benefits in detail.The 3D-VEBRM parameters can quantify the spatial distribution of the ecological benefits of a single tree.Then, combined with terrain factors, spatial distribution of ground features, and human spatial perception, the formula of the VEBI is given as follows: where a represents the maximum radiation radius of vegetation ecological benefits, d represents the horizontal distance between the vegetation and the observer point, DDEM represents the DEM difference between the observer point and vegetation, ht represents the height of the observer (1.70 m), and i represents the i-th vegetation pixel that is not blocked by buildings from the observer.

Vegetation extraction results
The experiments in this section mainly verified the vegetation extraction results of the HSVI.The overall accuracy (OA) was used to evaluate the accuracy of the Vis (Liu, Fang, and Li 2011), as shown below: where TP is the number of true vegetation pixels classified as vegetation pixels, FP is the number of true vegetation pixels classified as non-vegetation pixels, TN is the number of non-vegetation pixels classified as non-vegetation pixels, FN is the number of non-vegetation pixels classified as vegetation pixels, and TV is the total number of true vegetation pixels in the reference maps.Before extracting the vegetation, we used iterative calculations to determine the optimal threshold for each index to achieve binary classification.In this study, three experimental areas were selected, and each experimental area was set up with the interference samples shown in Figure 2. The vegetation extraction results for each index are presented in Table 1.
Based on the results in Table 1, the classification accuracy of OHSVI is better than that of the other algorithms in all cases, and the index stability is greater than that of the other indices.This shows that OHSVI exhibits excellent performance in accurately extracting vegetation, and its extraction solution is sufficient to meet the requirements of this study.

Vegetation ecological benefits
The experiments in this section mainly verify the advantages of the VEBI.We studied three separate patterns to verify the ecological benefits of the vegetation.High-resolution RGB imagery is shown in Figure 7(a)-(c).Figure 7(d)-(f) show the NDVI of the playground, academic building, and parking lot in the research area.The ecological benefits of the vegetation in the playground, academic building, and parking lot are depicted in Figure 7(d)-(f).The ecological parameters of urban vegetation are summarized in Table 2.
The NDVI in Figure 7(d)-(f) indicates that the parking lot includes a lot of vegetation, and the academic building and playground include the same area of vegetation.Although the NDVI results are consistent with the statistical results of the vegetation area and green biomass in Table 2, there are some discrepancies.In fact, although the parking lot has the most vegetation coverage, its green biomass is not much different from that of the academic building.This is very different from the results shown in Figure 7(e) and (f), which indicate that the NDVI vegetation statistics process will cause a large deviation, and the results are not accurate.In addition, the quantification results of the vegetation ecological benefits in Figure 7(g)-(i) show that the spatial radiation area of the ecological benefits of vegetation in the parking lot is not optimal and there is a gap compared to the academic building.This conclusion can be confirmed from the total ecological benefit radiation areas in Table 2. Academic buildings have far more vegetation areas than parking lots and include more forest gardens than parking lots.This is the main reason the ecological benefits of academic buildings radiate widely.This demonstrates that the composition and area of vegetation have an important influence on the ecological benefits of vegetation in the same area.More interestingly, as shown in Table 2, the playground is significantly better than the academic building and parking lot based on the ecological benefit radiation area produced per unit vegetation area.This shows that the ecological benefits of large lawns are superior to those of trees, which was unexpected.However, a careful analysis of Figure 7(g) and (h) shows that the ecological benefits of trees are far superior to those of lawns.In addition, we found that in the playground, the vegetation ecological benefits of each square meter were the highest, because the ecological benefits of vegetation were not blocked owing to the relatively broad environment.However, as shown in Figure 7(h) and (i), the academic building and parking lot include a large number of high-rise buildings surrounding the vegetation, resulting in ecological benefits being limited to a small space.This reveals that the spatial distribution of the ecological benefits of vegetation has a direct impact on the surrounding 3D landscape.In summary, the VEBI can consider the 3D structure of vegetation and the 3D spatial distribution of surrounding buildings to achieve a more accurate spatial quantification of vegetation ecological benefits.

Three-dimensional landscape structure
To more clearly analyze the impact of the 3D landscape structure on the ecological benefits of vegetation, we show the detailed landscape structure in Figure 8.  Owing to the difference in the vegetation density in the sports field, the ecological benefits brought by the vegetation inside the stadium are consistent with the same degree of differences.Figure 8(e) shows the ecological benefits of the island formed by trees better than grass.Additionally, different tree densities incur different ecological benefits.Although the intensity of ecological benefits brought about by dense forests is relatively high compared with the ecological benefits brought about by sparse forests, the amount of radiation and area are far less.As shown in Figure 8(f), although the ecological benefits produced by trees are far greater than those of grasslands owing to the landscape layout, the ecological benefits are difficult to radiate outward, and their actual contribution to ecological benefits is much smaller than that of grasslands.Under the same degree of vegetation dispersion, the 3D building pattern has the most serious impact on the ecological benefits of vegetation.In conclusion, VEBI can comprehensively consider the three-dimensional landscape pattern and the 3D structure information of vegetation to evaluate the ecological benefits of vegetation.
Additionally, the quantification of the ecological benefits of vegetation based on VEBI has obvious advantages in terms of unique features.We further compared the vegetation ecological benefits of VEBI and NDVI in three areas: a plastic court, withered grasslands, and building shadows, as shown in Figure 9.
In the synthetic turf stadium shown in Figure 9(a), NDVI shows strong vegetation features in the elliptical region of Figure 9(d).However, the synthetic turf stadium does not produce the slightest ecological benefit, which is consistent with the VEBI results shown in Figure 9(g).Similarly, as shown in Figure 9(b), there is a large number of withered grasslands in the rectangular area, and its ecological benefits are almost zero.However, the rectangular area in Figure 9(e) shows that the vegetation characteristics are clear and include strong ecological benefits.This is unreasonable and not true.The VEBI result in Figure 9(h) is different from the NDVI result, which effectively suppresses the impact of withered grassland and accurately quantifies the ecological benefits of vegetation in this region.Based on this conclusion, it is difficult to distinguish the ecological benefits of different disturbance factors, which is directly related to its simple modeling, resulting in insufficient comprehensive information.
In addition, building shadows are even more puzzling.As shown by the ellipse in Figure 9(c), there were impervious surfaces and withered grass under the shadow of the building as well as lower temperature.However, the analysis showed that the ecological benefit of cooling is not caused by vegetation, but by the blocking of solar radiation by tall buildings.In general, NDVI often indicates that an ecological benefit is derived from vegetation.However, this conclusion was rejected because Figure 9(f) confirms that the withered grassland has little ecological benefit.Therefore, the VEBI results in Figure 9(i) are more in line with the actual situation because they quantify the ecological benefits from the main body of vegetation.Effectively suppressing the interference caused by shadows is the advantage of VEBI, which is closely related to its comprehensive 3D modeling.
Under different landscape patterns, the VEBI can comprehensively consider building information in 3D space and effectively quantify the ecological benefits of vegetation with different structures.At the same time, the VEBI can focus more on the quantification of the ecological benefits of vegetation in the face of interference factors such as shadows and plastic courts.The VEBI possesses a unique perspective on the quantification of vegetation ecological benefits in 3D space, which is not available or considered in existing methods.

Discussion
In traditional urban vegetation research, the research content has gradually shifted from 2D vegetation areas to 3D urban green biomass.Quantitative research on the quantity and quality of vegetation in cities is also increasing.Therefore, the quantity and quality of vegetation have also become important indicators of urban ecological assessment.However, these conclusions are one-sided.
By proposing the VEBI, the influence of the 3D landscape was considered in the modeling for the first time.Owing to the quantitative problem of 3D landscapes, their direct application in urban vegetation research has been challenged.This challenge was addressed for the first time in the present study.Through this research, we found that the spatial radiation of vegetation ecological benefits is not only related to vegetation but also to the architectural landscape pattern of the area where the vegetation is located.In Figure 7(a) and (b), we can see that these two areas include almost the same vegetation area, but there is a large gap in the green biomass, which leads to a significant increase in the final vegetation ecological benefit radiation area.In addition, as shown in Figure 8(e) and (f), although the vegetation area and green biomass of academic buildings are less than those of parking lots, the radiated area of vegetation ecological benefits in academic buildings is larger than that of parking lots.This indicates that the architectural pattern around vegetation has a significant effect on the spatial impact of the ecological benefits of vegetation.The proposed VEBI can effectively integrate three-dimensional vegetation green quantity information and three-dimensional landscape pattern information around the vegetation to quantify the spatial radiation of the ecological benefits of the vegetation.This is more comprehensive than the traditional urban ecological construction evaluation based on a two-dimensional area of vegetation and three-dimensional green biomass.
In addition, as the original data are difficult to collect, VEBI cannot be applied on a large scale.This direction will be addressed in future research.Fortunately, the calculation of vegetation ecological benefits based on the block scale can meet the needs of urban ecological research, particularly in urban landscape planning.The use of the VEBI to calculate the spatial distribution of vegetation ecological benefits in real time can help urban construction provide optimal solutions, which is beyond human subjective cognition.

Conclusions
In this study, we demonstrated that a 3D-VEBRM-based VEBI can be developed to model vegetation ecological benefits using remotely sensed and lidar datasets.The novelty of this study is threefold: (a) the presentation of a 3D-VEBRM for modeling vegetation benefits using ecological field theory and remote sensing data; (b) the presented VEBI shows the ability to model the vegetation benefit status in both absolute and relative terms; (c) the challenge of an unclear relationship between landscape and urban ecology is solved by applying VEBI.The emergence of the VEBI has provided a new evaluation index and construction advice suited to urban green construction projects.This can describe the effectiveness of urban green construction more comprehensively.
Future studies should explore the feasibility of modeling night-time vegetation ecological benefits, because vegetation activities can be more intricate in the evening.VEBI can conduct refined research in combination with temperature, humidity, and other indicators and describe the impact of vegetation on ecological indicators in the evening.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Figure 1 .
Figure 1.Overview of the study area from Houston University (very high resolution RGB imagery): (a) playground, (b) parking lot, and (c) academic building.
the shadow (Figure 2(a)), and withered vegetation (Figure 2(b)), while non-vegetation objects included the plastic court (Figure 2(c)), impervious surfaces under shadows (Figure 2(d)), and steel roofs (Figure 2(e)).As shown in Figure 2, during vegetation extraction, interference factors, such as shadows (Figure 2(a) and (d)) and plastic courts (Figure 2(c)), often exist.In addition, we found that steel roofs were often misclassified as vegetation, as shown in Figure 2(e).As withered vegetation (Figure 2(b)) no longer possesses ecological benefit radiation, it is regarded as non-vegetation and needs to be culled.

Figure 3 .
Figure 3. Spectral curves derived from land cover types: vegetation, withered vegetation, shaded vegetation, plastic court, steel roofs, and shaded impervious surfaces.

Figure 4 .
Figure 4.The three-dimensional vegetation ecological benefits radiation model.

Figure 5 .
Figure 5. Landscape distribution in the 3D space of the observer.

Figure 6 .
Figure 6.The schematic diagram of the vegetation ecological benefits index.

Figure 8
Figure8demonstrates the advantages of VEBI in detail.The island phenomenon of vegetation ecological benefits is shown in Figure8(d).Owing to the difference in the vegetation density in the sports field, the ecological benefits brought by the vegetation inside the stadium are consistent with the same degree of differences.Figure8(e) shows the ecological benefits of the island formed by trees better than grass.Additionally, different tree densities incur different ecological benefits.Although the intensity of ecological benefits brought about by dense forests is relatively high compared with the ecological benefits brought about by sparse forests, the amount of radiation and area are far less.As shown in Figure8(f), although the ecological benefits produced by trees are far greater than those of grasslands owing to the landscape layout, the ecological benefits are difficult to radiate outward, and their actual contribution to ecological benefits is much smaller than that of grasslands.Under the same degree of vegetation dispersion, the 3D building pattern has the most serious impact on the ecological benefits of vegetation.In conclusion, VEBI can comprehensively consider the three-dimensional landscape pattern and the 3D structure information of vegetation to evaluate the ecological benefits of vegetation.Additionally, the quantification of the ecological benefits of vegetation based on VEBI has obvious advantages in terms of unique features.We further compared the vegetation ecological benefits of VEBI and NDVI in three areas: a plastic court, withered grasslands, and building shadows, as shown in Figure9.In the synthetic turf stadium shown in Figure9(a), NDVI shows strong vegetation features in the elliptical region of Figure9(d).However, the synthetic turf stadium does not produce the slightest

Table 1 .
The vegetation extraction accuracy for each vegetation index.

Table 2 .
Statistical results of urban vegetation.