2020.08.7 Friday

Standardized Green View Index: Measuring and mapping urban greenery by the machine’s eye

Standardized Green View Index

Urban green vegetation or canopy provide shades on the street, mitigation of heat island effect, psychological comfort and so on. However, the presence of the trees has not been well captured. This study employs the Green View Index, an indicator of visibility of greenery at street level, and proposes an area-level estimation of the visibility of green vegetation. First, calculation sites of Green View Index (GVI) are positioned along streets. Second, Google Street View images on these sites are retrieved. Third, the proportion of green pixels is estimated. Forth, hereby estimated GVI is aggregated at area-level, using Voronoi tessellation in order to consider the network structure in the area.

1. Define sites along streets

Position sites for image retrieval at 100m interval along streets in the study area (the minimal interval is 20m).

2. Retrieve images from Google Street View

For each site, retrieve Google Street View images (in six different angles covering 360˚ panoramic view) via Google Street View API. It is possible to designate year and month when the image is taken, if such metadata is available the GSV database.

3. Calculate the proportion of green pixels

Green View Index (Li et al. 2015) is calculated from the GSV images. Green View Index here is the proportion of green pixels over all pixels. Average value of the six images at one site is the Green View Index of the site.

Figure 4.15. from Li (2016)

4. sGVI: Voronoi tessellation

If you want to aggregate the Green View Index at zone level (e.g. city block, administrative boundary, etc.), calculate standardized GVI using Voronoi tessellation in the given zone. This mitigates the bias from densely located sites, and provides fairer estimation of visibility of greenery.

Comparison with other metrics

Another conventional method of measuring green vegetation is Normalized Difference Vegetation Index (NDVI), which is calculated from satellite images (see Figure 1).

Figure 1. NDVI in Nishi-ku and Kanagawa-ku in Yokohama city

Since NDVI captures the green vegetation from satellites (top-down), the estimated value does not necessarily correspond to what people perceive on the street. In order to compare the characteristics of sGVI and NDVI, geographical distributions of the indices are illustrated in Figure 2.

Figure 2. Comparison of sGVI and NDVI in Yokohama city (Nishi-ku and Kanagawa-ku)

It turned out that sGVI is more sensitive to green vegetation in well-developed area (e.g. city center), rather than in suburban area with abundant green space (e.g. parks, forests).

How to make use of this tool?

This tool enables us to measure green visibility, especially in urbanized zones, in an objective manner. Also, by combining sGVI estimation with other factors (e.g. public health, socio economic, etc.), association studies can be performed in order to observe relation between green vegetation and human beings. The works in this orientation will contribute to enhance the Quality of Life, through strategically creating green spaces.


Codes are available here


Kumakoshi, Y., Chan, S. Y., Koizumi, H., Li, X., & Yoshimura, Y. (2020). Standardized Green View Index and Quantification of Different Metrics of Urban Green Vegetation. arXiv preprint arXiv:2008.00229.


Li, X., Zhang, C., Li, W., Ricard, R., Meng, Q., & Zhang, W. (2015). Assessing street-level urban greenery using Google Street View and a modified green view index. Urban Forestry & Urban Greening, 14(3), 675-685.

Li, Xiaojiang. (2016). Investigating Environmental Inequities in Terms of Street Greenery using Google Street View. Doctoral Dissertations. 1307.

▲ Back to Top