Spatio-temporal monitoring of urban street-side vegetation greenery using Baidu Street View images

Street-side vegetation greenery contributes substantial health benefits for pedestrians. Multi-year street view images are expected to enable the monitoring of dynamic street-side vegetation greenery changes and the development of targeted urban landscape plans. However, the potential of multi-year street view images used for the assessment of street-side vegetation greenery has not been evaluated yet. Besides, complicated urban landscapes may make it difficult to accurately quantify vegetation greenery. This study developed a framework to assess the spatio-temporal variation of street-side vegetation greenery using the Baidu Street View images and a new Vegetation Greenery Index (VGI). The proposed analytical framework was applied to Tai'an city, a highly populated city where urbanization has been rapid in China. The level of vegetation greenery estimated using the proposed framework was compared with ground truths randomly collected at sampling sites along the road networks in 2014 and 2019 to assess the applicability. Results demonstrated that the proposed VGI method could accurately quantify street-side vegetation greenery. The comparison of multi-year VGI layers could identify locations where vegetation greenery substantially changed and quantify the overall change in urban greenery. Vegetation greenery estimates were well agreed with the ground truths. Spatio-temporal variations in the urban vegetation greenery were attributed to trees that were newly planted or removed, the natural growth of the existing vegetation, and new building construction. The proposed framework is expected to be a useful tool to evaluate urban vegetation greenery and help urban landscape planning.