An Image-Processing Approach for Computing the Size of Green Areas in Cities

The delimitation of green areas from urban ones is a fundamental operation for a correct estimate of green spaces within cities. Specific GIS applications assist users in the tracing of boundaries and the calculation of lengths and areas. However, when using such techniques, the main limitation concerns the analysis of irregularly shaped areas. For them, tracing outlines that overlap the original ones is performed manually, hence difficult, error prone and time consuming. This paper proposes a novel approach that is automatically performed by means of a Python-based application to recognize boundaries and compute the main geometric parameters of urban and green areas. Such operations are based on new algorithms that give an innovative character to the entire application. The application was tested in the territory of Kamakura, a Japanese city that embodies the “City Country Fingers” design pattern characterized by extensive intersections between the vegetation, coming from the peripheral areas, and the urban center. Their complexity and irregularity are a proper testbed to evaluate the behavior of the application in the presence of irregular areas. The obtained results confirmed the reliability of the proposed approach and the efficiency of the application.