QUANTITATIVE VEGETATION MAPPING OF URBAN AREA USING HIGH-RESOLUTION MULTISPECTRAL SATELLITE IMAGERY

Abstract

Maintenance of green spaces within the city is very important under current ecological conditions. In the present context, assessment of green spaces held by visually determining indicators plantations clarifying its instrumental measurements. The accuracy of this method strongly depends on the qualifications and professional experience of the employee conducting an assessment. High spatial and temporal resolution satellite systems provide information for operational monitoring of environmental conditions in the city and promote management decision making. Normalized Difference Vegetation Index (NDVI) – index of plant “greenness” or photosynthetic activity – is widely used for vegetation studying. However, a more important phytometric parameter of vegetation is Leaf Area Index (LAI). Actually LAI is defined as the one sided green leaves area per unit of ground area in broadleaf canopies, or as the projected needle leaves area per unit of ground area in needle canopies. Multispectral imagery processing to evaluate the NDVI and LAI provides informative and reliable indicators of vegetation quantity. High-resolution image from Pléiades-1A satellite was used in the study. Complete geoinformation technology for integrated quantitative assessment of vegetation in urbanized territory provides a distribution of LAI within the study area. The first group of operations was used for preprocessing of multispectral satellite image. Output of these operations is a scene surface reflectance. The second group of operations comprises the steps of LAI calculations based on vegetation mask and LAI versus NDVI regression dependence. Final group of operations provides visualization and interpretation of the results. The methodology for calculating the vegetation area within the land registry boundaries gives biased values of green mass and its status. At the beginning of the growing season the regulatory area of green zone is much greater than the actual amount of vegetation. It is possible to determine more accurately the boundaries and the total area of vegetation cover using the NDVI. Also the described LAI-based geoinformation technology provides tuning out the current phase of the phenological cycle. As a result, more accurate estimation of vegetation amount at moment and its changes over time are available too. Further studies aimed at assessing not only the quantity but also the quality of vegetation. The increase in number of ground control points and number of in-situ measurements are planned to clarify the regression confidence. There will also be studying the state of plant communities at different phases of the phenological cycle. High-resolution image from Pléiades-1A satellite was kindly provided by TVIS-INFO LLC (TVIS) Company, which is one of the leading geospatial and GIS solutions provider in Ukraine.