Analysis of Urban Green Space Extraction Based on UAV Images

High-resolution image data of Beichuan County were obtained by unmanned aerial vehicles (UAV). After a series of pretreatments such as three-space encryption and orthophoto correction,ENVI software was used to calculate the visible-band difference vegetation index, normalized green-red difference index (NGRDI),and normalized green-blue difference index (NGBDI).The urban vegetation was extracted and analyzed by object-oriented image classification,and the accuracy was evaluated. The results showed that all the three planting cover indexes could extract the urban green space well, and the overall extraction precision was above 83%, among which the VDVI extraction effect was the best, and the overall precision was up to 89.5%.Therefore,it was feasible to extract statistics of the urban green space by UAV remote sensing technology. Based on the VDVI statistics, the classification results of the urban green space were corrected by removing small patches and visual interpretation. The area of urban green space was 2.3948 km² in Beichuan County, and the coverage rate of the urban green space was 40.04%.