Research on Vehicle Identification Based on High Resolution Satellite Remote Sensing Image

Abstract

Traditional traffic information acquisition methods had many limitations, so the acquisition of road traffic information from high-resolution satellite images had become a research hotspot in intelligent traffic system. In this paper, the object oriented classification method is used to establish the vehicle detection and processing process of high resolution satellite image. First, the high resolution satellite image was processed by median filtering and denoising, and the enhancement of vehicle information in remote sensing images was realized by using the stretch of gray histogram. Secondly, the image was segmented to the optimal scale, and the optimal segmentation scale was determined by the method of the maximum average area of the vehicle. Then according to different color characteristics of vehicles with dark establish a separate classification rule set, based on the object-oriented classification of vehicle classification, by extracting feature threshold classification more bright color vehicles, using relationship between classes of vehicles using fuzzy classification method to extract the dark color vehicles, finally formed a high-resolution satellite remote sensing overall framework of vehicle detection.