The Target Recognition Technology For High Resolution Satellite Image

Since the twenty-first century,along with China's rapid economic development,the increasing purchasing power of the people,the increasing demand for traffic travel,rising car purchases,lead to more and more serious road congestion situation.In order to mitigate this situation,it is important to provide accurate road vehicle information for the transportation planning and monitoring department.With the launch of China's NO.2 High Resolution Satellite,China's high-resolution satellite for road vehicle information extraction provides an important means.In this paper,the realization and verification of various methods of vehicle extraction on NO.2 High Resolution Satellite image are taken as the main research contents,and the extraction of congested sections is realized on the basis of this.By analyzing the spectral characteristics,spatial features and texture features of the vehicle on the remote sensing image,it provides the basis for the vehicle extraction,analyzes and studies the threshold segmentation method,the gradient comparison method,the background difference method,the object-oriented method and the vehicle extraction method And the verification of morphological processing,edge detection,image segmentation and image enhancement are also studied.In addition,the object extraction function of ENVI FX module is introduced.On the basis of vehicle extraction,12 * 12 mobile window method is designed to identify and locate the position of vehicle on the road,and the method of Hough transform is used to estimate the length of car section.A method of extracting congested road is designed.High Resolution Satellite Image in the extraction of road vehicles has a more extensive application,as China's transportation management departments to provide more scientific and effective means in emergency disasters and major accidents monitoring has a more far-reaching significance,lay a good foundation for the future research work.