Research On Road Extraction And Traffic Analysis Based On High-resolution Remote Sensing Images

With the development of remote sensing technology in China,the information extraction technology of ground object information for remote sensing is becoming more and more extensive.It’s strong demand to apply remote sensing for public security works,which is practical significance in anti-terrorism,emergency as well as traffic monitoring and management.The high resolution remote sensing images data is used to extract road information where terrorist threaten frequently in Xinjiang municipality,aiming at providing comprehensive,complete and accurate distribution information of road network for the public security department,which is convenient for police arrangement and police route planning.In order to optimize the selection of police route planning,this paper focuses on the study of road conditions,and uses the vehicle information extracted from the road in remote sensing images to develop spatial analysis and divide the road congestion level by vehicle density.This level can be given in the planning algorithm of the weight based path and provide the optimum path choice support.In addition,the results of road analysis can also be applied to traffic monitoring,providing analysis and solution for the existence congestion,illegal parking and other causes in large cities.Based on this purpose,this paper first carried out on the high-resolution remote sensing image of the road and vehicle information extraction research.After preprocessing for the original image,the image is enhanced,binarized,segmented and merged according to the spectral and geometric characteristics of the target.Finally,the road and Vehicle information are extracted based on the rule-based classifying to control the distribution and update of the city,township,mountain and other regional road network.Then,the vehicle information extraction results are converted into point information for spatial analysis,achieving the classification of congestion and master the traffic conditions.All the information would be used to provide scientific choice of emergency routes and keep the safety of urban traffic.In the experiment,the roads are divided into three levels: county road,rural road and simple road.The extraction accuracy can reach more than 80% for the first two kinds of road;due to the geometric characteristics of complex and difficult to complete the complete extraction,the extraction accuracy of simple road is about 60%,requiring manual intervention to optimize.The results of the road analysis are significant that the recognition precision of the vehicle extraction is higher than 80%.Second,the spatial analysis of vehicle information extraction can effectively obtain the vehicle density in the specific road area and used it to quantify traffic congestion level,making a more intuitive and accurate monitoring.In this paper,we achieve the road and vehicle information extraction based on high score remote sensing image,monitoring road condition successfully through spatial analysis.The research can not only provide basis for the choice of emergency route,but also apply to the general city Traffic monitoring and management,also has positive effect for the application of remote sensing in the field of public security.