Research On Video Spatial Iation Method And Application Of Urban Road Traffic Based On Video GIS

As an important part of modern urban development,urban traffic is a very complex dynamic system,and its operation efficiency plays an important role in urban economic and social development.However,with the rapid development of China's economy,private car ownership has been increasing year by year,and traffic congestion has become more and more frequent.At present,most map apps can display the road condition information,but the road congestion degree is based on the collection of GPS data of users and the data of purchased floating cars,and the expansion and estimation based on the relevant model to obtain the road condition information of road operation.This kind of information can be used to show or even predict the traffic jam situation in real time and inform the users,but it cannot accurately obtain the actual traffic volume of the road,and there are some road condition information errors caused by incomplete data collection.The road monitoring video can show the road operation in real time and reflect the road service level.In this paper,video GIS technology will be used to make better use of the spatial-temporal expression ability of surveillance video,dig the road operation information contained therein,optimize the judgment result of road traffic congestion,and provide more real,accurate and reliable road condition information.Based on the above research goal,this article take the video of GIS technology,combining with the video image identification technology,focusing on road traffic information extraction and classification,this paper studies the spatial technology and vehicle identification technology,video and apply this method to direction of road traffic information extraction,construction of information platform of urban road traffic.The main research contents and work include the following aspects :(1)research and implement a universal video spatial method.Parameters through the camera's longitude and latitude,height thereof and internal calibration of camera,in this article puts forward the vision of rectangular pyramid model to calculate the video taking a pixel corresponds to the actual space coordinates,and according to the coordinates to video projection,video taken by the camera to obtain spatial reference video given,as the data source for GIS system to provide real-time,real data;(2)a classifier for vehicle recognition is designed and implemented,and the target tracking algorithm is realized based on it.HOG feature extraction was carried out on the pre-processed sample images and sent to the SVM classifier for learning.The classifier needed in this study was obtained for vehicle recognition in the video.Combining with the target tracking algorithm,the number and trajectory of vehicles in the video can be counted effectively.(3)summarize the evaluation index of road congestion to grade the level of road congestion.After the indicators are determined,the congestion classification model is constructed and compared with the road condition function of Amap,Baidu map and other map apps for verification.(4)integrate various functional modules to build the urban road traffic information platform.The development of desktop software,in which in addition to translation,zoom,ranging and other basic GIS functions,the realization of video projection,vehicle identification,congestion classification and display functions.This study effectively solved the relationship between video data and spatial location,and applied it to the field of urban transportation.