Research Of Camera Self-Calibration Detection Technology Based On Highway Surveillance Video

With the development of world economy,the application of highway is in rapid rise.So far the highway in China mileage has been over thirty thousand kilometers,ranking second in the world.But the trouble with it is that cars on the highway are highly susceptible to bad weather such as rain,snow,wind and frost.Especially being faced with mass of fog or dust storms,emergency serial collisions are more likely to happen,which leads to huge casualties and economic losses.So in the situation of the rapid development of the highway,detecting the highway on real-time is necessary to reduce traffic accidents and improve travel security.To real-time detect highway atmospheric visibility quickly,camera calibration is made some changes,where no longer multiple images,based on the principle of multiple view geometry,are adopted to calibrate camera.In this paper,vanishing point of single image to calibrate camera is investigated.First,the frame difference method is used to extract the background of the video capture and get rid of running vehicles on the lane mark line.Then the image is weighted average and binaries to get the gray image,which reduce the amount of image information.After this,it’s found that the image possibly appears holes or pseudo target points.According to the morphological image processing technique,it can effectively reduce the image impurities;At last,the edge of the image is extracted by Canny algorithm,and the two points of the intersection of the image’s parallel mark are found.The geometric model of the three coordinate system,based on the system of world coordinate,camera coordinate and image plane coordinate is established.Camera is calibrated by means of matrix,formula derivation and other mathematical operations,and using known condition of a single image pixel coordinate information and highway lane marking standard.After calibration,the camera parameters error is very small,which can be used as the camera self calibration method and provide the premise of real-time accurate detection of visibility.The entire calibration process is less than 1s,which has a great convenience in practical application.This paper introduces the importance of the security of people’s travel,releasing the self calibration of camera.The visibility monitoring system of highway adopts image processing,automatic calibration of the road surveillance cameras and other electronic industry means.It replaces the traditional manual measurement or through the erection of large instruments to measure the visibility and other cumbersome,inefficient means,whichbrings the possibility for the rapid detection of visibility.Finally,the test results for the calibration of a good picture show that the algorithm is feasible,has a good application value,and can be used in the real time detection system of Expressway visibility in the country.