Research On Exterior Parameter Detection System Of ETC Freight Car Based On Image Processing

Transportation industry has developed rapidly and the mileage of highway construction has continuously increased in China in recent years.The development of the transportation industry not only promoted economic growth,but also brought problems such as frequent road traffic accidents and urban traffic congestion.The issue of quick charges at highway entrances and exits is one of them.ETC technology can solve the problem of slow charging for small cars.However,some owners have illegally refitted the trucks to increase their revenue,which causing serious overload problems.Then the truck shape become variability,which cannot be quickly charged through ETC technology,and only rely on manual measurement.Because of this,the efficiency is extremely low,and congestion at highway entrances and exits is easy to occur.In order to improve the transportation capacity of highway,an ETC truck shape parameter detection system based on image processing was studied,and the rapid measurement of shape parameters of trucks was achieved without stopping.The main research work is as follows:(1)Camera calibration.In the camera imaging model,the relationship between the world coordinate point and the pixel coordinate point is obtained.The internal and external parameters of the camera were obtained using the Zhang Zhengyou calibration method.By comparing the parameters obtained from the calibration charts of different numbers and different positions,we know how many calibration charts can be used to get the best parameters.(2)Research on image processing algorithms.First get the background image and the foreground image before and after the truck enters the detection area.Remove the background and extract the target vehicles image by using background differential methods.By comparing the detection results of several different edge detection operators and selecting the Canny edge detection operator to extract the exterior contour of the truck,the exterior rectangle boundary of the image was drawn.Then determine the pixels on the edge of the truck by extracting the corners of the external rectangle using Harris corner detection method.Finally calculate the relevant dimensions of the width and length of the measured truck in the image.(3)Establishing the measurement scheme for the detection system.According to the calculation of data such as camera angle of view,placement height,and the maximum appearance of the truck,the scheme of image acquisition based on multiple monocular cameras is determined.By using multiple photoelectric switches triggering multiple cameras at the same time,the image of the front and the rear of truck are acquired.The width of the truck and the length of the front of the truck are calculated by the image processing algorithm.The length of the truck is finally calculated by considering the fixed interval of the camera.The height of the truck is measured using a laser ranging sensor.(4)Experimental verification.According to measurement scheme of the system,the suitable hardware device is selected to test the detection system.First,the parameters of the camera at different heights are calibrated.Then the image was collected for different size trucks.The trucks shapes are calculated by using the parameter of the camera and the software algorithm of the system.Finally,the cause of the measurement error is also analyzed.