Research On Appearance Parameter Detection System Of Goods Vehicle On ETC Channel Based On Embedded Platform

At present,the method for over-limit online detection of high-speed intersection trucks is limited.The trucks can only perform manual over-limit detection and manual charging by stopping at the inspection station,which leads to frequent traffic congestion at high-speed intersections.In order to realize the over-limit online detection of trucks and the non-stop ETC charge,this paper studies an embedded ETC channel truck shape detection system.Under the embedded platform,the image processing technology is used to realize the trucks under the low speed driving condition of ETC channel.The rapid measurement of the external dimensions,the main research work is as follows:(1)System measurement program research.The detection system sets up two cameras for image acquisition on the front and side of the road according to the camera lens parameters,working distance and maximum size of the truck.The photoelectric switch is used to detect the state of the vehicle,and the camera is controlled to continuously collect the side view and the overhead view image of the truck.The image processing algorithm is used to calculate the characteristic parameters such as height,length and width of the truck.(2)Research on image processing algorithms.The feature point matching algorithm is studied,and the splicing of vehicle panoramic images is realized by ORB and SURF algorithm combined with F-SORT algorithm.The vehicle symmetry axis is extracted using a SURF algorithm transformed with symmetric feature matching.The body extraction is achieved by the background difference method and the method oflocal binarization according to the vehicle axis of symmetry.The wheel axis is detected by the Canny edge detection algorithm combined with the Hough circle transformation.Finally,the gray scale mean statistical method is used to determine the measurement points of the vehicle length,width and height.(3)System hardware platform construction and software implementation.The Linux operating system was ported using the Exynos 4412 development board of the Cortex-A9 core architecture.The C++ language is combined with the OpenCV visual library to implement image processing algorithms and system main program design.The measurement data is collated and saved through the Sqlite database,and the remote view of the detection results is realized by using the Boa server and the designed system interface.Finally,the camera was calibrated using the Zhang Zhengyou calibration method.The experimental system hardware platform was built,and the dimension measurement experiments were carried out on various types of trucks.The experimental results show that the detection system can realize the measurement of the outer dimensions of the truck stably and quickly,and the measurement results meet the requirements of the national standard for the maximum error of the automatic dimension measuring instrument.