基于FPGA和DSP的高速实时轨道巡检图像采集处理系统

基于FPGA和DSP的高速实时轨道巡检图像采集处理系统由光学系统、FPGA模块、DSP图像处理模块及上位机组成。为提升图像采集的质量,提出并设计线阵相机加激光光源的组合方案,用以有效滤除阳光干扰,避免图像过度曝光。针对高速实时采集需求,设计FPGA采集模块,实现巡检图像采集控制和传输。针对实时智能检测需求,开发基于DSP模块的JPEG压缩、钢轨区域识别和扣件定位算法以及基于工控机的扣件缺陷识别算法。设计上位机软件,实现用户交互及数据存储。试验测试结果表明:该系统对钢轨定位的准确率在99%以上,对扣件定位的准确率在90%以上,对异常扣件识别率在80%以上,系统检测的最高速度可达160 km·h-1。

High-Speed Real-Time Acquisition and Processing System of Track Inspection Images Based on FPGA and DSP

The high-speed real-time acquisition and processing system of track inspection images based on FPGA and DSP is composed of optical system,FPGA module,DSP image processing module and upper computer.In order to improve the quality of image acquisition,an integrated scheme of linear array camera and laser light source is proposed and designed to effectively filter out the sunlight interference and avoid the overexposure of images.For high-speed real-time acquisition requirements,the FPGA acquisition module is designed to realize the control and transmission of inspection image acquisition.For real-time intelligent detection requirements,the algorithm of JPEG compression,rail area identification and fastener positioning based on DSP module and the fastener defect identification algorithm based on IPC are developed.The upper computer software is designed to realize user interaction and data storage.The test results show that the accuracy of the system for rail positioning is more than 99%,the accuracy of fastener positioning is more than 90%,and the recognition rate of abnormal fastener is more than 80%.The maximum detection speed of the system can reach 160 km·h?1.