Research Of Real-Time Image Acquisition And Preprocessing System Based On FPGA

In rencent years, with the rapidly development of high-speed rail, more and more people pay attention to the high-speed rail’s safety. The high speed and high density running of high-speed rail increases the wear of the surface of the steel rail, and it will cause the suface of steel rail uneven and affect the smooth running of train, even may cause the train derailment accident. Usually, the railway authorities take the way of manual detect to check the steel rail, it is not only inefficient, but also low performance in real-time. The railway workers can not replace the steel rail in time, which has already been damaged.This paper designed a real-time image acquisition and preprocessing system, which is based on field programmable gate array (FPGA), as to be the front-end system of steel rail detection system. In the front-end system, it takes full advantage of FPGA’s parallel data processing and rich logic resources characteristic to realize the image acquisition and preprocessing, so as to be further detecting by the upper system, thus to make up the shortcomings of manual detection. After learning the basic theory about video and image, then designed the system block diagram. This paper used the top-level design idea to finish the system design, which used Verilog HDL programming language to accomplish the functional modules, from top to bottom. Those modules include the image acquisition module, reading and writing module, video graphics array (VGA) module, etc. At the same time, used median filtering algorithm and Sobel edge detection algorithm to preprocess those video images, so as to remove the noise and extract the steel rail’s contour. At last, did the experiment debugging and performance analysis of the system, and the video image was displayed on VGA monitor in real time, which has already been preprocessed.Through the VGA monitor, the experiment results show that the system not only can realize the video image acquisition in real time, and extract the steel rail’s contour quickly, but also remove the noise effectively, thus to realize the video image’s preprocessing. Also through the performance analysis and comparison, the system has characteristics of low-cost, low-power, high speed, flexibility, etc. And the system has the high real-time ability, with strong scalability. It can meet kinds of need of functionality and speed, which has great application value.