Real-time locating system for high speed and dim target based on FPGA

摘要

In test-range tracking, traditional machine vision systems are composed of a frame grabber and a host PC, which cannot meet the requirement of real-time processing. In this paper, we present a real-time locating system based on FPGA (Field Programmable Gate Array) to solve problems caused by high-speed and low-sharpness of the target in field testing. FPGA, as the system processing core, captures the high resolution images from TDI-CCD (Time delay and integration charge coupled device) line scan camera which is based on Camera Link interface. Two SRAM chips are used in this system to store line images alternately. The high frame frequency image data are stored in one SRAM chip by stitching and finishing while the data from the other are processed to analyze the target. Adaptive threshold and smooth filtering methods are proposed to get rid of the environmental or false target interference. For identifying the dim target from the mass sequential images, energy accumulation and frame difference are presented to reduce the false target alarm rate. Also, the sub-pixel grayscale centroid algorithm is presented in this paper to improve the location accuracy. Experimental results demonstrate that the pipelined architecture system can fulfill data acquisition and processing on frame frequency of 15000 frame/s with the line scan images of 8192 × 1 pixels in real time.