深空背景弱小目标实时检测算法研究

Research on real-time dim target detection algorithm in deep space

根据深空背景的特点,提出一种基于FPGA和DSP的检测算法,并设计相应的硬件系统,能对大小为512×512像素、帧频率为40帧/s的视频图像进行实时处理。先在FPGA中用最大中值滤波抑制空间成像噪声,解决了图像预处理运算量大对处理速度要求高的难题,再在DSP中用使用改进的双差分法进行运动目标分割,克服了单差分提取目标轮廓不准的缺点,最后结合候选目标的特征,采用基于逻辑的最近邻关联方法提取目标航迹。试验结果证明该算法满足了深空背景弱小目标实时检测的要求。

A real-time dim target detection algorithm in deep space is presented that is based on FPGA and DSP, and the corresponding hardware is designed. With this system, the video image, whose frame frequency reaching 40Hz and the size reaching 512 × 512 pixels, is processed in real-time application. Max-median filter is realized in FPGA to depress noise. The problems of mass computation and high speed request in image pre-processing are solved. Segmentation of the moving target is done with an improved dual image difference to overcome the disadvantage of the single image difference algorithm. An algorithm of track initiation and data association based on logic principle and the characteristic of target is presented. Practical application approves that the system perfectly meets the requirements of the real-time dim target detection and recognition in deep space.