基于帧差法与边缘信息的红外目标检测算法研究

根据红外序列图像中运动目标具有连续性和一致性，本文提出了帧差法和边缘检测相结合的检测方法。首先采用连续帧间差分法处理图像得到运动区域，然后对当前帧进行canny边缘检测得到边缘信息，两者检测结果相与得到运动目标边缘；最后进行形态学运算得到精确的目标边缘。仿真结果表明，该方法克服了帧差法和边缘检测的不足，对复杂背景下的运动目标能够进行准确检测。

Infrared target detection algorithm based on frame difference and edge information

According to the moving target in an infrared image sequence with continuity and consistency, this paper proposes a detection method combining the frame difference and edge detection. First, get the moving regions using the continuous frame difference method to process the image, and get the edge information using the canny edge detection to process the current frame; then, get the edge of the moving target by and both of the detection results; finally, get the precise edge of target using morphological operations. Simulation results show that the method overcomes the deficiencies of the frame difference and edge detection, and can accurately detect moving targets in complex background.