一种基于红外图像的弹孔识别方法

Bullet Hole Recognition Method Based on Infrared Image Processing

为了满足在室外复杂环境下,快速精确的识别弹孔位置,提出一种基于红外图像的弹孔识别方法。采用帧间差分法,利用帧间的巨大差异产生的灰度值变化即可判定是否有子弹上靶,使用最大类间方差法得到阈值并对图像进行二值化,通过二值化后的图像找到连通区域重心即弹孔位置。实验结果表明,该方法可以实时精确地对新增弹孔进行快速识别,对不同光线环境有较高的适应能力,并能识别部分重叠的弹孔。

In order to identify quickly and accurately the location of bullet holes in outdoor complex environment,a bullet hole recognition algorithm based on infrared image processing was proposed.Inter-frame difference method was used to subtract the current frame image from the previous one.The change of gray value caused by the huge difference between frames would be used to determine whether a bullet was on target or not.The threshold was selected by the method of maximum inter-class variance for binarization.The center of mass of the connected region,i.e.the bullet hole position,was found through the binary image.The experimental results show that the algorithm can identify new bullet holes accurately and in real time,and has high adaptability to different environments,and can identify some overlapping bullet holes.