激光成像引信的目标识别方法研究

为了高效准确地滤掉云烟雾等悬浮粒子，减少对激光成像引信工作的影响，采用了改进的Harris+最小核值相似区域(SUSAN)角点检测算法与矩形度结合的目标识别方法。改进算法在原有Harris和SUSAN算法基础上，利用8邻域模板标准差对目标像素点进行初次筛选获得候选角点，经高斯滤波后，利用改进的角点响应函数值进行二次筛选，再通过非极大值抑制得到最终角点，最后利用矩形度对目标与干扰进行二次区分。通过理论分析和实验验证可知，95%的目标能被有效地识别出来。结果表明，该方法能高效准确地区分目标与干扰，同时满足实时性要求，为激光成像引信抗干扰方面提供了一定的理论参考。

Target recognition method of laser imaging fuze

In order to efficiently and accurately filter out the influence of suspended particles such as cloud smoke on laser imaging fuze, a target recognition method based on the combination of improved Harris+ smallest univalue segment assimilating nucleus (SUSAN) corner detection algorithm and rectangularity was proposed. Based on the original Harris and SUSAN algorithms, the 8-neighborhood template standard deviation was used to perform the initial screening of target pixels to obtain candidate corner points, and after Gaussian filtering, the improved corner point response function value was used for secondary screening, and then the final corner point was obtained by non-maximum suppression, and finally the target and interference were quadratically distinguished by the rectangularity. Through theoretical analysis and experimental verification, 95% of the targets can be effectively identified. The results show that the proposed method can efficiently and accurately distinguish between target and interference, while meeting the real-time requirements, and provides a theoretical reference for the anti-interference of laser imaging fuze.