军事知识服务平台
GM-APD激光雷达动态小目标超分辨率测距方法
Super-Resolution GM-APD LiDAR Method for Dynamic Small Targets
摘要 针对GM-APD单光子激光雷达在远距离、强噪声环境下对动态小目标测距易受噪声干扰、精度不足的问题,文章提出一种动态小目标超分辨率测距方法.该方法通过连续帧分块重构、密度筛选噪声过滤、最小距离值特征点群质心提取、十字邻域滤波,结合强度像与距离像优化加权平均,实现超分辨率时间间隔测距.在半仿真数据的实验测试中,针对静态小目标的测距精度提升 1.896 bin,置信度为 64%;针对动态小目标的测距精度改进量为 1.554 2 bin,置信度为75%.该统计模型模拟了飞行器进近时与单光子激光雷达运动耦合导致的复杂随机回波光子事件,这对后续算法的深入研究以及进一步验证系统性能边界具有参考价值.
Abstract:
GM-APD Single-photon LiDAR is susceptible to noise interference and offers insufficient accuracy when ranging dynamic small targets at long distances and in high-noise environments.This paper proposes a super-resolution ranging method to overcome these issues.This method achieves super-resolution time-interval ranging by combining continuous frame block reconstruction,density screening and noise filtering,centroid extraction of feature point groups with minimum distances,cross-neighborhood filtering,and optimized weighted averaging of intensity and distance images.Experimental tests using semi-simulated data demonstrate that the proposed method improves the ranging accuracy by 1.896 bin(64%confidence level)for static small targets,and by 1.554 2 bin(75%confidence level)for dynamic ones.This statistical model simulates the complex,random echo photon events caused by the motion coupling between an approaching aircraft and a single-photon LiDAR,thus facilitating subsequent algorithm development and system performance boundary verification.
刊 名 半导体光电 ISTICPKU
年,卷(期): 2026, 47(5)
分类号: TN958.98
机标分类号: TN958.98 TP391.41 P225.2
在线出版日期: 2026年11月4日