反无人机图像导引头远距空中目标探测技术  

Long-Distance Aerial Target Detection Technology of Counter-UAV Image Seeker

作　　者：赵飞 娄文忠[1,2] 冯恒振 苏子龙 汪金奎 宣炜琨 ZHAO Fei;LOU Wenzhong;FENG Huanzhen;SU Zilong;WANG Jinkui;XUAN Weikun(School of Mechatronical Engineering,Beijing Institute of Technology,Beijing 100081,China;Chongqing Innovation Center,Beijing Institute of Technology,Chongqing 400000,China)

机构地区：[1]北京理工大学机电学院,北京100081 [2]北京理工大学重庆创新中心,重庆400000

出　　处：《兵工学报》2023年第4期1023-1033,共11页Acta Armamentarii

基　　金：机电动态控制重点实验室开放课题基金项目(2021年)。

摘　　要：以低成本反无人机制导弹药为研究背景,开展针对远距无人机目标探测的研究工作,并设计一种局部视觉显著聚类测量算法。将局部对比测量思想引入可见光图像中,通过度量局部成像域的光谱聚类性实现对目标的检测,更具体的是度量局部像域的平均光谱值与相邻像素光谱值的最小距离。为解决多尺度目标问题,设计相应的多尺度滑窗测量方法。对原始RGB图像帧进行分频中值滤波,将滤波后的RGB图像转换到Lab颜色空间;通过滑窗模型进行无人机成像域搜索;使用显著性检测方法度量光谱差异性,得到显著测量图;利用阈值化算法获得潜在无人机目标的像素位置。根据无人机目标成像条件,开展远距无人机图像数据集的实地拍摄和人工合成工作。定性实验结果表明局部视觉显著聚类测量算法可在复杂背景下将小尺度无人机辨识,定量实验结果表明该算法的检测准确率可达到100%。Taking the low-cost counter-UAV guided munitions as the research object,this study focuses on long-distance UAV target detection,and the design of a local visual saliency-based clustering measurement algorithm.In this algorithm,the idea of local contrast measurement is introduced into the visible light image for the first time,and the detection of the target by measuring the spectral clustering of the local imaging domain is realized.More specifically,it measures the minimum distance between the average spectral value of the local image domain and the spectral values of adjacent pixels.In addition,to solve the multi-scale target problem,a corresponding multi-scale sliding window measurement method is designed.The brief flow of the whole algorithm is as follows:frequency-division median filtering was performed on the original RGB image frame;to better measure the spectral difference,the filtered RGB image was converted to the Lab color space;the sliding window model was used for UAV imaging domain search;the saliency detection method was adopted to measure the spectral difference to obtain a saliency measurement map;finally,a thresholding algorithm was used to obtain the pixel position of the potential UAV target.According to the UAV target imaging conditions,field shooting and artificial synthesis of long-distance UAV image datasets were carried out.The experimental results showed that the algorithm can successfully separate the UAV target from the background under various complex meteorological conditions.

关 键 词：反无人机技术 图像导引头 显著性检测 像素聚类 

分 类 号：V279[航空宇航科学与技术—飞行器设计]