改进Hausdorff距离和量子遗传算法在激光制导中的应用  

Application of improved Hausdorff distance and quantum genetic algorithm in laser image guidance

作　　者：张腾飞 张合新[1] 孟飞[1] 强钲捷 杨小冈[1] 

机构地区：[1]第二炮兵工程大学控制工程系,西安710025

出　　处：《激光技术》2016年第3期320-325,共6页Laser Technology

基　　金：国家自然科学基金资助项目(61203189)

摘　　要：为了在激光成像制导中提高目标识别的精度和实时性,并在遮挡条件下进行有效识别,采用基于改进Hausdorff距离和量子遗传算法的激光图像匹配算法,选择图像的局部边缘特征为特征空间,针对传统Hausdorff算法及几种改进Hausdorff距离存在的问题,提出了一种新的改进Haussdorff距离作为相似性度量;在搜索策略上,选择量子遗传算法进行并行搜索,为防止种群过早收敛,提出了种群灾变策略,并应用动态的量子旋转角调节收敛的速度和方向。通过理论分析和实验验证,取得了不同参量条件下的目标识别对比数据。结果表明,该算法可以消除激光图像中局部遮挡、噪声以及出格点等因素影响,鲁棒性好、匹配精度高、计算速度快。In order to achieve high matching precision , good real-time performance and availability of target recognizing under shade condition in laser imaging guidance , a laser image matching algorithm was proposed based on improved Hausdorff distance and quantum genetic algorithm .In terms of the traditional Hausdorff algorithm and the problems of improving Hausdorff distance, the local edge feature of the image was selected as feature space .A new algorithm of improving Hausdorff distance was proposed to use it as a similarity measure .In the search strategy , the quantum genetic algorithm was chosen for parallel search . In order to prevent premature convergence of the population , the population catastrophe strategy was proposed and the speed and direction of convergence were adjusted by applying dynamic quantum rotation .Through theoretical analysis and experimental verification, target recognition contrast data under the condition of different parameters was obtained .The results show that the new algorithm, with good robustness , high matching precision and fast computing speed , could eliminate the effect of partial occlusion, noise and outlier .

关 键 词：激光技术 目标识别 HAUSDORFF距离 量子遗传算法 

分 类 号：TN911.73[电子电信—通信与信息系统]