基于随机搜索算法的异源激光图像匹配方法  

A matching method of heterologous laser image based on random search algorithm

作　　者：王荣荣[1] 王团部[1] 吕林涛[1] 李洵 WANG Rongrong;WANG Tuanbu;LV Lintao;LI Xun(School of Information Engineering,Xijing University,Xi’an 710123,China)

机构地区：[1]西京学院信息工程学院

出　　处：《激光杂志》2019年第6期64-68,共5页Laser Journal

基　　金：陕西省工业科技攻关项目(No.2016GY-141);西安市科技产学研项目(No.2017087CG/RC050(XJXY001))

摘　　要：传统激光图像匹配方法在匹配亮度和对比度差异较高的异源图像时,存在匹配效果差的弊端,为此提出基于随机搜索算法的异源激光图像匹配方法,采用基于相位一致性的特征点和边缘检测算法,采集异源图像特征点和边缘图像,利用基于随机轮廓检索的图像配准算法,通过轮廓直线化算法采集边缘图像特征点后,通过随机搜索算法对提取的特征点实施鲁棒的匹配,完成异源激光图像配准。实验结果说明,所提方法对不同异源激光图像的匹配效果较佳,匹配速度增长率最大值是45%,正确率和效率均值都高于95%,该方法是一种高效率、高精度的异源激光图像匹配方法。The laser image matching method based on Harris corner detection has the disadvantage of poor matching effect when matching with different source images with higher brightness and contrast.In this paper,a method of matching heterologous laser images based on random search algorithm is proposed.The feature points and edge detection algorithm based on phase consistency is used to collect heterologous image feature points and edge images,and the image registration algorithm based on random outline retrieval is used.After the edge image feature points are collected by the contour linear algorithm,the extracted feature points are robust matched by the random search algorithm to complete the heterologous laser image registration.The experimental results show that the proposed method has a better matching effect for different heterologous laser images.The maximum speed growth rate is 45%,and the accuracy rate and efficiency mean are higher than 95%.This method is an efficient and high-precision heterologous laser image matching method.

关 键 词：随机搜索算法 异源激光图像 匹配 边缘图像 轮廓特征 配准 

分 类 号：TN247[电子电信—物理电子学]