一种预处理OWLCM算法下的图像特征匹配研究方法  

A pre-processing OWLCM algorithm for image feature matching research

作　　者：秦蒙[1] 郑冬 冯鹏[2] 龚识懿 杨泳 任照富 QIN Meng;ZHENG Dong;FENG Peng;GONG Shiyi;YANG Yong;REN Zhaofu(School of Information Engineering,Chongqing Electric Power College,Chongqing 400053,China;School of Optoelectronic Engineering,Chongqing University,Chongqing 400044,China)

机构地区：[1]重庆电力高等专科学校信息工程学院,重庆400053 [2]重庆大学光电工程学院,重庆400044

出　　处：《重庆理工大学学报(自然科学)》2025年第1期117-124,共8页Journal of Chongqing University of Technology:Natural Science

基　　金：重庆市教育委员会科学技术研究计划青年项目(KJQN202302605);重庆市科委技术创新与应用发展专项(cstc2021jscx-gksbX0056)。

摘　　要：原始图像的质量直接影响图像匹配的精度。传统图像增强方法往往不能与后续图像匹配算法有机融合,难以大幅度提高匹配精度。基于此,提出了一种基于OWLCM算法对图像进行预处理进而提高特征匹配精度的方法,该算法以LCM算法为基础,通过引入局部增强因子和非线性变换因子达到像素动态增强的目的,能够更有效地增强匹配图像质量。为了验证该算法的可行性,利用Middlebury立体视觉测试集进行了初步验证,开展了偏转、弱光、旋转、普通等多因素工况环境下的图像匹配。实验结果表明:与传统LCM、LCM-SURF特征匹配算法相比,所提算法图像平均正确特征匹配对数为524.25,相比LCM-SURF算法提高了277.50。The quality of the original image directly affects the accuracy of the image matching.Traditional image enhancement methods often cannot be organically integrated with subsequent image matching algorithms.Thus,it is difficult to improve the matching accuracy.To address the issue,we propose a method based on OWLCM algorithm to preprocess images and improve the accuracy of feature matching.Based on the LCM algorithm,the algorithm achieves pixel dynamic enhancement by introducing local enhancement factors and nonlinear transformation factors,which can more effectively enhance the matching image quality.To verify the feasibility of the method,the Middlebury stereo vision test set is employed for preliminary verification and the image matching is made under multi-factor working conditions such as deflection,low light,rotation and ordinary conditions.Our results show the average number of correct feature matching pairs in the proposed algorithm is 524.25 compared with that of the traditional LCM and LCM-SURF feature matching algorithms,which is 277.50 higher than that of the LCM-SURF algorithm.

关 键 词：图像匹配 图像预处理 视觉质量 多因素工况 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]