正交小波与纹理一致性融合的图像识别算法研究  

Research on image recognition algorithm based on orthogonal wavelet and texture consistency fusion

作　　者：唐鑫 巫茜[2] 邹航[2] 刘恩 TANG Xin;WU Qian;ZOU Hang;LIU En(DINGZHAO(Chongqing)Packaging Technology Co.,Ltd.,Chongqing 400050,China;Chongqing University of Technology,School of Computer Science and Engineering,Chongqing 400054,China)

机构地区：[1]鼎兆(重庆)包装科技有限公司,重庆400050 [2]重庆理工大学计算机科学与工程学院,重庆400054

出　　处：《自动化与仪器仪表》2021年第6期26-30,共5页Automation & Instrumentation

基　　金：重庆市科技局重点项目(No.cstc2019jscx-fxydX0047);重庆理工大学教改项目(No.2020YB27)。

摘　　要：为解决一般图像特征识别算法难以兼顾图像细节与图像整体视觉感,探讨了一种基于双正交小波变换与纹理一致性测度融合的图像特征识别算法。该算法按照正交小波变换对源图像分别进行小波分解,按一定比例和纹理一致性测度确定融合图像的低频和高频小波系数矩阵,再经处理即可获得融合图像。实验仿真表明,提出算法不仅可较好地分辨出图像的伪边缘与丰富图像的细节信息,而且能兼顾图像整体视觉感,因此,可以比较显著地提高图像的识别效果。研究结果表明,采用上述提出的图像融合识别方法是可行和有效的。In order to solve the difficulty of considering image details and overall visual sense in general image feature recognition algorithms,an image feature recognition algorithm based on biorthogonal wavelet transform and texture consistency measure fusion was discussed.In this algorithm,the source images are decomposed by orthogonal wavelet transform,and the low frequency and high frequency wavelet coefficient matrices of the fused images are determined according to a certain proportion and texture consistency measure,and then the fused images are obtained by processing.The simulation results show that the proposed algorithm can not only distinguish the false edges of the image and enrich the details of the image,but also can take into account the overall visual sense of the image,so the recognition effect of the image is obviously improved.The research re-sults show that the proposed image fusion recognition method is feasible and effective.

关 键 词：图像识别算法 正交小波变换 纹理一致性测度 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]