基于Zernike矩与灰度计算的微弱光学图像智能识别方法  

Intelligent recognition method of weak optical image based on Zernike moment and grayscale calculation

作　　者：胡贵恒 张震[2] 陈翠红[1] HU Guiheng;ZHANG Zhen;CHEN Cuihong(School of Information Engineering,Anhui Business and Technology College,Hefei 231131,Anhui,China;School of Application Engineering,Anhui Business and Technology College,Hefei 231131,Anhui,China)

机构地区：[1]安徽工商职业学院信息工程学院,安徽合肥231131 [2]安徽工商职业学院应用工程学院,安徽合肥231131

出　　处：《上海电机学院学报》2022年第2期100-105,共6页Journal of Shanghai Dianji University

基　　金：安徽省职业与成人教育学会重点资助课题(azcg21);安徽省高等学校自然科学研究重点资助项目(KJ2020A1096);安徽省质量工程省级教学团队资助项目(2020jxtd032)。

摘　　要：为了准确完成微弱光学图像的智能识别,提出基于Zernike矩与灰度计算的微弱光学图像智能识别方法。结合Zernike矩和灰度矩阵共生阵法,提取微弱光学图像整体轮廓特征和多种图像有效纹理特征,并采用加权融合方法融合这两种特征的向量,获取两者融合后的特征描述子,采用模糊数学法识别微弱光学图像。测试结果表明:Zernike矩的阶数为5时,可完成图像整体轮廓特征和纹理特征的准确描述,并可保证融合后特征的全面性,在噪声的干扰下依旧可准确完成目标图像识别,并且识别时间较短。In order to accurately complete the intelligent recognition of weak optical images, an intelligent recognition method of weak optical images based on Zernike moment and gray calculation is proposed. Zernike moment and gray matrix co-occurrence matrix method are combined to extract the overall contour feature description vectors and various image texture effective features of weak optical image. The weighted fusion method is used to fuse the two feature vectors to obtain the feature descriptors after the fusion. The weak optical image is recognized by the fuzzy mathematics method. The test results show that when the order of Zernike moment is 5, it can complete the accurate description of the overall contour features and texture features of the image, and ensure the comprehensiveness of the fused features. Under the interference of noise, it can still accurately complete the target image recognition, and the recognition time is short.

关 键 词：ZERNIKE矩 灰度计算 光学图像 智能识别 特征融合 

分 类 号：TP1833[自动化与计算机技术—控制理论与控制工程]