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

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.