Research And Application Of Texture Extraction Algorithm Based On Multi-scale Gabor Wavelet

This paper mainly studies the origin of Gabor wavelet texture extraction and its application on different scales.It studies wavelet analysis theory and artificial intelligence theory and its key techniques in texture feature extraction.With the design and realization of multi-scale Gabor filter as the purpose,the paper studies and discusses the key algorithms and key technologies involved in its application in theory and practice.The paper included the following:This paper summarizes the present situation of texture feature extraction,wavelet transform theory,Gabor wavelet artificial intelligence algorithm and its application in various kinds of data,and analyzes the existing problems.At the same time,the basic theory and key technology directly related to the thesis are described.The principle of Gabor transformation is elaborated,and the expression of Gabor transformation is given.According to theoretical analysis and practical needs,a multi-channel Gabor filter set and a modulated Gabor filter are designed to extract the texture characteristics of the image.The texture feature classification algorithm based on multi-channel Gabor filter and the texture feature classification algorithm of adaptive Gabor filter are given accordingly.The advantages and disadvantages of multi-channel Gabor filter and modulated Gabor filter in texture feature extraction and its suitable application range are discussed from two aspects.The main research idea of this paper is to use Gabor wavelet to extract texture features of various image data.Research ideas: Analyze the problems that Gabor wavelet faces in the texture feature extraction process from theory,learn the existing improvement methods,and combine the advantages of each method organically.The application range of Gabor filters under different combinations is discussed,and the optimization theory and modulation technology are combined to enable Gabor filters to obtain effective texture characteristics.The main contents and achievements of this study are:(1)Analyze the advantages and disadvantages of Gabor filter in texture feature extraction,analyze the characteristics of natural image texture information,and optimize the parameters of Gabor filter for the texture characteristics of the image,making it more suitable for the texture feature extraction of the target image.(2)Based on the principle of Multi-Resolution analysis,a multi-channel filter based on Gabor wavelet is designed.The filter set can extract features of hyperspectral images and has good robustness.(3)In order to meet the requirements of real-time in industrial applications,a modulated Gabor filter is used to extract texture features of image data.An adaptive method for detecting and locating defects of Jacquard fabric by comparing local and local texture features is proposed.