Research On Target Extraction Based On Dyadic Wavelet Transform

In the image target extraction,the main content of the research is to extract the edge of the image.The edge is the basic feature of the image,which contains the most useful information in the image.Aiming at the problem of ignoring some important information of low-frequency sub-images in target extraction based on dyadic wavelet transform.First,a fusion target extraction method combining wavelet modulus maximum method and Canny edge detection method is proposed.In the wavelet domain,the edges of high-frequency subimages are extracted(detected)by solving the maximum points of the local wavelet coefficient model,and the edges of low-frequency sub-images are extracted(detected)by the Canny edge detection method.The edges of the two sub-images are fused according to the weighted average fusion rule.Secondly,another fusion target extraction method combining wavelet modulus maximum method and Gabor edge detection method is proposed.In the wavelet field,the edges of high-frequency sub-images are extracted by solving the maximum points of the local wavelet coefficient model,and the edges of lowfrequency sub-images are extracted by using Gabor edge detection.The edges of the two sub-images are fused according to the absolute value fusion rule.Theoretical analysis and experimental results show that the two methods can not only effectively enhance image edges,but also solve the problem of target image extraction under complex backgrounds.