Research On Image Edge Detection Models Based On Wavelet Transform

This thesis studies image edge detection techniques and uses the wavelet transformto construct some new image edge detection models with the advantages of singularvalue decomposition and local binary pattern. The main work of this thesis is asfollows:Chapter1introduces the research background and development history of imageedge detection technology both at home and abroadChapter2gives a brief overview of the history and theory of the wavelet andwavelet edge detection, and explains some classical edge detection operators inwavelet’s view.Traditional image edge detection algorithms do not perform well in noisesuppression. For solving the problem, chapter3presents a new method based onwavelet and local gradient field singular value decomposition to detect image edge.First, the wavelet transform is applied to the image in order to utilize the time-frequencylocalization characteristic of WT to extract the accurate edge character. Then the localgradient field singular value decomposition (SVD) is used to the gradient fieldcalculated by the wavelet. The feature and good stability of the singular values (SVs)enhance edge character and suppress noise. The experimental results of edge detectionillustrate that the proposed method can not only exact clear and complete single-pixeledge in the images without noise, but also find ideal edge from noised images.Chapter4gives an improved local binary pattern (LBP) edge detection model basedon wavelet with the advantages of wavelet and local binary pattern. The model firstlyuses the wavelet transform to find the gradient magnitude image, then applied theimproved LBP operator to the original image with the adaptive dual-threshold Otsu toextract the desired edge.