Denoising Based On Wavelet Analysis, Image Fusion Applications

The image is often corrupted by noise in its collection, acquisition or transmission, which influences the image quality and greatly affects to extract information from it, so the noise must be removed before the image can be analyzed and utilized. The image fusion technology is a fresh image processing technology, which is developed on the bases of the information fusion technology, it can integrate the image information from multi-sensor and provide more clear and effective pictures about scene, so it is widely used in the aspect of image processing and application in satellite remote sensing, medical, aviation, radar and so on.Wavelet analysis is a local analysis method in time-frequency domain and has a multi-resolution characteristic, which is based on the Short Time Fourier Transform. It is a useful tool to analyze the unstable signal that implements multi-scale analysis to the signal by the translation and dilation of the mother wavelet and extract effective information. Wavelet is widely applied in the field of image processing and pattern recognizes because it has a local character, lower entropy and it can get rid of pertinence, so it is a cogent tool, and its application is getting more and more importance.We will research the image denoising and image fusion technology based on the wavelet analysis theory. The main works of this dissertation are as fellows:1. Generally introduce the history and research actuality of image denoising and image fusion, point out the advantages of the wavelet methods than the traditional methods, and understand the meanings to research the image denoising and image fusion based on wavelet analysis.2. Elaborate the wavelet analysis theory, the continuous and discrete wavelet transform, the multi-resolution analysis, the single-wavelet decomposition and reconstruction algorithm, the multi-wavelet theory, the filter banks construction and promote the multi-wavelet decomposition and reconstruction algorithm based on the single-wavelet.3. Summarize the mean filtering method, median filtering method, the mean combination with median filtering method. Download the original sinsin map, add the noise, de-noise by the three traditional methods, find that the de-noising image have a poor quality though the traditional de-noising methods are simple. Introduce the wavelet threshold de-noising algorithm, and select belmont2 index map, respectively, remove the noise by the wavelet soft-threshold and hard threshold de-noising method.4.Introduce the principle of multi-wavelet de-noising, and download the original sinsin image, add noise, and then decompose and filter on the second floor by a single wavelet Sym4 and multi-wavelet GHM. With those mean, standard deviation and information entropy results data to verify the conclusion that the multi-wavelet de-noising is better than a single wavelet denoising.5. Summarize the simple image fusion methods and the wavelet image fusion methods and their advantages and disadvantages. And then analysis and compare the IHS+WT image fusion method and HSV+WT image fusion method, by examples we find the HSV+WT method is better than the IHS+WT method.