| Owing to the rapid growth of population and technology, resources consumption, marine has gradually become a new world for human development. The rich marine resource can ease the predicaments of China at a certain extent. So the development of marine is a very important and influential project. Owing to limited terrestrial resource and rich energy, mine, biological resources in marine, so human paid attention to marine development long time ago. But human know little about marine, so how to use the infinite precious treasure effectively is a problem for the scientists to deal with.Because of complicated underwater environment and bad imaging conditions, images from underwater tend to be flooded in all kinds of noise. Imaging under the water is difficult and pollution images seriously affect people to explore the ocean. Therefore, effective denoising of underwater image and reflecting real underwater information clearly is one of the most important key topics in underwater research. The underwater images'pollution process and mechanism is very complex. Different underwater environment, the noise is not the same. The paper'goal is to try to use wavelet to denoise from underwater images, in order to improve the images'quality.Firstly, the paper introduces imaging characteristics of underwater images, analyses the main factors of underwater imaging. The paper also introduces forward scatter,backscatter and applications of the underwater images, introduces several common methods of denoising.Secondly, the paper studies the fundamental theory of wavelet transform ,analyses wavelet transform's characteristics and later expounds and compares three main methods for denoising in wavelet domain. The paper emphasizes threshold denoising methods in wavelet domain and introduces several classical methods of thresholds estimation.Thirdly, the paper analyses briefly the property of underwater images'background noise. Based on analyzing the property of underwater backscatter noise,decomposing the underwater images with wavelet. Then the paper takes the method of filtering the low-frequency coefficients with LUM filter and using the Neighshrink and its improved method ENS to deal with high-frequency coefficients. Because of the correlation between the wavelet coefficients, a new threshold denoising method which based on intra-inter scale dependencies between wavelet coefficients proposed in this paper. The results show the methods whit used in the paper can denoise the underwater images effectively.Finally, the paper analyses the problems that come across in the study and need to be addressed in future. |
