Study To Threshold Function De-noising Of Image Based On Underwater Image Wavelet Transform

Human understands the marine by underwater filming, but because of thecomplexity of the underwater environment, harsh imaging conditions, the underwaterimages captured are often submerged in a variety of noise. The difficulties ofunderwater imaging and serious pollution of underwater images affect people onmarine exploration. Therefore, it is one of the key issues de-noising underwater imageto reflect a clear true underwater information. The process of underwater imagepollution is very complex, underwater environment is different, the noise is differenttoo. The purpose of this paper is combining wavelet and high-pass filter method forunderwater image de-noising to improve the image quality.Wavelet transform has many advantages, so it is used in image de-noising moreand more today. One of the wavelet threshold function de-noising algorithm’advantages is that principle and realize is very easy and it is very practical. manyscholars in our home and abroad put attentions on it. This paper that bases oncommonality and real-time has a detailed study about the problem of underwaterimage de-noising, and has a comparative study on several image de-noising methodsto make sure that the wavelet threshold function de-noising is one of the good methodto de-noising the underwater image.Firstly, the paper introduces imaging characteristics of underwater images,analyses the main factors of underwater imaging. The paper also introduces forwardscatter、backscatter and applications of the underwater images, introduces severalcommon methods of de-noising.Secondly, the paper studies the fundamental theory of wavelettransform,analyses wavelet transform’s characteristics and later expounds andcompares three main methods for de-noising in wavelet domain. The paperemphasizes threshold de-noising methods in wavelet domain and introduces severalclassical methods of thresholds estimation. Thirdly, the paper analyse prime elements of underwater image,and introduceshard threshold and soft threshold de-noising methods,on the basis of theoreticalanalysis, the methods between hard and soft threshold de-noising for underwaterimage backscatter de-noising is proposed,it will choose different threshold to highfrequency and low frequency, which overcomes the lack of the traditional waveletwith one threshold de-noising in underwater image processing. Experiment shows themethods can improve signal-to-noise ratio.Finally, the problem in the research and the problem needing to be solved in thefuture are analysesed.