Methods And Technologies Of Underwater Range-Gated Image Enhancement

Because of the excellent imaging effect, influences by surrounding light uneasily and long distance transmission, the underwater range-gated imaging technology has became an important role in the development of marine resources exploration, marine construction, marine defense and so on. But the long-range target image is degrad ated seriously and needs image postprocessing operation. This paper carries out the researches about the methods and technologies of underwater range-gated image enhancement. The main work and achievements of this paper are as follows:1. This paper has expounded the present situation about the domestic and international development on range-gated imaging system. We have designed and analyzed the scheme of our underwater range- gated imaging system and has analyzed the reasons of underwater range-gated images’ quality degradation. Some contrast experiment has been taken using underwater images getted from the long sink, and we have get some results.2. A denoising method based on gradient and wavelet transform has been proposed. Firstly, use the cosine wavelet transform to decompose the noisy image into many different frequencily spacal image sets. The low frequencily spacal image uses a new image gradient enhancement method to improve the whole image’s texture information. The LH or HL space images which have the information of non-uniform strips use the surface fitting method to eliminate the whole image’s non-uniform strips. In the HH space denoising process, the lower level space images use the non- local means method to preserve the whole image’s similarity information, and the upper space images use the fractional integral method to preserve the whole image’s more details. Finally, use the inverse wavelet transform to obtain the result image. O ur denoise method can smooth the noise and preserve the image’s texture at the same time. The objective evaluating index is improved by 6%.3. A two dimensional double platforms histogram equalization algorithm has been proposed. Firstly, add a window to any pixel and calculate the mean value of all the pixels in the window. Then, put the relative information of the pixel and its adjacent pixels into the one dimensional histogram information statistics. Keep the relationship about the window center pixel and its adjacent pixels as another information in the two dimensional histogram. Equalize the histogram after proper calculation. Finally, calculated the enhanced image according to the histogram.4. A new estimation deviation suppressed retinex algorithm has been proposed. Using the guided filter as the kernel function, the illumination part from the primitive image is calculated. Then using the illumination part, the weight factors of the reflectance part are calculated. Combining the weight factors and the reflectance part, a new reflectance part is calculated. Finally, add the illumination part into the reflectance part to get the final enhanced image using a nonlinear fusion method. O ur new image enhancement method can enhance the target, background and boundary information respectively. The method can also do the background noise suppression and brightness balance. The objective evaluating index is improved by 12%.Finally, this paper has made the summary and prospects.