Research On Underwater Image Enhancement Method Based On Histogram Equalization

Compared with the images captured in the ordinary natural environment,the images obtained in the water are affected by complex factors such as low illuminance,floating objects,water flow and so on,which leads to the serious degradation of the underwater image quality,such as serious color deviation,low contrast,blurred details and so on.These problems have become a big challenge in the follow-up research.In order to restore the good visual feeling of the underwater images and facilitate the follow-up research work,aiming at the problems of color deviation,low contrast,blur and lack of detail in the underwater images,an underwater image enhancement method based on dynamic histogram equalization and improved exposure value is proposed in this dissertation.This method processes the underwater images step by step in order to obtain high quality underwater images.Using color correction,detail enhancement and other processing methods to improve the quality of underwater images and make the images clearly discernible.The main research are as follows:The background and the significance of underwater image enhancement researches are briefly described in this dissertation.The existing underwater image enhancement methods are introduced and reproduced in detail,and these methods in theory and practice are analyzed,all of these provided a basis for the enhancement method that proposed in this dissertation.In order to solve the problems of serious color deviation,low contrast,image blur and lack of details in underwater images,after studying the histogram distribution characteristics of each channel on underwater images and the defects of dynamic histogram equalization technology,an optimized underwater image enhancement method based on optimizing dynamic histogram equalization is proposed in this dissertation.In the stage of histogram segmentation,the segmentation method based on exposure value is improved by considering the distribution range of histogram,which avoids the failure of segmentation.In the stage of constructing CDF,the reconstruction parameters are introduced to reconstruct the sub-histogram to achieve the purpose of adaptive constructing CDF.The over-enhancement phenomenon is suppressed while preserving the details of the enhanced images,and a reasonable remapping interval is allocated through the reconstructed sub-histogram.At the same time,the method of homomorphic filtering is introduced since the uneven illumination of underwater images.Together with the dynamic histogram equalization algorithm based on exposure value and limited contrast proposed in this dissertation to construct an improved histogram equalization method to enhance underwater images.The underwater image enhancement method that proposed in this dissertation is compared with several other common underwater image enhancement methods on the same data set.The superiority of this method in enhancing underwater images is verified by subjective visual evaluation and objective index experimental evaluation.In addition,a parametric comparison experiments are carried out using the underwater image data set with clear reference images to further verify the stability of the algorithm proposed in this dissertation.Through the comparison of different parameter values,the robustness of the algorithm is verified.Finally,the effectiveness of the algorithm is further proved by the application experiment of feature point matching.An underwater image enhancement system is developed based on the proposed algorithm.This system includes image enhancement module,image management module and image quality evaluation visualization module.The proposed algorithm is applied and tested in the developed system.Images from the data set are used as the system's input,and then users can select the different parameters to detect different images.Finally,as the output of the test results,the contrast effect images before and after enhancement are displayed.