Research On Enhancement Methods For Sea Image In Complex Sea Conditions

Image enhancement is an important part of image preprocessing,whose purpose is to improve the image contrast,highlight the features of the region of interest area,and improve the image quality and visual effect.This dissertation is dependent on the National Science&Technology Pillar Program "Airborne Infrared Search and Location Technology for Distressed Targets on Open Sea Surface",the program mainly studies infrared image data collection,image enhancement processing,target detection,target location and tracking,etc.,this dissertation mainly studies the enhancement technique of visable sea image and infrared sea image.The complex sea conditions in this dissertation's title refer to the normal sea condition,the light environment of backlight,the extremely dark environment caused by haze,and the interference environment by heavy sea waves.In emphasis this dissertation studies the effective methods of sea image enhancement in the histogram and gradient domain.HE(Histogram Equalization)is a common algorithm in image enhancement.The histogram reflects the distribution of image pixels at different grayscale levels,the advantage of the image enhancement algorithms based on histogram transforming is simple and easy to implement.Aiming at the characteristics of the histogram of the visable sea image,which is near to the normal distribution,this dissertation proposes the algorithm of three sub-histogram equalization,the algorithm divides the histogram into three sub-histograms,and then three sub-histograms are equalized respectively,which can preserve the image's brightness and enhance the image's contrast.This dissertation proposes the algorithm of histogram equalization with edge fusion.The algorithm of Laplacian filtering enhances the image by the differences from the adjacent gray value's change.Fusing with the classical HE result and Laplacian filtering result can enhance the visable sea image's details with low exposure and high exposure effectively.The image gradient reflects the change of gray value,and the human eye is more sensitive to the image gradient.As for the infrared sea image with less noise,this dissertation proposes the image details enhancement algorithm by gradient field bi-interval equalization.The preprocessing of infrared sea image with sea waves interference is to remove the sea waves interference and improve the details of target region,this dissertation proposes the algorithm of image gradient histogram's transforming,which can reduce the sea waves' interference and improve target region's contrast and entropy.The validity of image enhancement algorithm this dissertation proposes is verfied by the experimental visable sea images and infrared sea images.