Research On Enhancement Algorithms Of Degraded Images In Harsh Environment

Images quality of image acquisition systems is susceptible to environment in the real life. Images lose details, color information and contrast in harsh environment, it seriously affects the subsequent image processing and application. Therefore, it is important to enhance the degraded image in harsh environment. This thesis analyzes the degradation reason and characteristics of images in harsh environment, studies the common image enhancement algorithms and makes improvement and perfection of existing processing methods. The main points of the research are as follows.Firstly, a haze removal algorithm based on bright and dark region segmentation is proposed. When the haze removal algorithm which is based on mid-value filtering processes images, blackening phenomenon usually exists in the enhanced image. A region segmentation of bright and dark region in the junction of depth of field is taken, and the dark channel is amended according to the segmentation. The haze-free images are restored by the amended dark channel. Besides, as the haze-free images have the problems of color distortion and small brightness in dense haze regions, a post-processing algorithm is proposed. The algorithm increases the brightness of dense haze regions, and makes image fusion with the haze-free images. The experiments show that the algorithm could slove the blackening problem and improve visual effect on dense haze regions.Secondly, an improved Retinex algorithm is proposed. Traditional algorithms have the problem of excessive enhancement and the enhanced image usually looks unnatural. To solve this problem, the algorithm adds a decision function, and combines the reflection components and incident components as the enhanced images. Besides, considering the improved Retinex algorithm could not obtain good visual effects in different illumination regions while processing backlight images, illumination segmentation model is taken to segment images and appropriate processing methods are used in different illumination regions.Lastly, a comprehensive evaluation for image enhancement is implemented. This method chooses appropriate image quality evaluation indexes according to the type of degraded images, calculates corresponding weight coefficients and constructs synthetic evaluation function to measure the image quality. The experiment shows that the method could make accurate evaluation for enhanced images in harsh environment.