Research On Image Clearing Methods Based On Rough Sets In Foggy Weather

As a new mathematical tool, rough sets theory has been successfully applied to the pattern recognition, image processing and other fields. Especially, rough sets theory shows its unique superiority for analyzing and handling vague and uncertain knowledge. Therefore, the thesis puts forward an improved clearing method based on rough sets, combined with the fog-degraded image's comprehensive image features and optimal division algorithm. The method is mainly based on the decision table to generate intelligent decision rules, divide the image, and finally enhance the sub-images to achieve effective fog-degraded image clearing effect.After analyzing traditional degraded image clearing methods in foggy weather, the thesis introduces rough sets as a new mathematical tool to analyze and process degraded images in foggy weather. The main research work of the thesis is as follows:(1) According to the vagueness and uncertainty of the degraded images in the foggy weather, the thesis analyzes and discusses the disadvantages of traditional image clearing methods based on the model for the fog-degraded images, and selects the image enhancement method. In addition, the thesis analyzes traditional global image enhancement methods and compares them with the local ones, and proposes the local histogram equalization algorithm based on mean segmentation.(2) The thesis studies the basic principle of rough sets, and applies it to the foggy image enhancement methods, achieving the good results. Then, the thesis tests and verifies the superiority and feasibility of the image clearing methods based on image classification of rough sets'indiscernibility relation in image clearing application.(3) In order to overcome the problems that the above algorithms based on rough sets have the fuzziness of image features information, the subjectivity of classification threshold selection and the inaccuracy of image classification for the degraded images, the thesis puts forward an improved clearing method based on rough sets to ensure the accuracy of fog-degraded image classification, combined with the comprehensive image features and optimal division algorithm. The method firstly builds a decision system and derives classification decision rules from the decision system on basis of the data analytical ability of rough sets, and then divides the image according to the matched decision rules and enhances respectively the sub-images to get the clear image. Finally, the thesis tests and verifies that the method satisfies the requirement of the real-time system and makes the fog-degraded images get much clearer.The thesis describes the detailed algorithm design flow, and realizes the video monitoring system. Our main research work is to prove effectiveness of the thesis's algorithm through the system's testing. Finally, after summing up the full thesis, we also put forward some further discussion and research work.