Research On Real-time Enhancement Of Fog-degraded Image Based On FPGA

Defects such as poor visibility and low contrast are often existed in fog-degradedimages, which have some negative influence in public transportation, communitymonitoring, and production of human life. Therefore, the sharp processing offog-degraded images becomes a critical issue for image processing and machine visionarea. With the constant and rapid development of computer science in recent years, theprocessing of fog-degraded images has already become reality, which further putsforward higher requirements for this processing. The real-time image processing systemis quite an important indicator in practice, and it can be massively used in practice onlywhen the collected images are processed by the real-time image enhancementprocessing. The research of sharp processing of fog-degraded images and thepretreatment of image acquisition system are abundant both at Home and abroad now.The concrete scheme of sharp processing of fog-degraded images is relatively few, thushow to deal with fog-degraded images in real time proves to be a vital researchdirection.The enhancement and restoration of fog-degraded images are the main approachesin the sharp processing of fog-degraded images nowadays. According to the real-timerequirements of fog-degraded images, the author in this paper focuses on the contractenhancement of fog-degraded images, combines the characteristics of these images andfurther puts forward a simple and new effective algorithm in term of the present defectsof contrast enhancement algorithm. Finally makes it come true based on fieldprogrammable gate array (FPGA).In the first chapter，The author makes a massive research in the correlativeliterature review both at home and abroad, introduces the current situation and in theend determines to do the research based on FPGA, with image contrast enhancement asthe main process method.In the second chapter, the degrading mechanism and features of the fog-degradedimages is analyzed and the evaluation criterion on image quality is given, which providefoundation and basis for the following image processing.In the third chapter, a improved algorithm based on histogram equalizationalgorithm is proposed, which aim to overcome the defects of traditional histogramequalization. The improved algorithm uses a threshold value to avoid the defects such as over-enhancement and the details lost, and enhance the contrast of image to the greatestbest, and keep the real-time ability in the same time. The simulations of the improvedalgorithm are done to analyze the advantages of the algorithm.In the forth chapter, A real-time system of image capturing and displaying isrealized, the detailed implementation schemes of image capturing module、imagestoring module and image displaying module are given, which providing a platform forthe following verification algorithms.In the fifth chapter，a detailed hardware implementing of algorithm is given, andthe scheme of the adaptive selection of threshold value using the system onprogrammed chip(SOPC). The logic verification and resources consume is analyzed inthe last.