Research And Implementation Of FPGA Based Video Dehazing Algorithm

Recently, the fast development of industrial techniques induces the declining airquality. The fog and haze weather also becomes more frequently. The ashes in the airwill reduce the visibility, which induces image distortion captured in the fog scene.Thus, it is important to study the dehazing method.The traditional image dehazing methods usually try to improve the image qualityby enhancing the contrast of the distorted image. However, the enhanced image is notfacility. Recently, some methods try to improve the image quality by formulating thedehazing problem as an image restoration problem. These methods have greatrestoration performance and facility. However, the complexity of these methods isusually very high. In this thesis, we study the image dehazing methods with the imagerestoration framework.For video and image dehazing, our work focus on the following aspects:1. Proposed an algorithm of single image dehazing based on image texture. Thealgorithm is simulated by MATLAB. The experiment shows that thealgorithm can achieve image restoration efficiently with fast speed.2. Proposed an optimized video dehazing algorithm.The MATLAB simulationresult shows that the algorithm can avoid flicker effectively with fastprocessing speed.3. Implemented an image dehazing system based on FPGA. The experimentalresult based on FPGA is consistent with MATLAB simulation.4. Implemented PAL video dehazing system based on FPGA. Experiment showsthat the result is effective and the maximum frequency can reach about200MHz which can satisfy the requirements of real-time processing.