Dehazing System For Digital Haze Image With Unsharp Masking and Color Attenuation Prior

Fog is one of the major problems in the world of computer vision because it can cause degradation or loss of important visual information in images or videos. Fog can cause decreased contrast, loss of detail, and color distortion in images or video recordings. This can hinder the ability of computer systems to perform object detection, pattern recognition, or other visual analysis accurately. To overcome this problem, research in digital image processing often includes techniques to clean or repair images affected by fog. Therefore, in this study, the author created a system for removing fog using the unsharp masking algorithm and color attenuation prior (CAP). The test carried out on the same image repeatedly obtained different MSEi, PSNR and time values with a small difference in value, with an average MSE of 0.25581, an average PSNR of 54.0531 and an average time of 329.1487 from ten trials with the same image. This is caused by several factors including the implementation of the algorithm, the parameters used, hardware conditions, system conditions, and variability in the algorithm itself. From the research conducted visually, the proposed method on the system is able to remove noise in the form of fog in the image, so this method is suitable for use in fog removal processing.