Research And Application Of Image Defogging Algorithm Based On Dark Channel Prior And Bilateral Filtering

Haze is a special kind of weather phenomena, and pure fog is configured by a large number of tiny water droplets which are suspended in the air. Fog makes a poor visibility and seriously affects outdoor devices of image acquisition based on computer vision.The fog also causes degradation of image quality, with the contrast and resolution being reduced to a different extent. In order to minimize the fog’s influence to these equipment, various types of defogging algorithm gradually came into being.The defogging algorithm proposed in this paper is based on the physical model.Compared to the vast majority of the similar methods, the algorithm in this paper has a better target or purpose.In this algorithm, the overall image is defogged not only to get a better subjective vision, but the defogged image can be effectively applied to the current detection and identification of activities which are based on computer vision.That is to say, it puts forward higher standards on the algorithm on whether it can keep more information of the original image, for example, keep more edge details of the image. The main contents and innovations in this paper are briefly described from three aspects as follows:(1)Improved algorithm for image removing using dark channel prior. The algorithm is used to defog comprehensively with greater pertinence, that is to restore protectively emphatically local details of fog-degraded images, for instance, restore protectively the edge and other details of fog image based on specific color and depth characteristics. After the image is restored, the edge and other details of local areas from the fog image can be detected and identified more effectively to be applied to some detection equipment based on computer vision.(2)An approach based on dark channel prior and bilateral filtering is proposed for image defogging. In order to simplify the time complexity,we solve passively atmospheric transmittance with bilateral filtering. The conventional mainstream defogging algorithms are mostly based on the model of atmospheric physics, they are accustomed to solving the atmospheric transmittance with direct methods. First obtain the estimated value, then refine the intermediate process images with software matting way, however, matting process requires a lot of computation.while the way of solving transmittance with bilateral filtering is reversed, and it solves the transmittance indirectly by estimating the atmospheric scattering function and atmospheric light valuefirstly. Accordingly, Largely reduce the computational complexity of the algorithm and Improve the timeliness of the algorithm.(3)Combine defogging algorithm with applications based on the image detected. Thus the algorithm is more targeted,with being applied to applications of image edge detection,color segmentation and landslides monitoring based on computer vision.