Methods To Image And Video De-hazing

The digital image processing and computer vision has become focuses of academics research, such as computer science, information science, medicine and so on, and obtain the widespread application in the military, transportation and many other fields. Because of the atmosphere scattering phenomenon, bad weather (fog, haze, rain etc) affects the normal operation of outdoor vision systems, and then degrades the color and contrast of the images and video captured dramatically. Thus the vision systems (monitor, track, navigation etc) based on extracting image feature does not work very well, due to the internal characteristics of the image degraded by the bad weather. In this paper, based on the physical mechanisms and attenuation model, the scattered light is eliminated to increase scene visibility and recover haze-free scene colors and contrasts.In this paper, through analyzes the essential characteristic of the transmission, we propose a effective technology - non-linear bilateral filter method - for single image de-hazing, based on the physical mechanisms and attenuation model. And then image-based relighting technology is achieved by using the result of the de-hazing. After analyzing the video sequence characteristics and Computational complexity and utilized the high redundancy in video sequence reasonably, we present a method for estimating the non-key frames motion displacement between the key frames in the haze video sequence using bi-LK optical flow, solving the occlusion problem in video processing. Finally, it implements the video de-hazing. However, this measure dose not calculate the video volume spatiotemporal coherence. The Energy function is defined at the time domain and spatial domain can solve this problem. Using the energy function, and then it is optimized by the Graph Cut and obtain the optimum solution. In the end spatiotemporal coherence video de-hazing sequence is achieved.A large number of results are presented, detailed analyze and compare in our article, which adequate explain the de-hazing technology reasonable and effective. In the end, we summarize all researches in this paper, and give the direction of following research in future.