| With the rapid development of technology, the picture information become a reality to obtain information from the process of the main means, but also become the people in real life an indispensable source of information, but the main expression of image information is through the video display in front of people. A lot of aspects of video technology has gradually been used in people’s daily life, such as traffic control, aerospace, and other community warning most industries have video technology incorporated therein. But in the process of the captured video easily by many factors, such as: the resulting video image is vulnerable to the impact of light, which causes the video there were shortcomings uneven illumination, a greater impact to the visual effects brought about, so through image enhancement algorithms for video processing to reduce the impact of the video image by light interference, it is on technological development in this area has its important significance and economic value. Also in the transmission of video, it may be due to the influence of noise, signal interference and the amount of data transfer ones and other reasons, so that the video communication efficiency during transmission can not be well protected.This article mainly analyzes the classic image enhancement algorithm to solve the problem of light and and studys the H.264 coding standard used in the process of transmission, in view of the traditional Retinex algorithm’s existing problems of halo, color distortion and video transmission efficiency in the communication, this aticle puts forward the improved algorithm as follows.1、Aiming at the shortcomings of the traditional Retinex algorithm, proposes Retinex image enhancement method based on the improvement of bilateral filtering, to enhance the brightness of the above information of HSI space through a separate process, and then get brightness estimate image through the original brightness information of bilateral filtering with dynamic range compression processing and then get the reflection image and enhance processing and adaptive adjustment of saturation, finally transfer the enhanced HSI image into RGB image. Experiments show that the improved Retinex algorithm can effectively solve the problem of video lighting and solve the drawbacks of traditional algorithm, and has better effect on detail enhancement and color fidelity.2、In video communication, video of the user’s attention is generally concentrated in the interest area. We select one’s head shot of the video to process, according to the principle of MARR vision, on the premise of guarantee the subjective quality of video communication, this paper proposes a head video preprocessing method based on the interested region, which is a combination of face detection algorithm, H.264 coding and bilateral filtering, using the proposed algorithm can effectively improve the efficiency of video communications. The experimental results show that the proposed method can effectively improve the efficiency of video coding and can also ensure the quality of video’s subjective.3、There is a phenomenon of uneven illumination for videos which is collected in the case of complex light, in order to solve the existing problems in the above image videos, We use two improvement methods presented on the video acquisition system, and combine them together to preliminary processing of video system, through the experiment results from two aspects of histogram subjective evaluation and objective evaluation, it proves the feasibility and applicability of the system. |
