Research On Enhancement Algorithms For Image And Video In Dust Environment

With the computer vision technology and multimedia technology developing, the video surveillance system and imaging system are capable of capturing the rich image information. For example, toll station monitoring system, under the premise of sunny weather can get an accurate and clear the basic characteristics of the vehicle license plate number and other information, which can be used as query the vehicle information or an evidence of traffic violation etc. However, in the dust weather, due to the influence of dust particles, the outdoor monitoring system's video images will be affected in different degrees of degradation, leading to color distortion, lower contrast image, we need to reinforce image and video in such a condition to extract useful information.This paper through the algorithm of image processing, from both static and dynamic aspects, narrated clearly dust figure and processing video images. The main contents are as follows:1. To the degradation of dust image do disdust processingMethod 1:Using the multiscale Retinex algorithm enhance dust image, to some extent, this algorithm can remove dust, but image color become dark and details are not very obvious after enhancement, so we could use histogram equalization algorithm to improve image contrast. This paragraph according to the properties of the histogram equalization algorithm, color image is decomposed into three components firstly, then for each color channel do histogram equalization algorithm processing, it can better enhance image brightness. Finally, subjective evaluation and objective evaluation are used on the result of the experiment.Method 2:For image color attenuation due to sand and dust weather, so you need to do to the attenuation of each color compensation, restore the original image color to enhance the image in dust environment. First, according to the principle of scattering and color attention using experiential model calculated each color attenuation values produced by atmospheric propagation attenuation and to compensate; Secondly, using histogram equalization algorithm enhance the dust image contrast, then using the guide filtering method enhance image detail and filtering noise. Finally, with the purpose of verify the validity of the algorithm in the paper, subjective evaluation objective evaluation and noise testing would be used on the result.2. To the dust video sequences do disdust processingUse the sand image enhancement method based on experiential model to dust video to remove dust processing. Firstly, frame differential cluster algorithm is used to extract key frames, using the weighted Euclidean distance, which can accurately extract the key frames; Use the sand image enhancement method to enhance each key frames. Eventually, the clear key frames of video sequence are connected into a new clear video, the purpose of enhance degraded video is achieved.