Research On Object Clarity Of Video Surveillance Under Mist

With the continuous developm ent of image processing technology,video surveillance technology is widely used in various fields,which requires the system for a variety of occasions to have good adaptabi lity.However,many outdoor vision systems,such as ou tdoor security monitoring,road monitoring,tar get tracking,intelligent navigation,are very s ensitive to weather conditions.For example,in the weather conditions of fog and other adverse,due to the scattering of airborne particles in the air,resulting in blurred images captured,contrast and color are degrade d.Therefore,it is very important to study the method of image clarification in severe weather conditions.At present,there are tw o methods for clarifying the image in the m ist condition: firstly,from the perspective of image processing,enhancing the contrast of the image to achieve the purpose of image clarity,and the second one is based on the physical m odel of atmospheric degradation im age recovery method.In this thesis,from the im age enhancement point of view,the existing im age clarification methods are amended and improved,the specific work is summarized as follows:Firstly,the research status and developm ent trend of image clarification processing in fog conditions are introduced.Then,the reason why the im age is unclear under the mist is explained by the princi ple that the light intensity is weakened by the scattering and absorption of the air par ticles during the propagation of the light.Then,the theory basis of image enhancement under the mist video monitoring image is explained.Some traditional m ethods of m oving object detection are analyzed,including optical flow m ethod,frame difference method and background dif ference method.Combined with the research of this subjec t,background subtraction m ethod is used to detect the target and get the accurate target region.And the morphological processing is used to extract the target.The existing method of image fog removing enhancement-histogram equalization method,which is the most common appli cation,is pointed out.And the problems existing in these algorithm s are derived.In order to so lve this p roblem,this pap er proposes an improved POSHE algorithm based on the difference of information loss in the image collected un der the m ist,and th en makes a correc tion to the traditio nal algorithm.Firstly,the extracted image target is subjected to high frequency details of the reservation,and then target clear treatment will be conducted.On the basis of the ab ove theoretical research,a clear v ideo processing target under m ist is designed and implemented on the MATLAB platform,and the video files are processed clearly.The results show that the alg orithm can improve the performance of removing the mist of the target image to optimize the image information purposes.