Industry News
    Home > News
  • Video Surveillance DFG defogging algorithm effect previews

        DFG algorithm is implemented in real-time IP camera side, without any additional cost , which greatly improves the security monitoring quality in the fog and haze . We believe in the haze enveloped weather, use DFG algorithm is still able to implement IP camera security monitoring ; the technology on the basis of atmospheric transmission model combines image enhancement technology to achieve ideal image effect .
        In recent years , due to the deterioration of air quality , haze phenomenon of increased harm and aggravated . Especially from last year , " haze " has become the annual keywords. In the fog and haze of video surveillance equipment is mainly affected by the following three aspects:
        1 , the reflected light scattering produced by the decay of atmospheric particles , resulting in degradation of contrast ;
        2 , part of the larger particle size of atmospheric particles shading, resulting in loss of image detail ;
        3 , unrelated to participate in natural light through refraction imaging, resulting in image saturation, contrast and hue shift lower .
        Outdoor scene image contrast and color will be changed or degraded , making the image contains many of the features are covered or obscured , resulting in video surveillance products can not capture a clear image of the scene , the city of peace -building and security in various places implementation and management have a serious impact . Therefore, the security of the HD IP camera can have real-time defogging functions proposed requirements.
        DFG to fog algorithm efficiency previews
        To fog technology can be divided into physical camera to fog and mist two figures go . Physically go through the fog mist that is mainly achieved by the optical camera lens , high Refreshing fog lens on general industry was able to achieve great power zoom lens , expensive , generally used in ports, such as the high point of forest scenes .
        Digital defogging is on camera or back-end software implementation is based on human visual perception model design backend image restoration technology, low cost, ease of deployment and other characteristics suitable to promote widely used in urban monitoring .
        A security solution provider , developed a system to facilitate the realization of embedded DFG to fog algorithms that can quickly achieve real-time defogging framing effect. The technology based on the model of the atmospheric transmission combines image enhancement technology to achieve ideal image effect , and has achieved commercialization .
        In the image to the fog in this area , almost no one does not know "Single.Image.Haze.Removal.Using.Dark.Channel.Prior" this article. This article from nearly 1,500 contribution in 2009 CVPR Computer Vision and Pattern Recognition Conference emerged as the best paper in the field of computer vision in this top international conferences. In this paper , the authors found a very simple but surprising statistical laws , and proposed an effective approach to fog - fog algorithm based go dark tunnel. Because the mist is always gray , so once the image is affected mist , then those who have very dark objects will become gray. And the mist is formed on the physical formulas , but also according to the degree of white haze of these objects to determine the concentration . So dark path algorithm can effectively eliminate the impact on the fog generated to restore the true video .
        In recent years, almost all the papers to the fog algorithm based on the theoretical study of this thesis . The industry also has security manufacturers on the basis of this principle to the further development of real-time DFG fog algorithms, real-time and pay more attention to its effectiveness, and can be realized in the embedded side .
        In the field of computer vision , the people affected by the following equation to describe the image of fog and haze caused .
        I (x) =. J (x) t (x). +. A (1? T (x)). (1)
        Where I is the observed image fog , J is to be restored image without fog , A is a global atmospheric light component , t is used to describe the transmission of light through the media during the camera portion is not scattered .
        Order L (x) = A (1-t (x)), is called ambient light. Model then becomes:
        I (x) =. J (x) (1-L (x) / A) +. L (x) ... (2)
        J (x) = (I (x)-L (x)) / (1-L (x) / A) ... (3)
        Demister is observed from the I (x) estimated in L (x) and A, and then recover J (x).
        After removal of the gray mist , image will appear darker , and the contrast and saturation than normal scenes are much lower . Therefore, the image restoration algorithm need to mist concentration , and then adjust a contrast and saturation.
        DFG algorithm advantage
        1, the color reproduction is good , there is no local histogram equalization algorithm is prone to false colors and halo phenomenon ;
        2 , real-time, to ensure that the 1080P 30fps, effects and get the best papers CVPR close to fog algorithms , but much smaller amount of computation and enhance the image of the real-time ;
        3 , detail and edge retention is good, strong permeability . E.g. FIG rope and text can be clearly displayed , the image is relatively transparent ;
        4 , the image will not darken . After the general algorithm to remove the white fog , often causing the entire image is dim, visual effects . However, DFG intelligent algorithm to enhance the luminance can be achieved according to the environmental conditions , improve dim phenomenon ;
        5 , applicability, equipment open to fog function automatically detects there is no fog and mist concentrations , so it does not change the image tone when no fog , but when the concentration fog foggy enhance contrast and saturation , improve visual quality of the image.
        Epilogue
        DFG algorithm is implemented in real-time IP surveillance camera side, without any additional cost , which greatly improves the security monitoring quality in the fog and haze . We believe in the mist shrouded the weather , using DFG algorithm IP camera security monitoring can be implemented still play an important role in maintaining social stability , safeguard the normal order of society , to build a harmonious society .