All types of cameras, such as digital cameras, surveillance cameras, and car-mounted cameras, have been applied in many aspects of life, with the advances in technology and improvement of living. However, bad weather (fog, haze, dust, rain, and snow, etc.) will reduce the performance of outdoor cameras, and make them hard to keep stable visibility. For surveillance cameras and car cameras, the images will be blurred under the situations of heavy fog, dust, and air pollution. People cannot identify clearly moving cars, pedestrian, and car plate number. To solve such problems, we have developed a fast high performance defogging technology. It calculates respectively the approximate fog density in single pixel and local area, with single input image. Then by combining them together, it will estimate precise fog density for each pixel, and restore the defogged image. After defogging, the brightness can also be improved adaptively. By software implementation on PC, we can achieve the run speed of 720x480@30fps. This technology is suitable for outdoor surveillance and navigation assistance.
The basic idea of defogging is, foggy image = clear image + fog density. That means foggy images are clear images polluted by fog. If we estimate the fog density and remove it, then the clear image can be got. The foggy image model is shown as follows.
The procedure of defogging is as follows.
In traditional methods, the step of getting transmission map will cost much time. We have developed a fast method for this. It calculates respectively the approximate fog density of single pixel and local area, and then combines them to get the precise fog density for each pixel.
The atmospheric light is estimated as follows. Firstly, we think the color of fog and haze is consistent, without color offset. However, the colors of natural objects are different. Secondly, we think that the farthest part of background is the area where the fog density is highest, and the place has the best consistency is the sky.
The restored image often seems a little dark. We will use a transform function to improve the brightness. The function is a quadratic Bezier curve controlled by defogging parameter.
1. Single image based defogging
2. Simple and efficient filter operations based on local pixel, suitable for parallel processing
3. Adaptive brightness enhancement
Run speed：720x480@30fps (CPU: 2.53 GHz, Intel i5)
Zhiming Tan: firstname.lastname@example.org
Akihiro Higashi: email@example.com
Share this page