A new distance detection algorithm for images in deflecting angle

With the development of transportation industry, there are increasing numbers of surveillance cameras set up along the road. When taking camera calibration in these cameras, there are always no suitable calibration targets. Moreover, it is impossible for the cameras which have been installed to meet the requirements such as the locations, the directions of movement and so on. Thus, the traditional method of camera calibration cannot be used as usual. This paper puts forward a new distance detection algorithm for images in deflecting angle based on rectangular pattern. When it exists a certain angle of difference between the projection direction of camera optical axis and the road direction, it is also feasible to realize the transformation between image pixel distance and actual distance. This method establishes a three-dimensional model for camera calibration, which can be divided into two plane models and be used to calculate the distance in two directions. This improvement can make the distance detection more accurate and be free from angle of inclination. This method can be used in the application of foggy visibility distance detection and so on.