A height limit early warning method based on binocular stereo vision

Abstract

The invention relates to a height limit early warning method based on binocular stereo vision, which comprises the following steps: S1, obtaining a disparity map of an original image based on the original image shot by a binocular camera, and obtaining a current driving path of a vehicle in a world coordinate system; S2, calculating a first coordinate of each pixel point projected to a world coordinate system in the disparity map; S3, judging whether a corresponding pixel point in the disparity map is in a driving path according to the first coordinate, and if yes, projecting the pixel point to a vertical coordinate plane parallel to the image surface of the binocular camera in a world coordinate system; S4, whether an obstacle exists is judged based on a preset height area on the verticalcoordinate plane, and if yes, the average height of the obstacle is obtained; and S5, repeating the steps S1 to S4 to judge and track the obstacle in multiple frames, further confirming that the obstacle is a height limit mark, and giving an early warning. According to the method, the calculated amount is small, and the accuracy is high.

Description

technical field [0001] The invention relates to a height limit early warning method, in particular to a height limit early warning method based on binocular stereo vision. Background technique [0002] With the development of the economy, urban transportation is also booming. On some roads in urban traffic, in order to restrict the passage of super-high vehicles, height-limiting obstacles are often added to avoid accidents such as over-high vehicles getting stuck on bridges and crashing into bridges. occurrence, to ensure the normal passage of the road. However, such traffic accidents still occur from time to time. At present, there is still no effective device for accurately detecting the height-limiting obstacles in front of the vehicle. Accidents, but also prone to driver negligence. [0003] For example, the application publication number is CN105915846A, and the Chinese patent titled "a method and system for monitoring intrusions with single and binocular multiplexing...

AI Technical Summary

Problems solved by technology

Although this technical solution discloses a method for binocular vision detection of intruders, it can only detect the presence of intruders in video images, and in the complex environment of urban road traffic, this solution can only detect the presence of intruders on the road. There are obstacles ahead, but it cannot be used to detect height-limited obstacles while the vehicle is driving

Images