Integrated safe driving recorder system and method

Abstract

The invention discloses an integrated safe driving recorder system and method. According to a key point, an original image is converted into an overlooking image consistent with visual habits of a driver, the space around a vehicle body is provided, an actual distance proportion is presented, the corresponding overlooking image comprises lane line images, characteristics for identification can be acquired from the corresponding overlooking image, the advancing trajectory of a vehicle and the trajectories of the lane line images are calculated through a lane line deviation sensing unit to judge whether the advancing trajectory of the vehicle is close to the trajectories of the lane line images or not, a warning signal is sent to prompt the driver to pay attention to reduce accidents when the trajectories of the lane line images are close to the advancing trajectory of the vehicle, and the original image can be recorded as a basis for accident identification through an instant video signal coding / decoding unit after an accident happens.

Description

technical field [0001] The present invention relates to an integrated safety driving recorder system and its method, especially to a system capable of judging whether the vehicle track is close to the lane line image track, and when the lane line image track is close to the vehicle track, it has the function of issuing a warning The effect of the signal reminding the driver to pay attention to reduce the occurrence of accidents is applicable to the surrounding view driving recorder, driving recorder or similar devices. Background technique [0002] In recent years, the bird's-eye view system has been widely used in passive parking assistance systems to provide the driver with an understanding of the space around the vehicle and the real distance. The bird's-eye view image is converted, and finally the scene is combined. [0003] In addition, the current image-based lane detection technology can be divided into two types, the first is the original image direct detection lane...

AI Technical Summary

Problems solved by technology

[0003] In addition, the current image-based lane detection technology can be divided into two types, the first is the original image direct detection lane line technology, the second is the plane image detection lane line technology, and the first original image directly detects the lane The line technology uses the Hough transform to find the slope change of the lane line and the vanishing point of the lane line, etc., but the main disadvantage of the first type is that the slope of the lane line is difficult to define the lane offset. Image detection lane line technology can be divided into anti-perspective conversion technology and plane conversion technology, and the anti-perspective conversion technology uses the principle of perspective projection to correct multiple original images through the camera to obtain the internal and external coordinates of the camera in the world coordinates. External parameters, and then inversely transformed from the original image to obtain a bird's-eye view image, but the inverse perspective transformation technology is only applicable to the transformed plane image with a small transformation angle and a large field of view; while the other plane transformation technology uses the ground with known corresponding relationship Features and definition of a virtual image plane, and then use the corresponding features of the image plane and the virtual image plane to obtain the plane projection transformation matrix. This plane transformation technology is suitable for the conversion plane image with large conversion angle and small field of view

Images