Vision-based automatic detection method for front vehicle

Abstract

The invention discloses a vision-based automatic detection method for a front vehicle, which comprises the following steps of: acquiring an image sequence by a vehicle-mounted camera; before detection processing, performing calibration operation of drawing a calibration line on images to be detected, and manually setting a detection region; detecting by using vehicle bottom shadow features, plate number features or taillight features, and judging detected vehicles; during detection processing according to frames in sequence, performing snapshotting on traced vehicles. The mounting position of the camera is not limited any more due to the calibration operation, the camera can be arranged at any position in front of the vehicle, the operation is convenient, the applicable range is wide, and flexible mounting, use, and replacement are facilitated; part of unnecessary subsequent judgments are reduced by same vehicle judgment, operation time is saved, misdetection is effectively prevented, and the method has the advantages of high detection precision, high detection accuracy, small calculated amount, low processing cost and the like.

Description

technical field [0001] The invention relates to a detection method, in particular to a vehicle-mounted automatic detection method for a vehicle ahead based on vision, and belongs to the technical field of intelligent transportation. Background technique [0002] At present, the method of detecting the vehicle in front is generally to set a fixed vehicle detection area and detection threshold, which does not have enough discrimination between the vehicle and the fake car, and is processed for the case of a fixed camera, which not only limits the installation position of the camera but also The installation location is fixed and inconvenient to change; at the same time, the detection area and detection threshold are divided according to the lane lines detected in the image, which requires a large amount of calculation. [0003] And, the existing processing method usually uses various features to detect each frame when detecting a vehicle, so it cannot take into account process...

AI Technical Summary

Problems solved by technology

[0002] At present, the method of detecting the vehicle in front is generally to set a fixed vehicle detection area and detection threshold, which does not have enough discrimination between the vehicle and the fake car, and is processed for the case of a fixed camera, which not only limits the installation position of the camera but also The installation position is fixed and it is inconvenient to change; at the same time, the detection area and detection threshold are divided according to the lane lines detected in the image, which requires a large amount of calculation

[0003] And, the existing processing methods usually use various features to detect each frame when detecting vehicles, so it is impossible to take into account processing speed, detection accuracy, and image processing cost

Images