Machine vision-based intelligent vehicle right-turning target detection method

Abstract

The invention provides a machine vision-based intelligent vehicle right-turning target detection method. During the right-turning target detection process of an intelligent vehicle, an intelligent vehicle right-turning model is established. Based on the relative distance change rate of a current vehicle and a previous vehicle, and the relative distance change rate of the current vehicle and a roadboundary, whether the current vehicle drives from a straight lane into a curve lane or not can be judged. The optimal trajectory planning is carried out on the intelligent vehicle judged to enter thecurve lane, and then an area where the intelligent vehicle is judged to pass through within a short period of time is predicted. Meanwhile, the passing area of the intelligent vehicle is simplified into a parallelogram for processing. An effective target is determined according to the principle of the minimum distance from the lane, and then the effective target is updated in time. The detectionrange of the intelligent vehicle is reduced, and the target detection speed and the target detection accuracy are improved.

Description

technical field [0001] The invention provides a machine vision-based detection method for a right-turning target of an intelligent vehicle, which belongs to the field of intelligent driving. Background technique [0002] As a sensor that can obtain all road information in front of the vehicle, machine vision has the advantages of wide detection range, complete road information, and telemetry, and is widely used in the research of vehicle detection systems. At present, when machine vision technology is used to detect obstacles and vehicles in front, it is generally completed on a straight road with good light during the day, and the detection accuracy of obstacles or vehicles in front is also high. However, there are few studies on the front vehicle detection for effective target recognition when turning. In this case, the detection accuracy of the smart car to the effective target is low or the effective target cannot be detected at all, which increases the danger of driving...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problems of slow data processing speed and low real-time performance of the current camera, the present invention proposes a machine vision-based smart car right-turning target detection method, which detects effective targets on the basis of predicting the driving trajectory of the vehicle, and reduces data acquisition time. range, improving driving safety

Images