Binocular depth perception-based vehicle travelable area calculation method

Abstract

The present invention provides a binocular depth perception-based vehicle travelable area calculation method. According to the method, the categories of key targets and the movement behaviors of the key targets are obtained through detection; based on prior knowledge in some vehicle driving rules, mutual influence rules between the driving of related targets and the driving of a local vehicle are defined; the driving cost of each occupancy grid in a scene relative to the local vehicle is calculated according to the influence rules; and the travelable area of the vehicles is calculated to express a traffic scene with the local vehicle adopted as a center. With the method of the invention adopted, accurate identification of key targets such as pedestrians and vehicles in a traffic scene can be realized, and the movement trajectories of the key targets are predicted so as to be applied to behavior analysis. The binocular depth perception-based vehicle travelable area calculation method has high accuracy and high feasibility and can be applied to unmanned driving, driving assistance systems and the like in the intelligent transportation field.

Description

technical field [0001] The invention belongs to the technical field of video intelligent analysis, in particular to a method for calculating a drivable area of ​​a vehicle based on binocular depth perception. Background technique [0002] In the research on the calculation model of the driving area of ​​intelligent transportation unmanned vehicles, the current mainstream is the model for calculating the safe driving area based on the spatial position information of the vehicle's forward target, which mainly uses binocular vision depth matching technology or radar detection. The technology calculates the distance between the forward object and the own vehicle, and judges the obstacle position information, so as to calculate the safe driving area range of the own vehicle in the traffic scene. [0003] Although this model can obtain a relatively accurate drivable area in space through the calculation of depth information for static objects, in a more complex traffic scene, it d...

AI Technical Summary

Problems solved by technology

[0003] Although this model can obtain a relatively accurate drivable area in space through the calculation of depth information for static objects, in a more complex traffic scene, it does not take into account the motion information of other moving objects in the scene and the motion of the vehicle itself. Trajectory, when a vehicle in the side lane changes lanes at a higher speed or an obstacle that is completely difficult to perceive based on static analysis of the scene appears on the road ahead (blockage in the distance, impassable road, poor surface driving environment, etc.), This model is difficult to estimate and analyze in advance and re-plan the path, which makes it difficult for the control planning of the own vehicle

Images