Vehicle positioning method based on road surface features and application in visual positioning

Abstract

The invention provides a vehicle positioning method based on road surface features. The vehicle positioning method comprises the following operation steps: A) performing equipment preparation: installing a high-speed camera for ground-oriented shooting on a vehicle chassis; b) continuously inputting pavement images; c) calculating vehicle displacement through a machine vision algorithm: determining feature pixel points on the input image, performing feature recognition and matching on the feature pixel set of the road image through the machine vision algorithm, and calculating the displacementof the feature pixel points under the image scale; obtaining the actual positions of the pixel points according to the calibration conversion relation from the image scale to the actual scale, and obtaining the actual displacement of the vehicle according to the relative motion principle; D) finishing. According to the method, the ground image of the position where the vehicle passes through is collected in real time through the high-speed camera, the vehicle displacement corresponding to two continuous frames of images is estimated through a feature recognition and matching algorithm, the more accurate vehicle position is obtained through fusion with IMU measurement data, and higher vehicle positioning precision is obtained.

Description

technical field [0001] The invention belongs to the technical field of vehicle positioning and machine vision, and in particular relates to a vehicle positioning method based on road surface features and its application in visual positioning. Pavement, marble pavement, paved pavement, etc., which have image features extractable for current machine vision algorithms. Background technique [0002] At present, the on-board navigation chips used in passenger cars, commercial vehicles and other fields generally integrate MEMS (Micro-Electro-Mechanical Systems) inertial navigation unit IMU (Inertial Measurement Unit) to ensure that in the case of poor GNSS (Global Navigation Satellite System) signal The vehicle position solution can be obtained continuously. At the same time, parameters such as wheel speed and steering angle of the vehicle are often integrated as the input of the vehicle odometer model to obtain an estimate of the vehicle position. Specifically, under the scheme...

Images