A monocular vision vehicle distance measurement method using the Internet of Vehicles to obtain the characteristics of the vehicle in front

Abstract

The invention relates to a monocular vision vehicle range finding method for obtaining features of a vehicle ahead via Internet of Vehicles. The method comprises the following steps: Internet of Vehicles information of an object vehicle is obtained; image information of the object vehicle is obtained via a camera; information of spatial relations among characteristic points of the vehicle ahead is obtained based on the Internet of Vehicles information, a space model is built based on the information of spatial relations and a position of each feature point of the vehicle ahead in the image information, a position of the vehicle ahead is estimated based on the space model, a space rectangular coordinate system is built, positions of feature points of the vehicle ahead on a two dimensional image are converted into parameters for solving actual coordinates of the vehicle, the position of the vehicle ahead is preliminarily estimated, distance capable of representing the position of the vehicle ahead and all feature points of the vehicle ahead is calculated, difference between an estimated value of the vehicle ahead and a true value is calculated, and a value of difference is lowered via adjustment of a camera estimated position; whether the different is minimum difference is determined; if yes, relation of relative positions of the vehicle ahead and a following vehicle can be determined according to the position of the camera; if no, the camera estimated position is re-chosen, and re-estimation is conducted.

Description

technical field [0001] The invention relates to a vehicle ranging method, in particular to a monocular vision vehicle ranging method applied in the field of computer vision vehicle positioning, which uses the Internet of Vehicles to obtain the characteristics of the preceding vehicle. Background technique [0002] The intelligent development of vehicles is inseparable from environmental perception. The accuracy and stability of environmental perception have an important impact on vehicle safety, and the use of monocular vision to estimate the position of vehicles is an important basis for environmental perception. The process of monocular vision distance measurement needs to obtain the realistic scale of image information and actual size conversion, and the accuracy of this scale has a decisive influence on the result. Most of the products and methods used in distance measurement, including companies such as Mobileye, are based on the assumption that the vehicles are on the ...

AI Technical Summary

Problems solved by technology

Most of the products and methods used in distance measurement, including companies such as Mobileye, are based on the assumption that the vehicles are on the same plane, mainly using the height of the camera to achieve scale conversion, but this method is not suitable for complex road conditions such as bumpy roads or slopes. There will be a large error, and there is still room for improvement in the overall adaptability and accuracy

Images