Vehicle-mounted binocular camera calibration and parameter verification method

Abstract

The invention discloses a vehicle-mounted binocular camera calibration and parameter verification method. The method includes a calibration stage and a verification stage, wherein the verification stage includes static verification and dynamic verification. According to the static verification, standing conical barrel markers before an automobile are adopted; internal and external parameters of a camera obtained in a binocular ranging formula verification calibration stage are utilized; the conical barrels are distributed at view regions of the imaging of the camera; and the relative error rate of the view regions is tested. According to the method of the invention, based on binocular camera ranging principles, a vehicle-mounted binocular camera ranging model for an outdoor real scene is constructed, and an operation basis can be provided for the reconstruction of a real three-dimensional scene in the binocular camera visual processing field. Vehicle-mounted binocular cameras are one of key research directions in realizing intelligent driving; and intelligent driving usually adopts a high-precision radar to obtain high-quality perception effects, while, cost is also very high, and the vehicle-mounted binocular camera calibration and parameter verification method of the invention can be matched with an ordinary radar and camera so as to improve vehicle intelligent identification accuracy to a level within 1 m.

Description

technical field [0001] The invention relates to the field of visual processing, in particular to a method for calibrating and verifying parameters of a vehicle-mounted binocular camera. Background technique [0002] In image measurement and machine vision applications, there is a one-to-one mapping relationship between the three-dimensional geometric position of objects in space and the corresponding pixels of camera images, and binocular vision models are usually used for measurement. Establishing the camera imaging model is to solve the internal and external parameters of the camera. The process of solving for these parameters is called camera calibration. Zhang Zhengyou, Brown and others researched the camera calibration algorithm, making the camera calibration operation process simple and convenient. But their experiments were all in a laboratory setting. In industrial applications, due to the technical defects in the structural devices, the working environment, place...

AI Technical Summary

Problems solved by technology

Strictly speaking, the system must be recalibrated every time an impact occurs, which is not easy to achieve in industrial applications

Images