Acquisition method, calibration method and system for image coordinates of invisible position of camera

Abstract

The invention relates to an acquisition method, a calibration method and a system for an image coordinate of an invisible position of a camera. The acquisition method comprises the following steps: acquiring a first color image of a to-be-measured position invisible to a main camera and a corresponding depth image from a depth camera, wherein the depth camera is positioned to enable the to-be-measured position to be located in a view field range of the depth camera; extracting three-dimensional coordinates of the position to be measured from the first color image and the corresponding depth image; and converting the three-dimensional coordinates of the position to be measured into image coordinates of the position to be measured through a translation matrix and a rotation matrix which are used for converting the coordinate system of the depth camera into the coordinate system of the main camera, wherein the translation matrix and the rotation matrix are determined through calibration between the main camera and the depth camera. By means of the method, the image coordinates of the position outside the view field of the camera can be obtained in a simple mode, and errors are small.

Description

technical field [0001] The present application relates to the technical fields of image measurement and machine vision, and in particular to a method for acquiring pixel coordinates of an invisible position to be measured, a calibration method and a system. Background technique [0002] With the development of automatic driving, cameras (for example, IR cameras or RGB cameras, etc.) need to be installed on vehicles to provide assistance for the realization of automatic driving. However, due to the limited field of view of the camera, it is impossible to use a single camera to obtain its invisible position (that is, the position that is not within the field of view of the camera, such as the vehicle's rearview mirror, front window glass, instrument panel, etc.) etc.) in the camera coordinate system, but in the actual driving process, the image coordinate information about these positions may be very important for automatic driving. Currently, multiple auxiliary cameras or mi...

AI Technical Summary

Problems solved by technology

However, due to the limited field of view of the camera, it is impossible to use a single camera to obtain its invisible position (that is, the position that is not within the field of view of the camera, such as the vehicle's rearview mirror, front window glass, instrument panel, etc.) etc.) image coordinates in the camera coordinate system, but in the actual driving process, the image coordinate information about these positions may be very important for automatic driving

Currently, multiple auxiliary cameras or mirrors are usually used to help obtain such image coordinates, but this is more complicated to implement and can easily cause large errors

Images