Depth camera automatic calibration algorithm based on three-dimensional feature points

Abstract

The invention provides a depth camera automatic calibration algorithm based on three-dimensional feature points. According to the calibration algorithm, calibration work of a depth camera can be completed at a time by using two depth images and corresponding three-dimensional feature pixel points. The method comprises the following steps: firstly, combining a depth camera measurement model and a classical camera calibration model according to known coordinates of characteristic pixel points in a three-dimensional space, coordinates of corresponding points in a depth image and an original erroroffset to obtain a basic calibration model of a depth camera; secondly, obtaining an initial value of an internal parameter of the camera through an internal parameter constraint condition of the depth camera, and obtaining an external parameter according to the initial value of the internal parameter; and finally, optimizing the established depth image error control function to obtain an optimalparameter of the depth camera. According to the calibration algorithm, a traditional calibration model of the depth camera is simplified, the defect that classic calibration is affected by illumination conditions, scene uncertainty and the like is overcome, internal and external parameters of the camera are accurately obtained, a good restoration effect on edge distortion of the depth image is achieved, and the calibration algorithm has the advantages of being easy to operate, high in calibration precision, good in practicability and robustness and the like.

Description

technical field [0001] The invention belongs to the technical field of machine vision calibration, and relates to a depth camera calibration algorithm, in particular to an automatic depth camera calibration algorithm based on three-dimensional feature points Background technique [0002] With the development and progress of science and technology, machine vision is no longer an unfamiliar term when artificial intelligence technology pervades all major fields. Because the machine vision system has many advantages such as high precision, high efficiency, good practicability, safety and reliability, and low cost, the demand for depth vision in many industries is becoming more and more prominent, whether it is in scientific research institutions or in enterprise production. pivotal position. For example, in the field of industrial manufacturing, visual inspection, route planning, tracking and positioning, etc.; in life, face recognition technology, AR technology, holographic pr...

AI Technical Summary

Problems solved by technology

However, these methods can only be applied to ordinary optical cameras, and have the disadvantages of being easily affected by the environment, complicated calibration steps, time-consuming, labor-intensive, high calibration costs, and complicated solution processes.

The imaging methods of depth cameras are complex and diverse, so that scholars at home and abroad have not studied depth camera calibration algorithms deeply. Most of their algorithms have complex structures, low calibration accuracy, time-consuming and labor-intensive

Images