Image calibration method based on RGBD depth camera device

Abstract

The invention relates to an image calibration method based on RGBD depth camera devices, which is used for gait acquisition and analysis, and comprises the following steps: acquiring images of a humanbody by using an imaging device, selecting feature points corresponding to the images acquired by the RGBD depth camera devices at the same moment by using a GMS algorithm, and pairing the feature points; fitting a rotation and translation matrix by utilizing the paired feature points and a general analysis method; and performing rotational translation calculation on all points acquired by the target camera device by using the rotational translation matrix, and performing registration by using an ICP algorithm to obtain a fine rotational translation matrix. By means of the method, the situation that gait information cannot be completely and accurately collected due to the factors such as the side body of a subject and the arrangement position of the camera device can be avoided, and compared with the prior art, the method can meet the real-time requirement, can guarantee precision, and is high in efficiency and short in consumed time.

Description

Technical field [0001] The present invention relates to the field of image processing, and more specifically, to an image calibration method based on an RGBD depth camera device, a computer storage medium, and computer equipment. Background technique [0002] Gait is a periodic phenomenon that describes the characteristics of human or animal walking, and each cycle can be divided into multiple parts for analysis. Gait analysis is a process of using the concept of biomechanics to calculate gait time, space, dynamics and other gait parameters through the gait data collected by the human body during walking, and to research and analyze gait characteristics. With the development of computer technology and human biomechanics, in recent years, gait analysis has received more and more attention, both in clinical and scientific research, such as movement recognition, rehabilitation evaluation, disease monitoring and diagnosis, and artificial The design of joints and assistive devices ha...

AI Technical Summary

Problems solved by technology

The problem with using RGBD camera devices for gait acquisition and analysis is that a single RGBD camera device has a limited field of view due to its small viewing angle. During the 3D reconstruction process, it is necessary to move the camera to scan from multiple angles to reconstruct a complete human body or scene model. Synchronization of multiple RGBD cameras Acquisition becomes the best 3D reconstruction technology solution

However, the 3D point cloud collected by multiple RGBD cameras needs to be registered and fused before 3D reconstruction can be performed. The registration process of 3D point cloud requires a large amount of calculation and takes a long time. The calculation time will be further extended as the accuracy requirements increase, and The registration effect depends on the initialization parameters of the registration algorithm

Therefore, many researchers first use the color camera device in the RGBD camera device to calibrate the camera device, and then further calibrate the depth information (3D point cloud) based on the calibration results of the color camera device. This method can significantly reduce the calculation amount of point cloud registration and Calculation time, but this method needs to place camera calibration aids during the registration process, such as T-frame or black and white checkerboard, so that each camera can completely capture multiple frames of camera calibration aids from different angles

This method is complicated to operate and requires many auxiliary devices, so the practicability is not high

Images