Multi-Camera Calibration Using Skeleton Landmarks for 3D Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for three-dimensional shape estimation from a small number of camera images are inaccurate due to the difficulty in obtaining camera parameters without precise calibration, especially in environments where cameras have few common visual fields and cannot rely on natural features or fixed patterns.
Innovation Solution
A method for camera calibration using skeleton estimation of a target object, such as a person, to determine camera parameters based on synchronized image capturing, allowing three-dimensional reconstruction even with a small number of cameras, without the need for fixed patterns or extensive visual overlap.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional multi-view stereo or visual hull methods are used for three-dimensional shape estimation, then mathematical formalization of projection process is achieved, but many cameras with precise adjustment are required and recessed shapes cannot be handled
Solution Approach 1:
The patent introduces a predetermined part (such as a landmark or feature point) as an intermediary element that facilitates camera parameter estimation. By detecting the position of this predetermined part in captured images, the system can estimate camera parameters without requiring complex multi-camera setups or precise manual adjustment, thus resolving the contradiction between measurement precision and device complexity
Solution Approach 2:
The patent changes the approach from direct three-dimensional shape estimation to a two-stage process: first estimating camera parameters by detecting predetermined part positions, then performing three-dimensional reconstruction using these estimated parameters. This parameter change enables accurate shape estimation with fewer cameras by decoupling the camera calibration step from the reconstruction step
2Measurement precision
If deep learning model is used for three-dimensional reconstruction, then reconstruction quality improves, but camera parameter accuracy becomes critical and synchronized image capturing by great number of cameras is required
Solution Approach 1:
The patent performs preliminary estimation of camera parameters by detecting the position of predetermined parts in images before performing the actual three-dimensional reconstruction. This preliminary action provides accurate initial camera parameters that enable high-quality reconstruction without requiring synchronized capturing by numerous cameras, thus improving productivity while maintaining precision
3Ease of operation
If camera parameter estimation is performed without precise calibration, then manual calibration burden is reduced, but reconstruction accuracy deteriorates in environments with few common visual fields
Solution Approach 1:
The patent enables the system to automatically estimate camera parameters by detecting predetermined part positions in captured images, making the system self-calibrating without requiring manual calibration operations. This self-service approach maintains reconstruction accuracy even in environments with few common visual fields by using the predetermined part as a reliable reference that does not depend on visual field overlap
Data Source
AI summary
An information processing apparatus includes: an obtainment unit configured to obtain a captured image obtained by each of multiple image capturing apparatuses; a detection unit configured to detect a position of a predetermined part in the object from the captured image of each of the multiple image capturing apparatuses; an estimation unit configured to estimate a camera parameter indicating a position and an orientation of each of the multiple image capturing apparatuses by using the detected position of the predetermined part; an update unit configured to update the camera parameter of each of the multiple image capturing apparatuses by using the estimated camera parameter as an initial value; and a determination unit configured to determine the camera parameter of each of the multiple image capturing apparatuses based on a result of performing three-dimensional reconstruction of the object based on the updated camera parameter.