3D Model Reconstruction Using Correlation-Based Image Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to efficiently reconstruct elaborate and reliable three-dimensional models from images captured by image sensors, particularly in mixed reality systems like augmented and virtual reality, due to challenges in determining accurate pose information and handling large and small camera motions.
Innovation Solution
A method and data processing device that utilize color and depth data to determine a registration order for input images, generating pose information sequentially, and reconstruct a three-dimensional model by registering images based on correlation values, enhancing accuracy and robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods are used to reconstruct three-dimensional models from images, then the reconstruction process can be performed, but the accuracy and reliability of the three-dimensional model is insufficient
Solution Approach 1:
The patent segments the reconstruction process into distinct phases: determining registration order based on color data correlation, sequentially generating pose information, and progressively registering images. This segmentation allows each phase to be optimized independently, improving overall reconstruction accuracy and reliability
Solution Approach 2:
The patent performs preliminary determination of registration order using color data correlation before actual three-dimensional reconstruction. By pre-establishing the optimal sequence of image registration based on correlation values, the system ensures that images are processed in the most effective order, enhancing both accuracy and reliability of the final model
2Measurement precision
If images are registered without determining a specific order, then the processing can be simplified, but the accuracy of pose information generation deteriorates
Solution Approach 1:
The patent changes the parameter of image registration from arbitrary or simultaneous processing to sequential processing based on correlation values. By using correlation strength as the ordering parameter, the system achieves higher pose information accuracy while maintaining manageable process complexity through automated correlation-based sequencing
3Reliability
If all input images are processed equally, then the processing method is simple, but noisy images degrade the quality of the three-dimensional model
Solution Approach 1:
The patent applies local quality assessment by evaluating each image's correlation value with other images in the sequence. Images with low correlation (potentially noisy or irrelevant) are automatically identified and handled differently during the sequential registration process, allowing the system to maintain high model quality without requiring complex manual filtering
Solution Approach 2:
The patent incorporates feedback through correlation value calculation, where the registration order is determined based on measured correlation between images. This feedback mechanism automatically adjusts the processing sequence to optimize model quality, with higher correlation images processed earlier to establish a more reliable reconstruction foundation
Data Source
AI summary
Input images are obtained to include color data and depth data of a target. An image graph is generated based on the color data of the input images such that the image graph includes correlation values between the input images. A registration order is determined based on the image graph. Pose information of each input image is sequentially generated based on the registration order and the depth data of the input images. The input images are sequentially registered as registered images such that each registered image includes the pose information. A three-dimensional model of the target is reconstructed based on the registered images. The pose information may be generated and the elaborate three-dimensional model may be reconstructed by determining the registration order based on the color data and sequentially generating the pose information of the input images according to the registration order using the registered images.