RGB-D Camera Pose Tracking for Efficient 3D Model Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing camera tracking technologies, such as DVO SLAM and ORB-SLAM, suffer from inaccurate camera pose estimation and excessive computational burdens, necessitating a more efficient and accurate method for generating 3D models using RGB-D cameras.
Innovation Solution
A method and device for generating 3D models by determining initial and final camera poses through RGB-D data, selecting key vertexes, and generating 3D models using key-frames, which involves minimizing distance and intensity differences between frames, and utilizing a processor to perform these operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing camera tracking technologies (DVO SLAM or ORB-SLAM) are used, then camera tracking functionality is provided, but measurement precision of camera pose estimation deteriorates
Solution Approach 1:
The patent segments the camera tracking process into two distinct phases: coarse tracking using ORB-SLAM for initial pose estimation, and fine tracking using DVO for precise pose refinement. This segmentation allows each method to operate in its optimal performance range, with ORB-SLAM providing robust initial estimates and DVO delivering high-precision adjustments, thereby resolving the contradiction between measurement precision and reliability
Solution Approach 2:
The patent merges two different camera tracking methodologies (ORB-SLAM and DVO) into a unified hybrid system. The ORB-SLAM component provides global localization and mapping capabilities, while the DVO component delivers high-frequency precise pose estimation. By combining these complementary approaches, the system achieves both accurate measurement and reliable tracking performance that neither method could achieve alone
2Productivity
If existing camera tracking technologies (DVO SLAM or ORB-SLAM) are used, then camera tracking functionality is provided, but device complexity increases due to excessive calculations
Solution Approach 1:
The patent applies partial action by using ORB-SLAM only for initial coarse tracking and keyframe selection, rather than continuously running it at full computational intensity. The computationally intensive DVO is then applied selectively for fine tracking on selected frames. This partial application of each method reduces overall computational burden while maintaining tracking efficiency
Solution Approach 2:
The system implements periodic action by alternating between ORB-SLAM-based coarse tracking and DVO-based fine tracking in a structured sequence. ORB-SLAM operates periodically to provide global context and select keyframes, while DVO operates in between to refine pose estimates. This periodic alternation optimizes computational resource utilization and reduces overall system complexity
Data Source
AI summary
Disclosed are a method and device for generating a three-dimensional (3D) model through RGB-D camera tracking. A 3D model generation method may comprise the steps of: identifying RGB images and depth images; determining an initial camera pose of an RGB-D camera for each original frame of the depth images and the RGB images; determining a final camera pose for each of the original frames by updating the initial camera pose by using differences between an original frame, transformed according to the initial camera pose, and remaining original frame among two consecutive original frames of the depth images and the RGB images; selecting key vertexes capable of covering all original vertexes among original vertexes; and generating a 3D model by using key-frames, corresponding to the key vertexes in the original frames of the depth images and the RGB images, and final camera poses corresponding to the key-frames.