Single Camera Photogrammetry Structure From Motion 3D Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing photogrammetry methods require cumbersome processes, such as using two cameras or manual calibration steps, to obtain accurate 3D digital representations of objects, especially when measuring distant objects, which limits their usability and accuracy.
Innovation Solution
A method using a single image-capture device, such as a smartphone or tablet, that processes overlapping 2D digital images with a structure from motion algorithm to generate 3D digital representations, including 3D models, point clouds, and edge clouds, without the need for active sensors or manual calibration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional stereo/2D or 3D cameras are used to obtain two images simultaneously from two viewpoints, then measurements can be obtained, but the parallax is insufficient for high-quality measurement of distant objects
Solution Approach 1:
The patent transitions from static two-viewpoint photography to dynamic sequential photography, where the camera moves between multiple positions to capture images of the same object. This dynamic approach creates sufficient parallax for distant objects while using a single camera, eliminating the need for complex multi-camera setups.
Solution Approach 2:
The patent introduces a template or framework as an intermediary reference object with known dimensions. This template serves as a mediator that provides scale reference and geometric constraints, enabling accurate measurements from single-camera sequential images without requiring complex calibration procedures.
2Measurement precision
If a template or framework is used to guide orientation of the image-capture device, then accurate visualization can be obtained, but the process is cumbersome and prone to user error
Solution Approach 1:
The patent enables the image-capture device to automatically perform calibration and measurement functions without requiring manual template placement or complex user intervention. The system self-calibrates by capturing images of the template at different orientations and automatically computes the necessary geometric parameters, reducing user error and simplifying operation.
Solution Approach 2:
The patent performs preliminary calibration by capturing multiple images of the template framework at different orientations before actual measurement. This preliminary action establishes the geometric relationships and camera parameters in advance, so that subsequent measurements can be performed automatically without repeated manual calibration.
3Measurement precision
If multiple images are processed using structure from motion algorithms, then accurate 3D digital representations can be generated, but processing time is increased
Solution Approach 1:
The patent performs preliminary feature detection, matching, and camera pose estimation during the image capture phase itself. By pre-processing the images and extracting key geometric features before the actual 3D reconstruction, the system reduces the computational burden during final processing, maintaining accuracy while reducing processing time.
Solution Approach 2:
The patent segments the 3D reconstruction process into distinct stages: feature detection, matching, camera pose estimation, and final model generation. By dividing the processing into manageable segments and performing optimizations at each stage, the system achieves accurate results more efficiently than monolithic processing approaches.
Data Source
AI summary
The inventions herein relate generally to improvements in photogrammetry and devices suitable for obtaining such improvements. Some embodiments use only a single passive image-capture device to obtain overlapping 2D images, where such images at least partially overlap with regard to at least one object of interest in a scene. Such images can be processed using methods incorporating structure from motion algorithms. Accurate 3D digital representations of the at least one object of interest can be obtained. Substantially accurate measurements and other useful information regarding the at least one object of interest are obtainable from the methodology herein.