CT-Based Textured Surface Reconstruction for Complex Objects
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for generating textured surfaces of physical objects, especially those with complex structures or visual occlusions, fail to produce accurate 3D surface representations and texture data due to calibration issues and reliance on visible light data, which are inadequate for materials like transparent or shiny surfaces.
Innovation Solution
A multimodal X-ray system combining a CT scanner with optical cameras uses a phantom for accurate calibration, enabling precise alignment of camera positions relative to the scanner, and generates textured surfaces by projecting optical image data onto volumetric data reconstructed from CT scans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If optical cameras and photogrammetry techniques are used to generate textured surfaces, then texture data can be captured, but accurate surface representation fails for objects with visual occlusions, transparent surfaces, shiny surfaces, or complex structures
Solution Approach 1:
The patent combines X-ray CT scanning technology with optical camera systems to create a hybrid imaging system. The X-ray component penetrates opaque and transparent materials equally, while the optical component captures surface texture and color information. By merging these two modalities and fusing their data, the system overcomes the limitations of each individual technique and achieves accurate surface representation for complex objects including those with transparent, shiny, or occluded surfaces.
Solution Approach 2:
The patent introduces a specialized calibration phantom as an intermediary object that bridges the X-ray and optical domains. The phantom contains both X-ray detectable features (for calibration in the X-ray domain) and optically detectable features (for calibration in the optical domain). This intermediary enables precise registration and alignment between the two different imaging modalities, solving the calibration challenge of multi-modal systems.
2Measurement precision
If photogrammetry techniques are used for camera calibration relative to CT scanner, then camera positions can be located, but the process becomes computationally intensive and requires multiple images and repeated computations for each camera
Solution Approach 1:
The patent performs camera calibration using a dedicated phantom with known geometric features before actual object scanning. The calibration phantom contains precisely manufactured features with known positions and orientations that allow pre-computation of transformation matrices. This preliminary calibration step establishes the geometric relationship between the optical camera and X-ray scanner coordinate systems, enabling efficient processing during actual object scanning without requiring computationally intensive photogrammetry during production scanning.
Solution Approach 2:
The patent uses a calibration phantom that serves as a known reference copy with precisely defined geometric features. Instead of performing complex computations on arbitrary object surfaces, the system uses the phantom's known geometry as a template for calibration. The phantom replicates the coordinate transformation relationship between optical and X-ray domains, allowing efficient determination of camera positions and orientations without repeated photogrammetry computations.
3Measurement precision
If small angular errors occur in camera calibration, then calibration is slightly off, but large translation shifts occur at the object surface due to abbe errors
Solution Approach 1:
The calibration phantom acts as an intermediary reference that directly connects the optical camera and X-ray scanner coordinate systems at a known location. By using the phantom's precisely known geometry as the mediator, the system establishes an accurate transformation relationship that minimizes the lever arm for angular errors. This reduces the magnification of calibration errors at the object surface, preventing large translation shifts that would otherwise occur due to abbe errors in direct alignment methods.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach produces accurate 3D surface meshes with added texture data, overcoming calibration challenges and achieving superior surface accuracy for complex objects, including those with transparent, shiny, or overlapping surfaces, without requiring computationally intensive photogrammetry.
Implementation Method 1
two-dimensional radiographs of the physical object acquired using an X-ray source in a computed tomography scanner
Implementation Method 2
adding texture data to the three-dimensional surfaces represented in the computer data structure using images captured by at least one optical camera
Data Source
AI summary
Provided herein are methods, apparatuses, computer program products, and systems for generating textured surfaces using computed tomography. One method can include generating a computer data structure representing three-dimensional surfaces of a physical object from (i) segmentation of volumetric data reconstructed from radiographs, (ii) position information and orientation information of the physical object, and (iii) size information of the physical object; adding texture data to the three-dimensional surfaces represented in the computer data structure using images captured by at least one optical camera, the images being of the physical object at different rotational orientations, by projecting data of the images onto the three-dimensional surfaces using the size information, the position information, and the orientation information of the physical object and predetermined position, location, and parameters of the at least one optical camera; and providing the computer data structure for processing of the three-dimensional surfaces.