3D Object Alignment via Stereo Vision and Geometric Shape Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current machine vision systems face challenges in accurately and efficiently determining the three-dimensional alignment of objects due to the loss of depth information when converting 3D scenes to 2D images, leading to inefficiencies in processing and alignment accuracy, especially in complex geometries and industrial settings.
Innovation Solution
A system and method that utilize stereo camera heads and 3D sensors to generate 3D point clouds, extract higher-level geometric shapes, and employ a scoring process to determine the alignment of objects in six degrees of freedom by correlating these shapes with a 3D model, reducing processing overhead and increasing speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If two-dimensional images are used for alignment, then device complexity is reduced, but measurement precision deteriorates due to loss of depth information
Solution Approach 1:
The patent transitions from 2D image processing to 3D point cloud processing by introducing depth information through multiple camera views. The system captures images from multiple angles and reconstructs 3D coordinates, adding the z-dimension to the traditional x-y plane analysis, thereby resolving the depth information loss inherent in 2D imaging.
Solution Approach 2:
The patent introduces 3D point clouds as an intermediary representation between the captured 2D images and the final alignment determination. The point cloud serves as a mediator that preserves depth information from multiple views while enabling comprehensive 3D spatial analysis, bridging the gap between simple 2D imaging and complex 3D reconstruction.
2Measurement precision
If three-dimensional sensors are used to capture depth data, then measurement precision is improved, but processing time increases
Solution Approach 1:
The patent segments the 3D point cloud into distinct geometric features (planes, cylinders, cones, spheres) for independent analysis. By dividing the complex point cloud data into manageable feature components, the system can process each feature type with specialized algorithms, reducing overall processing time while maintaining high precision in alignment determination.
Solution Approach 2:
The patent performs preliminary classification and identification of geometric features before executing the full alignment calculation. By pre-identifying the types of geometric primitives present in the point cloud and their approximate locations, the system prepares data structures and selects appropriate algorithms in advance, significantly reducing the time required for the final precise alignment computation.
3Measurement precision
If complex geometric shapes are analyzed in 3D, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent transforms the complex 3D geometric analysis problem into a series of simpler parameter-based operations. By representing complex shapes as combinations of basic geometric primitives with defined parameters (plane normals, cylinder radii, cone angles, sphere centers), the system changes the problem parameters from full 3D surface analysis to discrete feature parameter extraction and matching, reducing processing complexity while maintaining precision.
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
The solution enables robust, efficient, and accurate 3D alignment of complex objects, improving alignment speed and accuracy in industrial settings by leveraging 3D sensors and geometric shape extraction to handle complex geometries and shading patterns.
Implementation Method 1
One technique measures the delay of time between transmission of a light pulse and receipt of the reflected light pulse-a technique called Light Detection and Ranging or LIDAR
Implementation Method 2
A particular depth data-determination technique employs triangulation. This technique locates a feature in the scene or on the object in two or more of the images respectively acquired from each of 2D cameras, and using the relative position of the feature in each of the images, triangulation is performed to recover the depth information for that feature
Data Source
Figure 1
Figure 2
Figure 3
AI summary
This invention provides a system and method for determining the three-dimensional alignment of a modeled object or scene. After calibration, a 3D (stereo) sensor system views the object to derive a runtime 3D representation of the scene containing the object. Rectified images from each stereo head are preprocessed to enhance their edge features. A stereo matching process is then performed on at least two (a pair) of the rectified preprocessed images at a time by locating a predetermined feature on a first image and then locating the same feature in the other image. 3D points are computed for each pair of cameras to derive a 3D point cloud. The 3D point cloud is generated by transforming the 3D points of each camera pair into the world 3D space from the world calibration. The amount of 3D data from the point cloud is reduced by extracting higher-level geometric shapes (HLGS), such as line segments. Found HLGS from runtime are corresponded to HLGS on the model to produce candidate 3D poses. A coarse scoring process prunes the number of poses. The remaining candidate poses are then subjected to a further more-refined scoring process. These surviving candidate poses are then verified by, for example, fitting found 3D or 2D points of the candidate poses to a larger set of corresponding three-dimensional or two-dimensional model points, whereby the closest match is the best refined three-dimensional pose.