3D Scan De-Skewing Using Planar Surface Angle Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
3D imaging systems based on light triangulation often introduce a skewing effect, causing scanned objects to appear skewed in 3D images, which conventional methods using point reference features struggle to accurately correct.
Innovation Solution
Determine a transform based on detected planar imaged surfaces with known angular relationships, applying it to align the angular relationships between imaged surfaces to match those of the real surfaces, eliminating the need for point reference features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional methods using point reference features are used to correct skewing, then de-skewing can be attempted, but accuracy is insufficient and algorithms become more complex
Solution Approach 1:
The patent changes the reference feature type from point-based to plane-based, utilizing the angular relationships between planar surfaces instead of point coordinates. This parameter change in the reference feature dimensionality improves measurement precision while simplifying the overall algorithm structure
Solution Approach 2:
Instead of using point reference features as conventionally done, the patent inverts the approach by using planar surface reference features with known angular relationships. This inversion fundamentally changes the reference system and improves both accuracy and simplicity
2Reliability
If point reference features are used for de-skewing, then correction can be performed, but scale and position invariance are lost and estimation accuracy decreases
Solution Approach 1:
The planar reference feature approach serves multiple functions simultaneously: it provides scale invariance, position invariance, and improved estimation accuracy. The angular relationships between planes are universally applicable regardless of scale or position, making the method more versatile and reliable
Solution Approach 2:
The patent transitions from 0-dimensional point features to 2-dimensional planar features, adding dimensional information that enables scale and position invariance. The angular relationships between planes provide additional geometric constraints that improve estimation reliability
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 provides robust, scale-invariant, and position-invariant de-skewing of 3D image data, simplifying algorithms and improving estimation accuracy compared to conventional methods.
Implementation Method 1
3D imaging based on light triangulation, where structured light, typically a light plane, or 'sheet of light', is used, and an object scanned through and/or by this light plane
Data Source
Figure 1
Figure 2A~2D
Figure 3A~3B
AI summary
Method and device(s) (605; 630; 640: 650; 900) for determining a transform having a de-skewing effect on 3D image data resulting from scanning by a 3D imaging system (605). It is obtained (701) 3D image data resulting from scanning by the 3D imaging system (605) of planar real surfaces (322a-c; 422a-c; 522a-c) that at least during the scanning were non-parallel with constant one or more angular relationships between each other, whereby said 3D image data comprises planar imaged surfaces (327a-c) corresponding to said planar real surfaces (322; 422; 522). The planar imaged surfaces (327) are detected (702) t in the 3D image data. The transform is determined based on the detected planar imaged surfaces (327a-c) and said one or more angular relationships between the planar real surfaces (322a-c; 422a-c; 522a-c), such that the transform, when applied to a description of the detected imaged surfaces (327a-c) in coordinates of said 3D data, results in that angular relationships between the detected imaged surfaces (327a-c) match said angular relationships between the planar real surfaces (322a-c; 422a-c; 522a-c).