Stereo Imaging Mapping Using De Bruijn Light Patterns
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for mapping surface points to pixels in stereo imaging, particularly in intra-oral applications, face challenges due to irregular surfaces and limited space, leading to difficulties in achieving accurate correspondence between cameras.
Innovation Solution
A method using a sequence of monochrome light patterns projected along lines, where each pixel is assigned a unique label through a de Bruijn or Hamilton sequence, enabling accurate correlation of pixels to surface points, even in complex topographies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional light patterns are projected onto irregular surfaces, then surface illumination is achieved, but correspondence accuracy deteriorates due to distortion from sharp contours and irregular topography
Solution Approach 1:
The patent segments the light pattern into multiple discrete lines that can be independently controlled and projected at different orientations. This segmentation allows the system to adapt to irregular surfaces by projecting lines that follow surface contours rather than using a single complex pattern, thereby maintaining correspondence accuracy without requiring overly complex overall pattern design.
Solution Approach 2:
The patent applies different pattern characteristics to different regions of the surface by projecting lines with varying orientations and spacings tailored to local surface geometry. This local adaptation ensures that each region is illuminated in a way that maximizes correspondence detection accuracy for that specific area, rather than using a uniform pattern across the entire surface.
2Reliability
If structured light patterns are used to solve the correspondence problem, then feature point identification is improved, but measurement ambiguity increases in regions with significant surface distortion
Solution Approach 1:
The patent introduces temporal dimension by projecting sequences of line patterns at different orientations over time. This transforms a 2D spatial pattern problem into a 3D spatiotemporal problem, where the same surface point can be identified across multiple time points with different line orientations, reducing ambiguity even in highly distorted regions.
Solution Approach 2:
The patent uses periodic projection of line patterns with varying orientations and spacings. By repeatedly projecting patterns in a systematic sequence, the system accumulates multiple measurements of the same surface features from different angular perspectives, allowing statistical resolution of correspondence ambiguity in distorted regions.
3Measurement precision
If multiple cameras are used for stereo imaging, then surface contour detection is improved, but device complexity and space requirements increase in confined areas
Solution Approach 1:
The patent makes the single camera system universal by implementing a computational framework that can handle both structured light pattern projection and stereo-like correspondence analysis. The system performs multiple functions including pattern projection, image capture, correspondence matching, and 3D reconstruction within a single integrated platform, eliminating the need for separate multi-camera hardware while achieving similar measurement capabilities.
Solution Approach 2:
The patent introduces a computational intermediary layer that processes single-camera images to extract stereo-like depth information. This computational mediator performs correspondence matching and 3D reconstruction algorithms that traditionally required multiple cameras, thereby achieving multi-camera functionality through software processing rather than additional hardware.
4Measurement precision
If spectral patterning is used to illuminate the surface, then feature differentiation is improved, but detection accuracy varies with tooth surface optical properties and wavelength sensitivity
Solution Approach 1:
The patent changes the spatial parameters of the light pattern (line orientation, spacing, and position) rather than relying on spectral/wavelength variations. By modulating the spatial structure of the illumination pattern and using monochrome light, the system achieves feature differentiation through geometric encoding that is independent of surface optical properties and camera wavelength sensitivity, ensuring consistent detection across different tooth surfaces.
Data Source
AI summary
A method for registering a first imaging detector to a surface projects a sequence of k images toward the surface, wherein k≧4, wherein each of the k images has a pattern of lines that extend in a direction that is orthogonal to a movement direction. The pattern encodes an ordered sequence of labels, each label having k binary elements, such that, in the movement direction, any portion of the pattern that is k equal increments long encodes one label of the ordered sequence. The method obtains, for at least a first pixel in the first imaging detector, along at least one line that is parallel to the movement direction, a first sequence of k signal values indicative of the k binary elements of a first label from the ordered sequence of labels and correlates the at least the first pixel in the first imaging detector to the surface.