Hierarchical Binary Structured Light Patterns for Depth Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Structured light imaging devices face limitations in capturing both detailed scenes and extending depth range, as they typically rely on single spatial frequency patterns that are either better for details or depth, but not both simultaneously, and require stationary scenes.
Innovation Solution
The generation and use of hierarchical binary structured light patterns, which are formed by iteratively scaling a lower resolution binary pattern to multiple successively higher resolutions, allowing for improved detail capture and extended depth range while maintaining pattern structure, enabling more accurate and comprehensive depth mapping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single high spatial frequency pattern is used, then detail capture is improved, but depth range is reduced
Solution Approach 1:
The single pattern is segmented into multiple binary patterns with different spatial frequencies. Each binary pattern captures depth information at different ranges, and their combinations enable both fine detail capture and extended depth range simultaneously.
Solution Approach 2:
Multiple binary patterns with different spatial frequencies are combined into a composite structured light pattern. This composite pattern integrates the advantages of both high-frequency (detail capture) and low-frequency (depth range) patterns, resolving the trade-off between the two parameters.
2Length of stationary object
If a single low spatial frequency pattern is used, then depth range is extended, but detail capture is reduced
Solution Approach 1:
The low-frequency pattern is segmented and combined with high-frequency binary patterns. The segmentation allows the system to process different spatial frequency components separately and combine them to achieve both extended depth range and fine detail capture.
Solution Approach 2:
The structured light pattern is formed as a composite of multiple binary patterns with varying spatial frequencies. This composite structure enables the system to simultaneously achieve the depth range extension benefit of low-frequency patterns and the detail capture benefit of high-frequency patterns.
3Length of stationary object
If dynamic multi-pattern projection is used, then both detail capture and depth range are improved, but the scene must be stationary
Solution Approach 1:
Multiple binary patterns are merged into a single composite structured light pattern that can be projected in one shot. This merging eliminates the need for sequential projection of multiple patterns, thereby eliminating motion artifacts and enabling operation in dynamic environments while maintaining both detail capture and depth range.
Solution Approach 2:
The multiple binary patterns are pre-computed and combined into a single projected pattern before execution. This preliminary preparation allows the system to capture all necessary depth information in a single shot, making the system adaptable to dynamic scenes where objects may move during capture.
4Length of stationary object
If fixed pattern projection is used, then dynamic scenes are handled, but either detail capture or depth range must be compromised
Solution Approach 1:
The fixed projected pattern is designed as a composite of multiple binary patterns with different spatial frequencies. This composite structure enables the single-shot fixed pattern system to simultaneously achieve both detail capture and extended depth range, eliminating the need to compromise between these parameters while maintaining compatibility with dynamic scenes.
Data Source
AI summary
A method of image processing in a structured light imaging device is provided that includes receiving a captured image of a scene, wherein the captured image is captured by a camera of a projector-camera pair in the structured light imaging system, and wherein the captured image includes a pre-determined hierarchical binary pattern projected into the scene by the projector, wherein the pre-determined hierarchical binary pattern was formed by iteratively scaling a lower resolution binary pattern to multiple successively higher resolutions, rectifying the captured image to generated a rectified captured image, extracting a binary image from the rectified captured image at full resolution and at each resolution used to generate the pre-determined hierarchical binary pattern, and using the binary images to generate a depth map of the captured image.


