Graphical Fiducial Marker Detection Using GPU Corner Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional fiducial marker detection methods rely heavily on central processing units (CPUs) and result in significant computing and energy resource drains due to inefficient image processing techniques, particularly at higher image resolutions.
Innovation Solution
The use of computer vision processing implemented partly on graphics processing units (GPUs) for identifying fiducial markers, which includes filtering and segmentation for boundary identification, and corner detection followed by efficient corner filtering to reduce the number of pixels processed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional CPU-based fiducial marker detection methods are used, then the system can operate at lower image resolutions, but the frame rate drops significantly and computing resources are overwhelmed at higher resolutions
Solution Approach 1:
The patent replaces CPU-based sequential processing with GPU-based parallel processing. The GPU's architecture with thousands of cores enables simultaneous execution of multiple image processing operations, dramatically increasing frame rates while reducing the computational burden on the CPU. This substitution allows the system to handle high-resolution images (1080p, 4k) at 30 fps or greater, resolving the contradiction between productivity and energy consumption.
2Difficulty of detecting and measuring
If graph-based image segmentation algorithms are used to identify all lines and combine them into polygons, then comprehensive polygon detection is achieved, but an overwhelming number of identified polygons creates a drain on computing resources
Solution Approach 1:
The patent extracts and prioritizes only the most relevant polygons for fiducial marker detection rather than processing all detected polygons. By implementing filtering mechanisms that identify and focus on polygons with specific characteristics (such as those matching fiducial marker patterns), the system eliminates the computational overhead of processing overwhelming numbers of irrelevant polygons, thereby improving processing efficiency while maintaining detection completeness.
Solution Approach 2:
The patent applies partial action by implementing early filtering steps that eliminate obviously non-fiducial polygons before full analysis. Rather than exhaustively analyzing every detected polygon, the system performs selective processing on a subset of candidate polygons, reducing computational load while still achieving complete fiducial marker detection.
3Measurement precision
If segmentation is used to identify boundaries and then analyze each pixel along the boundaries to determine corners, then accurate corner detection is achieved, but significant computing, time, and energy costs are incurred
Solution Approach 1:
The patent applies partial action by analyzing only a subset of pixels along boundaries rather than every single pixel. The system identifies candidate corner regions using efficient algorithms and then performs detailed analysis only on these localized areas, significantly reducing processing time and energy consumption while maintaining corner detection accuracy through targeted pixel analysis.
Data Source
AI summary
In various examples, image data may be received that represents an image. Corner detection may be used to identify pixels that may be candidate corner points. The image data may be converted from a higher dimensional color space to a converted image in a lower dimensional color space, and boundaries may be identified within the converted image. A set of the candidate corner points may be determined that are within a threshold distance to one of the boundaries, and the set of the candidate corner points may be analyzed to determine a subset of the candidate corner points representative of corners of polygons. Using the subset of the candidate corner points, one or more polygons may be identified, and a filter may be applied to the polygons to identify a polygon as corresponding to a fiducial marker boundary of a fiducial marker.