Cable Overlap Detection Using RGB Superpixels and Depth Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional cable detection methods are unsuitable for detecting stacked cables and often result in inaccurate grasping due to the inability to differentiate between cables that are overlapped and those that are not.
Innovation Solution
A cable detection method that involves acquiring RGB and depth images, performing image segmentation to form superpixels, determining target similarities between superpixels, and using depth images to identify overlapping relationships between cables.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional cable detection methods based on deep learning or machine learning are used, then scattered cables can be detected, but stacked or overlapped cables cannot be accurately detected
Solution Approach 1:
The patent segments the image processing task into multiple stages: first performing image segmentation to obtain superpixels, then detecting cables in non-overlapped regions using similarity comparison, and finally handling overlapped regions using depth information. This segmentation approach allows the system to accurately detect both scattered and stacked cables by applying different detection strategies to different spatial regions.
Solution Approach 2:
The patent transitions from two-dimensional RGB image analysis to three-dimensional detection by incorporating depth information from depth images. When cables are detected to be overlapped in the 2D image, the system uses depth values to determine spatial relationships and identify which cable is above or below, enabling accurate detection of stacked cables that would otherwise be indistinguishable in 2D.
2Measurement precision
If conventional detection methods are used, then detection speed is maintained, but detection accuracy for overlapped cables deteriorates
Solution Approach 1:
The patent divides the detection process into distinct modules: an image segmentation module that creates superpixels, a cable detection module that uses similarity comparison on non-overlapped regions, and an overlapped cable detection module that uses depth information. This modular segmentation allows each module to be optimized independently while maintaining overall system efficiency and accuracy.
Solution Approach 2:
The patent introduces depth images as an intermediary data source to resolve ambiguities in overlapped cable detection. The depth information acts as a mediator that provides spatial relationship data when RGB image analysis alone is insufficient, enabling accurate determination of cable stacking relationships without requiring complex computational models.
3Loss of information
If only RGB image analysis is used, then processing simplicity is maintained, but ability to distinguish overlapped cables is lost
Solution Approach 1:
The patent segments the image space into overlapped and non-overlapped regions, applying different processing strategies to each. In non-overlapped regions, simple similarity comparison on RGB images suffices. In overlapped regions, depth information is selectively utilized. This segmentation approach minimizes computational energy by avoiding unnecessary depth processing in regions where it is not needed.
Solution Approach 2:
The patent applies depth information selectively only to regions where cables are detected to be overlapped, rather than processing the entire image with full 3D analysis. This partial application of the more computationally intensive depth-based method reduces overall energy consumption while still capturing the necessary spatial relationship information where required.
Data Source
AI summary
The present application relates to a cable detection method, a robot and a storage device. The method includes: acquiring an RGB image of a cable and a corresponding depth image; performing image segmentation on the RGB image, and merging pixel points in the RGB image, which are adjacent in position and have similar pixel features, into super-pixels; determining a target similarity between at least some of the super-pixels and the adjacent super-pixels, and detecting cables in the RGB image according to the target similarity between at least some of the super-pixels and the adjacent super-pixels; determining whether the cables are overlapped; and when it is detected that the cables are overlapped, determining an overlapping relationship between the detected cables according to the depth image.