研究目的
To develop a wire detection tool using RGB-D images to aid robotic arms in machine alignment procedures, preventing damage to the wire during approach.
研究成果
The proposed RGB-D-based wire detection tool effectively aids robotic arms in machine alignment by automating wire detection, reducing the risk of damage. It shows promise for applications requiring fine line detection, though further optimizations for light robustness and execution time are needed.
研究不足
The system may fail due to light reflections and shadows affecting depth interpretation; robustness to light variations needs improvement. Detection is not fully reliable in all background conditions.
1：Experimental Design and Method Selection:
The system uses an RGB-D camera for depth segmentation and feature extraction to detect a thin wire in cluttered backgrounds. It involves capturing video frames, depth filtering, connected components analysis, and feature-based decision-making.
2：Sample Selection and Data Sources:
Experiments were conducted in a mock-up tunnel simulating the LHC environment, with videos recorded in different backgrounds (magnet and collimator).
3：List of Experimental Equipment and Materials:
Intel RealSense SR300 RGB-D camera, robotic arms, laser sensors, fiducial markers, and a stretched wire.
4：Experimental Procedures and Operational Workflow:
The camera is attached to a robotic arm; frames are acquired, pre-processed with masking, depth-filtered, analyzed for connected components, and features are extracted to detect the wire. The arm moves iteratively until detection.
5：Data Analysis Methods:
Loop time for detection is measured and averaged; performance is evaluated based on detection success and time efficiency.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
