Skill 审计报告

robot-perception 审计报告.

Comprehensive best practices for robot perception systems covering cameras, LiDARs, depth sensors, IMUs, and multi-sensor setups. Use this skill when working with RGB image processing, depth maps, point clouds, sensor calibration (intrinsic, extrinsic, hand-eye), object detection, semantic segmentation, 3D reconstruction, visual servoing, or perception pipeline optimization. Trigger whenever the user mentions OpenCV, Open3D, PCL, RealSense, ZED, OAK-D, camera calibration, AprilTags, ArUco markers, stereo vision, RGBD, point cloud filtering, ICP registration, coordinate transforms, camera intrinsics, distortion correction, image undistortion, sensor streaming, frame synchronization, or any computer vision task in a robotics context. Also covers multi-camera rigs, time synchronization across sensors, perception latency budgets, and production deployment of perception pipelines.