How do I install the Perception skill?

Run npx skillmds@latest add miuav/perception-3 in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

What does the Perception skill do?

四足机器人感知系统 - 视觉识别、深度相机、激光雷达、触觉传感器配置 It is listed under Coding & Dev Tools on SkillMD.

Is Perception safe to use?

This skill has not completed SkillMD's automated safety review yet. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Perception?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Perception free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Perception?

miuav (@miuav) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Perception by miuav — Agent Skill

Name: Perception
Author: miuav

File contents

四足机器人感知技能

用于配置和开发四足机器人的感知系统

何时使用

当需要以下帮助时使用此技能：

配置视觉传感器
设置激光雷达
实现目标识别
环境感知融合

快速参考

传感器配置

quadruped_perception:
  # 深度相机
  camera:
    type: realsense / astra / orbbec
    topics: [/camera/color/image, /camera/depth/image]
    
  # 激光雷达
  lidar:
    type: rplidar / livox / garfield
    topics: [/scan, /point_cloud]
    
  # IMU
  imu:
    type: BMI088 / CH110
    topics: [/imu/data]

视觉感知

深度相机配置

# Intel RealSense
ros2 launch realsense2_camera rs_launch.py

# Astra
ros2 launch astra_camera astra_launch.py

# ORBBEC
ros2 launch orbbec_camera orbbec_launch.py

目标检测

# YOLO 目标检测
class QuadrupedVision:
    def __init__(self):
        self.yolo = YOLOWrapper("yolov8n.onnx")
        
    def detect_objects(self, image):
        results = self.yolo.predict(image)
        return self.filter_quadruped_relevant(results)

激光雷达感知

雷达配置

# RPLIDAR
ros2 launch sllidar_ros2 sllidar_a1_launch.py

# Livox
ros2 launch livox_ros_driver livox_lidar_launch.py

# 思岚
ros2 launch rplidar_ros2 rplidar_a3_launch.py

点云处理

# 点云滤波
def filter_point_cloud(point_cloud):
    # 地面移除
    ground_filter = PointCloudFilter()
    ground_filter.set_optimized_coefficients([0, 0, 1, 0])
    
    filtered = ground_filter.apply(point_cloud)
    return filtered

传感器融合

多传感器融合

class PerceptionFusion:
    def __init__(self):
        self.camera_sub = ...
        self.lidar_sub = ...
        self.imu_sub = ...
        
    def fuse(self):
        # 相机目标检测
        objects = self.detect_objects()
        
        # 激光雷达距离
        distances = self.get_lidar_distances()
        
        # IMU 姿态
        orientation = self.get_orientation()
        
        # 融合结果
        return self.merge_sensory_data(objects, distances, orientation)

地形感知

足端感知

foot_sensors:
  - type: proximity    # 接近传感器
  - type: force       # 力传感器
  - type: contact     # 触地检测

地形分类

class TerrainClassifier:
    def classify(self, sensor_data):
        features = extract_features(sensor_data)
        
        terrain_types = {
            'flat': 0.9,
            'rough': 0.05,
            'stairs': 0.03,
            'slope': 0.02
        }
        
        return terrain_types

常用功能包

包	功能
`vision_msgs`	视觉消息类型
`darknet_ros`	YOLO 目标检测
`lidar_localization`	激光雷达定位

相关文档

---name: perception-3description: 四足机器人感知系统 - 视觉识别、深度相机、激光雷达、触觉传感器配置---# 四足机器人感知技能> 用于配置和开发四足机器人的感知系统---## 何时使用当需要以下帮助时使用此技能：- 配置视觉传感器- 设置激光雷达- 实现目标识别- 环境感知融合---## 快速参考### 传感器配置```yamlquadruped_perception: # 深度相机 camera: type: realsense / astra / orbbec topics: [/camera/color/image, /camera/depth/image] # 激光雷达 lidar: type: rplidar / livox / garfield topics: [/scan, /point_cloud] # IMU imu: type: BMI088 / CH110 topics: [/imu/data]```---## 视觉感知### 深度相机配置```bash# Intel RealSenseros2 launch realsense2_camera rs_launch.py# Astraros2 launch astra_camera astra_launch.py# ORBBECros2 launch orbbec_camera orbbec_launch.py```### 目标检测```python# YOLO 目标检测class QuadrupedVision: def __init__(self): self.yolo = YOLOWrapper("yolov8n.onnx") def detect_objects(self, image): results = self.yolo.predict(image) return self.filter_quadruped_relevant(results)```---## 激光雷达感知### 雷达配置```bash# RPLIDARros2 launch sllidar_ros2 sllidar_a1_launch.py# Livoxros2 launch livox_ros_driver livox_lidar_launch.py# 思岚ros2 launch rplidar_ros2 rplidar_a3_launch.py```### 点云处理```python# 点云滤波def filter_point_cloud(point_cloud): # 地面移除 ground_filter = PointCloudFilter() ground_filter.set_optimized_coefficients([0, 0, 1, 0]) filtered = ground_filter.apply(point_cloud) return filtered```---## 传感器融合### 多传感器融合```pythonclass PerceptionFusion: def __init__(self): self.camera_sub = ... self.lidar_sub = ... self.imu_sub = ... def fuse(self): # 相机目标检测 objects = self.detect_objects() # 激光雷达距离 distances = self.get_lidar_distances() # IMU 姿态 orientation = self.get_orientation() # 融合结果 return self.merge_sensory_data(objects, distances, orientation)```---## 地形感知### 足端感知```yamlfoot_sensors: - type: proximity # 接近传感器 - type: force # 力传感器 - type: contact # 触地检测```### 地形分类```pythonclass TerrainClassifier: def classify(self, sensor_data): features = extract_features(sensor_data) terrain_types = { 'flat': 0.9, 'rough': 0.05, 'stairs': 0.03, 'slope': 0.02 } return terrain_types```---## 常用功能包| 包 | 功能 ||----|------|| `vision_msgs` | 视觉消息类型 || `darknet_ros` | YOLO 目标检测 || `lidar_localization` | 激光雷达定位 |---## 相关文档- [Realsense ROS2](https://github.com/IntelRealSense/realsense-ros)- [Livox SDK](https://github.com/Livox-SDK)

Frequently asked questions

Perception