用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/MIUAV/vibe-coding-ros2 --skill perception命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | perception |
| description | 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 |
| argument-hint | 四足感知 OR 机器人视觉 OR 深度相机 OR 激光雷达 |
| user-invocable | true |
用于配置和开发四足机器人的感知系统
当需要以下帮助时使用此技能:
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]
# 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
# 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)
# 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
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)
foot_sensors:
- type: proximity # 接近传感器
- type: force # 力传感器
- type: contact # 触地检测
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
| 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 |