A high-performance computer vision toolkit for local webcam management featuring YOLO-based object detection and DA3Metric depth estimation.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install camera-yolo-operator
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install camera-yolo-operator using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Camera YOLO Operator?
The Camera YOLO Operator is a robust integration designed for Openclaw Skills that empowers AI agents to interface directly with local hardware for real-time visual analysis. It supports a diverse range of tasks including standard object detection, instance segmentation, pose estimation, and oriented bounding boxes (OBB) using the YOLO26s series.
Beyond simple detection, this skill integrates the DA3Metric-Large model for high-fidelity depth estimation, allowing developers to calculate real-world distances between the camera and detected objects. It is architected for cross-platform compatibility, leveraging NVIDIA CUDA on Linux/Windows and Apple's MPS on macOS for accelerated inference.
Camera YOLO Operator Use Cases
- Real-time pedestrian and vehicle tracking with unique ID assignment via ByteTrack.
- Monocular depth estimation for measuring object distance in robotic or automation contexts.
- Security monitoring through motion detection and automated frame capturing.
- Hardware debugging to ensure local camera backends (V4L2, DirectShow, AVFoundation) are functioning correctly.
How Camera YOLO Operator Works
- The operator initializes the local camera backend based on the host operating system (e.g., V4L2 for Linux).
- YOLO models are loaded from the centralized Openclaw Skills module cache to perform primary object classification and localization.
- If depth mode is active, the DA3Metric model processes the frames to generate a metric depth map.
- Detection boxes and depth data are fused to calculate the median distance for each identified object.
- Results are annotated with visual labels (class, confidence, and distance) and saved to the local project workspace.
Camera YOLO Operator Setup
First, navigate to your local Openclaw Skills directory and install the necessary Python dependencies:
cd ~/.openclaw/workspace/skills/camera-yolo-operator
pip install -r requirements.txt
pip install supervision
Next, initialize the AI models using the provided script:
bash scripts/download_models.sh
Camera YOLO Operator Data Schema & Taxonomy
The skill maintains a strict separation between code, model weights, and generated assets within the Openclaw Skills environment:
| Directory | Content Type | Purpose |
|---|---|---|
~/.cache/huggingface/modules/ |
Model Weights | Shared cache for YOLO (.pt) and DA3Metric models. |
~/.openclaw/workspace/projects/ |
Output Assets | Storage for captures, annotated videos, and detection logs. |
scripts/ |
Executables | Core logic for webcam capture, detection, and tracking. |
All output files follow a standardized naming convention such as frame_*.jpg or annotated_video.mp4 to ensure easy parsing by downstream agents.
Camera YOLO Operator Advanced Features
- Multi-platform GPU acceleration supporting CUDA (NVIDIA) and MPS (Apple Silicon).
- Depth Scaling Calibration: Fine-tune model distance outputs using the
--depth-scaleparameter to match physical measurements. - Intelligent Tracking: Integrated ByteTrack and Supervision libraries for de-duplicated object counting and trajectory plotting.
- Modular Model Search: Automated search path logic that prioritizes environment variables and centralized module caches.
SKILL.md
Loading
Related Openclaw Skills
Markdown Studio
Markdown Studio is a professional-grade document generator for creating structured technical, business, and academic documentation with full syntax support.
VAPT Intern Roadmap
A professional career development platform that generates personalized, evidence-based learning paths for aspiring Vulnerability Assessment and Penetration Testing professionals.
Yunxiao Work Items Skill
A powerful Openclaw Skills extension for managing Alibaba Cloud Yunxiao work items, tasks, and project communications directly through an AI agent.
Excel Studio
A professional spreadsheet generator for creating automated data tables, financial reports, and visual charts in Excel format.
Market Analysis Service
A comprehensive market intelligence tool providing automated enterprise trend analysis, competitor tracking, and deep user behavior insights.
Weather API Tool
A professional-grade AI skill that enables autonomous agents to fetch, parse, and utilize structured weather data from global providers.
Featured*
