CV Studio

A node-based computer vision platform for flow analytics, mapping, and spatial decision support.

🎯 Overview

CV Studio helps teams observe, count, localize, map, and compare activity over time through visual pipelines. It is positioned first as a platform for retail intelligence, urban analytics, and spatial decision support.

You connect video or field data to detection, tracking, analytics, and map layers, then publish results to dashboards, databases, or APIs without building custom glue code for every workflow.

🧭 Why Teams Use CV Studio

• Retail intelligence — Measure occupancy, footfall, hot zones, and in-store movement from existing camera feeds
• Urban analytics — Analyze density, trajectories, curb activity, and space usage across streets, stations, and facilities
• Spatial decision support — Combine detections, coordinates, maps, and satellite imagery for territory monitoring and site review

⚙️ How It Works

flowchart LR
    subgraph Inputs["📥 Acquisition"]
        A1["Webcam / Video"]
        A2["RTSP · HLS · WebRTC"]
        A3["GPS / JSON / MQTT / API"]
        A4["Satellite / Copernicus"]
    end
    subgraph Process["⚙️ Detection, Tracking & Analytics"]
        B1["Object Detection<br/>(ONNX · YOLO)"]
        B2["Tracking &amp;<br/>Trajectories"]
        B3["Heatmaps · Charts<br/>&amp; Spatial Analysis"]
    end
    subgraph Outputs["📤 Spatial Outputs"]
        C1["Maps &amp;<br/>Satellite Context"]
        C2["Overlays · Views<br/>&amp; Dashboards"]
        C3["MongoDB · MQTT<br/>API / JSON"]
    end

    Inputs --> Process --> Outputs

    style Inputs fill:#e3f2fd,stroke:#1976d2,color:#0d47a1
    style Process fill:#f3e5f5,stroke:#7b1fa2,color:#4a148c
    style Outputs fill:#e8f5e9,stroke:#388e3c,color:#1b5e20

Loading

✨ Core Workflows

• Retail flow analysis — Camera → detection → tracking → heatmap / chart → occupancy and movement insight
• Urban mobility analysis — Street or facility feed → trajectories → map overlay → density and path interpretation
• Spatial diagnostics — Coordinates + satellite imagery + detections → site context and territory monitoring

🧱 Core Platform

• Visual workflow builder for node-based pipeline design with real-time previews
• Video, stream, and field inputs including webcam, RTSP, HLS, WebRTC, JSON, MQTT, API, and satellite imagery
• Detection, tracking, and analytics with ONNX / YOLO models, MOT, trajectories, heatmaps, and charts
• Spatial outputs and integrations for map overlays, dashboards, MongoDB, MQTT, and API export
• Reusable pipelines saved as JSON for repeatable analysis and demos

🎥 See It In Action

The demos below are ordered to tell the product story first: overview, live analytics, workflow construction, model operations, and advanced pipelines.

▶️ Platform Overview	▶️ Live Video Analytics	▶️ Workflow Builder
▶️ Detection & Tracking Modules	▶️ Advanced Pipeline

💡 Click any thumbnail to watch the full video on YouTube.

🚀 Quick Start

git clone https://github.com/hackolite/CV_Studio.git
cd CV_Studio
pip install -r requirements.txt
python main.py

Need detailed setup?

• General installation: doc/INSTALLATION.md
• Windows install: English · Français
• Get the Windows executable: English · Français
• Build guides: Quick reference · Full guide

📚 Documentation Map

• Full documentation: doc/DOCUMENTATION.md
• Architecture overview: src/README.md
• Queue system details: doc/TIMESTAMPED_QUEUE_SYSTEM.md
• Examples and demos: examples/README.md
• Testing guide: doc/TESTING.md

🧩 Node Library

CV Studio includes a broad node library spanning:

• Inputs — image, video, webcam, RTSP, HLS, WebRTC, WebSocket, MQTT, API, GPS, microphone
• Vision processing — filtering, transforms, enhancement, thresholding, overlays
• AI inference — object detection, segmentation, classification, pose, depth, tracking, audio
• Spatial analytics — maps, coordinate streams, trajectories, heatmaps, charts, satellite layers
• Outputs and actions — visualization, JSON export, MongoDB, MQTT, APIs, workflow persistence

For the complete node catalog and detailed descriptions, see doc/DOCUMENTATION.md.

🧪 Testing

python -m pytest tests/ -v

For focused validation and testing notes, see doc/TESTING.md.

🛠️ Development

• Use the modern architecture in src/README.md for new work
• Review examples in examples/README.md
• Keep documentation and tests updated with changes

📋 Roadmap

• Improve graph editing and import reliability
• Continue GUI and plugin-system refactoring
• Expand type safety, monitoring, and export capabilities

👥 Authors & Contributors

Original Author:
Fork from Kazuhito Takahashi (@KzhtTkhs)

Repository Builder :
hackolite

We appreciate all contributions from the community!

📄 License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Important License Notes

• The source code of CV Studio itself is under Apache-2.0 license
• Each algorithm/node implementation is subject to its own license
• Please check the LICENSE file in each node directory for specific algorithm licenses
• Third-party dependencies have their own licenses

Image License

Sample images are sourced from:

• Free Material Pakutaso
• NHK Creative Library

🙏 Acknowledgments

• Original Image-Processing-Node-Editor project
• DearPyGUI for the GUI framework
• OpenCV for computer vision functionality
• ONNX Runtime for ML model inference
• MediaPipe for ML solutions
• All contributors and users of this project

📞 Support

• Issues: GitHub Issues
• Discussions: GitHub Discussions
• Documentation: See the docs in this repository

---

Made with ❤️ for the Computer Vision Community

⭐ Star this repo if you find it useful!