Introducing Our Enhanced DJI G2 + YOLOv10 Object Detection Project Repository!

At KCCI, we are thrilled to share a free public tutorial to help you transform your DJI drone into a real-time surveillance and object detection system with low-latency performance.

This tutorial demonstrates:

?How to use DJI G2 goggles with a Cosmostreamer box (Raspberry Pi 4B) for seamless video transmission. ?[**cosmostreamer_site**](https://cosmostreamer.com/products/djigoggles2/).

?Leverage YOLOv10 for accurate and fast object detection, Person detection, etc... (latency: 124ms end-to-end). ?

?A streamlined setup using NDI live streaming for effortless integration.

What’s Included?

Step-by-step setup instructions: From installing dependencies to running the project on your local machine. ?

Code repository: Explore our optimized Python scripts and Conda-based kernel integration. ?

Demo visuals: See the project in action with images and videos. ?

?New GitHub Repo: ?[**DJI-G2-YoloVX**](https://github.com/Ramkethar/Dji-G2-Yolovx/tree/main).

Key Highlights

?End-to-End Latency: ?

- 110ms (video transmission to PC). ?

- 124ms (with YOLOv10 inference).

?

?Hardware Used: ?

- DJI G2 Goggles ?

- Raspberry Pi 4B (Cosmostreamer Firmware) ?

- Windows 10/11 PC (Intel i5-i7, 7-12th Gen, rtx gpu at least 2070 or equivalent AMD).

?

?Optional Features: ?

- Free alternatives for video output via scrcpy (with a slight latency trade-off). ?

?Who Can Benefit?

- Developers building DJI-based AI pipelines. ?

- Industries looking for cost-effective surveillance solutions. ?

- Hobbyists eager to explore cutting-edge AI technology.

Get Started Today!

Explore the repository: [GitHub - DJIg2_YoloVX](https://github.com/Ramkethar/Dji-G2-Yolovx/tree/main) ?

Follow our page for more tutorials and insights.

We are proud to make technology accessible and impactful. Let us know your thoughts and tag anyone who might find this helpful!