AMD-AI-Engage

Agri-Edge: AMD-Accelerated Crop Disease Diagnosis

This project demonstrates a high-performance Edge AI solution for real-time agricultural diagnostics. It is specifically optimized to run on AMD Kria™ SOM and Zynq™ UltraScale+™ hardware.

🚀 Key Features

• Hardware Acceleration: Utilizes the AMD Deep Learning Processor Unit (DPU).
• Optimization: INT8 Quantization via AMD Vitis AI, reducing model size by 4x.
• Low Latency: Sub-20ms inference for offline field use.
• Practical Impact: Enables immediate localized treatment of crop diseases without cloud dependency.

🛠 Tech Stack

• Language: C++
• Framework: Vitis AI Runtime (VART)
• Libraries: OpenCV, XIR (Xilinx Intermediate Representation)
• Hardware Target: AMD Kria KV260 Vision AI Starter Kit

📂 Repository Structure

• /src: C++ source code for DPU execution.
• /model: Compiled .xmodel file for the AMD DPU.
• /tests: Sample images for verification.

📈 Performance Metrics

Metric	Result
Platform	AMD Kria KV260
Inference Time	18ms
Power Draw	< 5 Watts
Accuracy (INT8)	95.8%

⚖️ License

This project is licensed under the MIT License.