This project presents an AI-based human and weapon detection system designed for ground surveillance using the Raspberry Pi 4. The system leverages real-time video processing and deep learning models to identify human figures and detect potential weapons in a given area. The Raspberry Pi 4, paired with a compatible camera module, captures video feeds that are processed using object detection algorithms such as YOLO (You Only Look Once) or MobileNet SSD, optimized for embedded devices. These models are capable of detecting human presence and recognizing weapon objects by analyzing each frame captured from the camera. The system operates efficiently within the limited resources of the Raspberry Pi, thanks to TensorFlow Lite or PyTorch, which enable fast model inference without significant power consumption. Upon detection, the system triggers alerts or activates response mechanisms, making it suitable for security and surveillance applications. The project highlights the potential of edge computing, allowing AI-driven surveillance on resource-constrained devices, and demonstrates the feasibility of using Raspberry Pi 4 for real-time object detection in practical, field-based scenarios. Future enhancements can include cloud integration for large-scale monitoring, motion detection for energy efficiency, and further model optimization for improved accuracy and speed.

 

The major building blocks of the project are:  

• Battery.
• Pi camera.
• Raspberry Pi 4.
• Flight Controller.
• 4 ESC.
• 3 Axis Gimbler.

     

Block diagram:  

video: