Edge AI-Powered Surveillance: Intelligent Threat Detection

The Edge AI-Powered Surveillance system operates locally at the network edge to detect threats without relying on cloud infrastructure. At the core of the vision system is the Sipeed Maixduino RISC-V microcontroller board paired with a GC0328 camera module. The camera captures continuous visual frames, which are processed locally by the Maixduino using on-board machine learning models trained for object detection. A 128GB microSDXC card is integrated into the board to store high-capacity local image and data logs. When the AI model detects a specific intelligent threat, the Maixduino communicates the alert data to a connected NodeMCU (ESP8266) or utilizes the integrated SX-1278 LoRa modules. The SX-1278 transceivers establish a long-range wireless link, allowing the system to transmit critical threat alerts over extended distances to a remote receiver without needing cellular coverage or internet access. Power and housing are managed through a custom 3D-printed case designed to securely enclose the Maixduino hardware setup. Programming and deployment are split between the MaixPy IDE for configuring the edge machine learning environment on the RISC-V board and the Arduino IDE for handling auxiliary microcontroller logic and LoRa radio communication protocols using MicroPython and Arduino libraries.