Smart curve guardian: real-time vehicle threat detection

This study is about the development of a cheap AI powered unit that will monitor the blind spot of the road in real-time with the ESP32-CAM module. It is connected with a streaming camera of live images in which it sends data of frames to a server to identify the vehicles using YOLO (You Only Look Once) and categorize them into divisions of size and type using ResNet-50. The outcome of threat levels is obtained and is utilized to energize the dynamic warning system (of LEDs), which is kept at different speeds of LED blinking depending on the types of vehicles. It is the primary development to optimize data transmissions between resource-limited platform embedded systems and high-performance cloud models, to develop a hybrid intelligent system of transportation applications in remote areas with low infrastructure.