Embedded pedestrian detection based on the Alex net model

Abstract

A real-time pedestrian detection system based on deep learning is proposed for the safety of unmanned vehicles, which need to accurately identify pedestrians in front of them. Real-time pedestrian detection using the Raspberry pi 4B embedded system combined with the OpenCV framework, image processing such as frame difference and binarization, and the Alex net neural network model. The moving objects are marked with rectangular boxes. Experiments show that the system can better complete the real-time detection of pedestrians, and has good robustness and sensitivity, and higher image processing speed and accuracy compared to conventional pedestrian detection.