Jetson nano based YOLOV5 deep learning pavement pothole detection

As an important part of the entire logistics system, intelligent, mechanised and automated end-to-end logistics distribution has become a very challenging research direction in today’s industry. The timeliness and low cost of logistics distribution has become an important consideration for enterprises, and the service performance and efficiency of unmanned end-of-line delivery robots have a direct impact on the user’s satisfaction and experience, and indirectly affects the collective social trust in unmanned end-of-line delivery robots. The world’s unmanned delivery robots are developing rapidly, but still face various problems. The safety of unmanned terminal delivery robot’s working movement is the most basic guarantee for its successful completion of the work, this paper studies the unmanned delivery robot’s obstacle avoidance problem at work and realises the accurate detection of potholes on the road surface based on the YOLOV5 deep learning model.