Light Adaptive Room Clutter Tidy Smart Car

With the continuous progress of urbanization and social technology, the shift of family model makes the room clutter cleaning robot gradually become a necessary tool for modern families. In order to improve the robot’s target detection, automatic grasping and classification ability in dark light environment, this paper takes YOLOv4_tiny as the core, combines Ghost convolution and data enhancement algorithm for image brightness, and utilizes positional PID algorithm to realize accurate localization and tracking, and designs a highly efficient, intelligent, and user-friendly room clutter tidying intelligent vehicle. The actual test results show that the model accuracy is 95.2%, the recall is 97.9%, the frame rate is 25 fps, and the running speed of the cart is not affected by the ambient lighting, which is a good performance and has the potential to be applied in family life.