Research on the visual identification and classification-based perception method of trash bins for automated garbage trucks

Abstract:
With the development of autonomous driving technology, there is a trend toward automation in special vehicles including garbage trucks. Due to its specialized use, how to identify and classify trash bins in time can become the key problem of automated garbage trucks. In this research, we proposed a trash bin perception method for garbage trucks, which integrates a trash bin identification method based on conventional neural network (CNN) and a trash bin classification method based on color recognition. According to the relative position of target trash bin, the autonomous driving system can stop the garbage truck at the right position and grasp the target trash bin in time. Test shows that the method designed in this research can recognition the target trash bin rapidly, and realize the automatic operation cooperated with the autonomous driving system.