Perception and Control of Vision-Based Sorting Manipulators for Express Parcels

Abstract

Due to the rapid development of the e-commerce logistics industry in today’s society, the problems of low efficiency and high error rate in manual sorting of express deliveries have become increasingly apparent. Hence, it needs to add an express sorting robotic arm with a visual model. This paper intends to look at the necessity and present usage situation for these kinds of robotic arms now. This article is mainly presented in the form of a workflow through flow charts and text descriptions, giving an elaborate description of the workings and principal formulas of the mean filter and the Gaussian filter. YOLO is faster and more useful in places where higher real - time is needed, but SSD has better accuracy and precision, so they can be used for robot vision as well as videos. And it shows that it is not adaptable in situations like parcels stacked up, packages damaged, materials of different kinds of stuff, and complicated lights too. It says maybe researchers should change ideas or use more cameras so it knows what it’s holding better, gives some ideas on how to improve technology like this.