Image Recognition of Supermarket Shopping Robot Based on CNN

摘要

With the development of society, more and more goods are sold intelligently and without people. The purpose of supermarket shopping robots is to help shop assistants replenish goods and free shop assistants from tedious mechanical work. Based on the “Innovation Robot Design and Production Competition-Supermarket Robot Challenge Competition”, the competition rules are taken as the design criteria. On the basis of CNN (convolution neural network), we design a CNN image recognition algorithm used in supermarket shopping robot. This algorithm overcomes the problems of low recognition accuracy and slow recognition speed in image recognition, and enables the robot to complete the competition more efficiently and quickly in the process of operation. The experimental results also verify that the image recognition algorithm can be applied to supermarket shopping robots and meet the needs of the competition.