Chip Surface Character Detection Based on Convolution Neural Network

With the development of computer technology, semiconductor chips have been more and more widely used. They will print some important information in production, such as batch number, merchant information, chip information, etc. This paper presents a chip surface character detection system based on convolutional neural network OCR technology. An industrial camera is used to capture chip surface image, then the image is processed and identified in computer software. At the same time, the real-time information of the current chip is obtained from the database and compared with the identified information so as to achieve the testing purpose. What's more, a hardware communication controller is designed, according to the different results, it can produce different processing.