Tablet Surface Defect Detection Based on Convolutional Neural Network

• 1、College of Engineering, South China Agricultural University, Guangzhou Guangdong 510642, China

Abstract：Tablet defect detection is an important means to ensure the quality of drugs. There are many problems in manual detection, such as high cost, low efficiency and strong subjectivity. Although the traditional machine vision method based on pattern recognition has improved compared with manual detection method, it still has some problems, such as complex process and low detection accuracy. In order to solve the above problems, a method based on convolution neural network was used to detect the surface defects of tablets. First, the CCD industrial camera and photoelectric sensor are used to automatically collect the image of the tablet, and then the image of the tablet is pre-processed to remove the noise of the tablet image and improve the image quality of the tablet. Finally, a convolutional neural network is used to detect the eligibility of the tablets and to further detect the type of defects in the unqualified tablets. The experimental results show that the detection method based on convolutional neural network can be improved by nearly 10% compared with the traditional machine vision method.

Keywords： Deep learning Convolutional Neural Network (CNN) network structure defect tablets defect classification