Detection of capsule foreign matter defect based on BP neural network

Considering the fuzziness and diversity of the capsule foreign matter defect in the image, the BP neural network is applied to discern the capsule foreign matter defect Firstly, the capsule image is separated into three parts by vertical Sobel operator, and every part of image is processed by median filter to clear the noise. Then the histogram features of all the three parts of the image, namely smoothness, skewness, flatness, distortion, kurtosis and entropy are extracted and used as the input of the BP neural network. According to the inhomogeneity of the input data, a normalization method based on the clustering algorithm is proposed in this paper. Experiment results show that this method has high precision.