The Classification of Stored Grain Pests based on Convolutional Neural Network
Download as PDF
Dexian Zhang, Wenjun Zhao
With the advent of the era of big data, convolutional neural network (CNN) of deep learning has been widely used in the field of image recognition. The CNN has higher recognition rate and faster extraction speed of characteristic than traditional machines in learning methods. The CNNs theory is introduced for the recognition of stored grain pests in the granary environment. Firstly, stored grain images with pests are normalized and the implicit characteristics are extracted using a trained convolution kernel. Then, the maximum pool method is used to reduce the dimensionality of the extracted features. Finally, the Softmax classifier is used to classify the image of the test sample. The results show that CNN has a good ability to identify reserves and generalization ability.
Deep Learning, Convolutional Neural network, Stored-grain Pests Recognition, Feature Extration.