Research Article
Hyperspectral Recognition and Early Warning of Rice Diseases and Insect Pests Based on Convolution Neural Network
@INPROCEEDINGS{10.1007/978-3-030-67871-5_19, author={Heng Xiao and Cao-Fang Long}, title={Hyperspectral Recognition and Early Warning of Rice Diseases and Insect Pests Based on Convolution Neural Network}, proceedings={Advanced Hybrid Information Processing. 4th EAI International Conference, ADHIP 2020, Binzhou, China, September 26-27, 2020, Proceedings, Part I}, proceedings_a={ADHIP}, year={2021}, month={2}, keywords={Convolution neural network Rice diseases and insect pests Hyperspectral imaging Recognition and early warning methods}, doi={10.1007/978-3-030-67871-5_19} }- Heng Xiao
Cao-Fang Long
Year: 2021
Hyperspectral Recognition and Early Warning of Rice Diseases and Insect Pests Based on Convolution Neural Network
ADHIP
Springer
DOI: 10.1007/978-3-030-67871-5_19
Abstract
The traditional method of disease and pest recognition uses SVM to classify and recognize the image. Because of the large training convergence error, the recognition accuracy is not high. In view of the above problems, the paper studies the method of rice hyperspectral pest identification and early warning based on convolution neural network. By reducing the dimension of the collected spectral image, we can get more image information and extract image features. Based on alexnet, the structure of convolutional neural network is designed. The recognition database was established by collecting the spectrum images of rice diseases and insect pests, and the convolution neural network was trained by transfer learning, so as to realize the recognition and early warning of rice diseases and insect pests. The experimental results show that the convergence error of the method based on convolution neural network is small and the recognition accuracy is higher.
