No abstract

Pathogen infection has greatly reduced crop production. As the symptoms of diseases usually appear when the plants are infected severely, rapid identification approaches are required to monitor plant diseases at early the infection stage and optimize control strategies. Hyperspectral imaging, as a fast and nondestructive sensing technology, has achieved remarkable results in plant disease identification. Various models have been developed for disease identification in different plants such as arable crops, vegetables, fruit trees, etc. In these models, important algorithms, such as the vegetation index and machine learning classification and methods have played significant roles in the detection and early warning of disease. In this paper, the principle of hyperspectral imaging technology and common spectral characteristics of plant disease symptoms are discussed. We reviewed the impact mechanism of pathogen infection on the photo response and spectrum features of the plants, the data processing tools and algorithms of the hyperspectral information of pathogen-infected plants, and the application prospect of hyperspectral imaging technology for the identification of plant diseases.

IntroductionPrecise identification of crop insects is a crucial aspect of intelligent plant protection. Recently, with the development of deep learning methods, the efficiency of insect recognition has been significantly improved. However, the recognition rate of existing models for small insect targets is still insufficient for insect early warning or precise variable pesticide application. Small insects occupy less pixel information on the image, making it more difficult for the model to extract feature information.MethodsTo improve the identification accuracy of small insect targets, in this paper, we proposed S-ResNet, a model improved from the ResNet, by varying its convolution kernel. The branch of the residual structure was added and the Feature Multiplexing Module (FMM) was illustrated. Therefore, the feature expression capacity of the model was improved using feature information of different scales. Meanwhile, the Adjacent Elimination Module (AEM) was furtherly employed to eliminate the useless information in the extracted features of the model.ResultsThe training and validation results showed that the improved residual structure improved the feature extraction ability of small insect targets compared to the original model. With compare of 18, 30, or 50 layers, the S-ResNet enhanced the identification accuracy of small insect targets by 7% than that on the ResNet model with same layer depth.

No abstract