Deep learning is a branch of artificial intelligence. In recent years, with the advantages of automatic learning and feature extraction, it has been widely concerned by academic and industrial circles. It has been widely used in image and video processing, voice processing, and natural language processing. At the same time, it has also become a research hotspot in the field of agricultural plant protection, such as plant disease recognition and pest range assessment, etc. The application of deep learning in plant disease recognition can avoid the disadvantages caused by artificial selection of disease spot features, make plant disease feature extraction more objective, and improve the research efficiency and technology transformation speed. This review provides the research progress of deep learning technology in the field of crop leaf disease identification in recent years. In this paper, we present the current trends and challenges for the detection of plant leaf disease using deep learning and advanced imaging techniques. We hope that this work will be a valuable resource for researchers who study the detection of plant diseases and insect pests. At the same time, we also discussed some of the current challenges and problems that need to be resolved.

The intelligent identification and classification of plant diseases is an important research objective in agriculture. In this study, in order to realize the rapid and accurate identification of apple leaf disease, a new lightweight convolutional neural network RegNet was proposed. A series of comparative experiments had been conducted based on 2141 images of 5 apple leaf diseases (rust, scab, ring rot, panonychus ulmi, and healthy leaves) in the field environment. To assess the effectiveness of the RegNet model, a series of comparison experiments were conducted with state-of-the-art convolutional neural networks (CNN) such as ShuffleNet, EfficientNet-B0, MobileNetV3, and Vision Transformer. The results show that RegNet-Adam with a learning rate of 0.0001 obtained an average accuracy of 99.8% on the validation set and an overall accuracy of 99.23% on the test set, outperforming all other pre-trained models. In other words, the proposed method based on transfer learning established in this research can realize the rapid and accurate identification of apple leaf disease.

To perform accurate and synchronous detection of the soluble solid contents (SSC) in fresh jujubes at different stages of maturity, hyperspectral imaging was used to establish robust models. The combined data constituting four maturation stages were used to build the grid-search least squares support vector machine (GS-LS-SVM) model. The determination coefficient (Rp²), the root-mean-square error (RMSEP), and the residual predictive deviation (RPD) of the prediction set for samples of the overall stages were 0.98, 1.10%, and 7.85, respectively. Furthermore, a successive projections algorithm (SPA) was used to extract the characteristic wavelengths of the combined data. An artificial bee colony (ABC) algorithm (for the prediction set, Rp² = 0.98, RMSEP = 1.19%, RPD = 7.25) was used to improve the SPA-LS-SVM model, which was better than the SPA-GS-LS-SVM model (for the prediction set, Rp² = 0.98, RMSEP = 1.24%, RPD = 6.96). Lastly, visualization of the SSC distribution map was performed based on the SPA-ABC-LS-SVM model, which clearly showed that the SSC gradually increased during maturation. The results indicated that it was realistic to construct a detection model of the multimaturity stage. This research also demonstrated that the combination of hyperspectral imaging and the ABC had good application values in the testing of agricultural products.