No abstract

Background Learning from a few samples to automatically recognize the plant leaf diseases is an attractive and promising study to protect the agricultural yield and quality. The existing few-shot classification studies in agriculture are mainly based on supervised learning schemes, ignoring unlabeled data's helpful information. Methods In this paper, we proposed a semi-supervised few-shot learning approach to solve the plant leaf diseases recognition. Specifically, the public PlantVillage dataset is used and split into the source domain and target domain. Extensive comparison experiments considering the domain split and few-shot parameters (N-way, k-shot) were carried out to validate the correctness and generalization of proposed semi-supervised few-shot methods. In terms of selecting pseudo-labeled samples in the semi-supervised process, we adopted the confidence interval to determine the number of unlabeled samples for pseudo-labelling adaptively. Results The average improvement by the single semi-supervised method is 2.8%, and that by the iterative semi-supervised method is 4.6%. Conclusions The proposed methods can outperform other related works with fewer labeled training data.

In the area of plant protection and precision farming, timely detection and classification of plant diseases and crop pests play crucial roles in the management and decision-making. Recently, there have been many artificial neural network (ANN) methods used in agricultural classification tasks, which are task specific and require big datasets. These two characteristics are quite different from how humans learn intelligently. Undoubtedly, it would be exciting if the models can accumulate knowledge to handle continual tasks. Towards this goal, we propose an ANN-based continual classification method via memory storage and retrieval, with two clear advantages: Few data and high flexibility. This proposed ANN-based model combines a convolutional neural network (CNN) and generative adversarial network (GAN). Through learning of the similarity between input paired data, the CNN part only requires few raw data to achieve a good performance, suitable for a classification task. The GAN part is used to extract important information from old tasks and generate abstracted images as memory for the future task. Experimental results show that the regular CNN model performs poorly on the continual tasks (pest and plant classification), due to the forgetting problem. However, our proposed method can distinguish all the categories from new and old tasks with good performance, owing to its ability of accumulating knowledge and alleviating forgetting. There are so many possible applications of this proposed approach in the agricultural field, for instance, the intelligent fruit picking robots, which can recognize and pick different kinds of fruits; the plant protection is achieved by automatic identification of diseases and pests, which can continuously improve the detection range. Thus, this work also provides a reference for other studies towards more intelligent and flexible applications in agriculture.

Smart agriculture is inseparable from data gathering, analysis, and utilization. A high-quality data improves the efficiency of intelligent algorithms and helps reduce the costs of data collection and transmission. However, the current image quality assessment research focuses on visual quality, while ignoring the crucial information aspect. In this work, taking the crop pest recognition task as an example, we proposed an effective indicator of distance-entropy to distinguish the good and bad data from the perspective of information. Many comparative experiments, considering the mapping feature dimensions and base data sizes, were conducted to testify the validity and robustness of this indicator. Both the numerical and the visual results demonstrate the effectiveness and stability of the proposed distance-entropy method. In general, this study is a relatively cutting-edge work in smart agriculture, which calls for attention to the quality assessment of the data information and provides some inspiration for the subsequent research on data mining, as well as for the dataset optimization for practical applications.

No abstract