Rice pests are one of the main factors affecting rice yield. The accurate identification of pests facilitates timely preventive measures to avoid economic losses. Some existing open source datasets related to rice pest identification mostly include only a small number of samples, or suffer from inter-class and intra-class variance and data imbalance challenges, which limit the application of deep learning techniques in the field of rice pest identification. In this paper, based on the IP102 dataset, we first reorganized a large-scale dataset for rice pest identification by Web crawler technique and manual screening. This dataset was given the name IP_RicePests. Specifically, the dataset includes 8248 images belonging to 14 categories. The IP_RicePests dataset was then expanded to include 14,000 images via ARGAN data augmentation technique to address the difficulties in obtaining large samples of rice pests. Finally, the parameters trained on the public image ImageNet dataset using VGGNet, ResNet and MobileNet networks were used as the initial values of the target data training network to achieve image classification in the field of rice pests. The experimental results show that all three classification networks combined with transfer learning have good recognition accuracy, among which the highest classification accuracy can be obtained on the IP_RicePests dataset via fine-tuning the parameters of the VGG16 network. In addition, following ARGAN data augmentation the dataset demonstrates high accuracy improvements in all three models, and fine-tuning the VGG16 network parameters obtains the highest accuracy in the augmented IP_RicePests dataset. It is demonstrated that CNN combined with transfer learning can employ the ARGAN data augmentation technique to overcome difficulties in obtaining large sample sizes and improve the efficiency of rice pest identification. This study provides foundational data and technical support for rice pest identification.

The occurrence of pests at high frequencies has been identified as a major cause of reduced citrus yields, and early detection and prevention are of great significance to pest control. At present, studies related to citrus pest identification using deep learning suffer from unbalanced sample sizes between data set classes, which may cause slow convergence of network models and low identification accuracy. To address the above problems, this study built a dataset including 5182 pest images in 14 categories. Firstly, we expanded the dataset to 21,000 images by using the Attentive Recurrent Generative Adversarial Network (AR-GAN) data augmentation technique, then we built Visual Geometry Group Network (VGG), Residual Neural Network (ResNet) and MobileNet citrus pest recognition models by using transfer learning, and finally, we introduced an appropriate attention mechanism according to the model characteristics to enhance the ability of the three models to operate effectively in complex, real environments with greater emphasis placed on incorporating the deep features of the pests themselves. The results showed that the average recognition accuracy of the three models reached 93.65%, the average precision reached 93.82%, the average recall reached 93.65%, and the average F1-score reached 93.62%. The integrated application of data augmentation, transfer learning and attention mechanisms in the research can significantly enhance the model’s ability to classify citrus pests while saving training cost and time, which can be a reference for researchers in the industry or other fields.

Citrus pests pose a major threat to both citrus yield and fruit quality. The early prevention of pests is essential for sustainable citrus cultivation, cost savings, and the reduction of environmental pollution. Despite the increasing application of deep learning techniques in agriculture, the performance of existing models for small target detection of citrus pests is limited, mainly in terms of information bottlenecks that occur during the transfer of information. This hinders its effectiveness in fully automating the detection of citrus pests. In this study, a new approach was introduced to overcome these limitations. Firstly, a comprehensive large-scale dataset named IP-CitrusPests13 was developed, encompassing 13 distinct citrus pest categories. This dataset was amalgamated from IP102 and web crawlers, serving as a fundamental resource for precision-oriented pest detection tasks in citrus farming. Web crawlers can supplement information on various forms of pests and changes in pest size. Using this comprehensive dataset, we employed the SPD Module in the backbone network to preserve fine-grained information and prevent the model from losing important information as the depth increased. In addition, we introduced the AFFD Head detection module into the YOLOv8 architecture, which has two important functions that effectively integrate shallow and deep information to improve the learning ability of the model. Optimizing the bounding box loss function to WIoU v3 (Wise-IoU v3), which focuses on medium-quality anchor frames, sped up the convergence of the network. Experimental evaluation on a test set showed that the proposed SAW-YOLO (SPD Module, AFFD, WIoU v3) model achieved an average accuracy of 90.3%, which is 3.3% higher than the benchmark YOLOv8n model. Without any significant enlargement in the model size, state-of-the-art (SOTA) performance can be achieved in small target detection. To validate the robustness of the model against pests of various sizes, the SAW-YOLO model showed improved detection performance on all three scales of pests, significantly reducing the rate of missed detections. Our experimental results show that the SAW-YOLO model performs well in the detection of multiple pest classes in citrus orchards, helping to advance smart planting practices in the citrus industry.

The accurate identification of crop aphids is an important aspect of improving agricultural productivity. Aphids are characterised by small targets and a body colour similar to their surroundings. Even the most advanced detectors can experience problems such as low detection accuracy and a high number of missed detections. In this paper, a multi-stream target detection model is proposed for fast and accurate detection of crop aphids in complex backgrounds. First, inspired by the human visual system, we propose a bionic attention (BA) approach. Unlike previous strategies, we do not improve the model but input additional category labels as bionic information streams into the network at the network input stage to support mainstream recognition, which improves the detection effect. In addition, through web crawling and manual screening, we construct an aphid dataset containing 2059 images, named IP_Aphids, based on IP102. In this paper, we combine the proposed BA with a number of classical target detection models, including YOLOv5s, YOLOv7-tiny, YOLOv8n, SSD, and faster R-CNN, and we also compare the BA with the classical attention mechanisms CBAM and SE. This approach allows the effectiveness of the method to be evaluated from multiple perspectives. The results show that combining the bionic information flow of BA with the original mainstream information flow as inputs to the five target detection models improves the mAP by 2.2%, 7%, 2.7%, 1.78%, and 1.12%, respectively, compared to the models using only the original inputs. In addition, the mAP of the YOLOv8n_BA model is higher than that of the YOLOv8n, YOLOv8n_CBAM, and YOLOv8n_SE models by 4.6%, 3.3%, and 2.7%, respectively. This indicates that the BA proposed in this study is significantly better than the classical attention to improve crop aphid detection, which provides a reference for crop aphid-related research.