Multi-class pest detection is one of the crucial components in pest management involving localization in addition to classification which is much more difficult than generic object detection because of the apparent differences among pest species. This paper proposes a region-based end-to-end approach named PestNet for large-scale multi-class pest detection and classification based on deep learning. PestNet consists of three major parts. First, a novel module channel–spatial attention (CSA) is proposed to be fused into the convolutional neural network (CNN) backbone for feature extraction and enhancement. The second one is called region proposal network (RPN) that is adopted for providing region proposals as potential pest positions based on extracted feature maps from images. Position-sensitive score map (PSSM), the third component, is used to replace fully connected (FC) layers for pest classification and bounding box regression. Furthermore, we apply contextual regions of interest (RoIs) as contextual information of pest features to improve detection accuracy. We evaluate PestNet on our newly collected large-scale pests’ image dataset, Multi-class Pests Dataset 2018 (MPD2018) captured by our designed task-specific image acquisition equipment, covering more than 80k images with over 580k pests labeled by agricultural experts and categorized in 16 classes. The experimental results show that the proposed PestNet performs well on multi-class pest detection with 75.46% mean average precision (mAP), which outperforms the state-of-the-art methods.

Increasing grain production is essential to those areas where food is scarce. Increasing grain production by controlling crop diseases and pests in time should be effective. To construct video detection system for plant diseases and pests, and to build a real-time crop diseases and pests video detection system in the future, a deep learning-based video detection architecture with a custom backbone was proposed for detecting plant diseases and pests in videos. We first transformed the video into still frame, then sent the frame to the still-image detector for detection, and finally synthesized the frames into video. In the still-image detector, we used faster-RCNN as the framework. We used image-training models to detect relatively blurry videos. Additionally, a set of video-based evaluation metrics based on a machine learning classifier was proposed, which reflected the quality of video detection effectively in the experiments. Experiments showed that our system with the custom backbone was more suitable for detection of the untrained rice videos than VGG16, ResNet-50, ResNet-101 backbone system and YOLOv3 with our experimental environment.

Human driving behaviors are personalized and unique, and the automobile fingerprint of drivers could be helpful to automatically identify different driving behaviors and further be applied in fields such as auto-theft systems. Current research suggests that in-vehicle Controller Area Network-BUS (CAN-BUS) data can be used as an effective representation of driving behavior for recognizing different drivers. However, it is difficult to capture complex temporal features of driving behaviors in traditional methods. This paper proposes an end-to-end deep learning framework by fusing convolutional neural networks and recurrent neural networks with an attention mechanism, which is more suitable for time series CAN-BUS sensor data. The proposed method can automatically learn features of driving behaviors and model temporal features without professional knowledge in features modeling. Moreover, the method can capture salient structure features of high-dimensional sensor data and explore the correlations among multi-sensor data for rich feature representations of driving behaviors. Experimental results show that the proposed framework performs well in the real world driving behavior identification task, outperforming the state-of-the-art methods.

The recent explosion of large volume of standard dataset of annotated images has offered promising opportunities for deep learning techniques in effective and efficient object detection applications. However, due to a huge difference of quality between these standardized dataset and practical raw data, it is still a critical problem on how to maximize utilization of deep learning techniques in practical agriculture applications. Here, we introduce a domain-specific benchmark dataset, called AgriPest, in tiny wild pest recognition and detection, providing the researchers and communities with a standard large-scale dataset of practically wild pest images and annotations, as well as evaluation procedures. During the past seven years, AgriPest captures 49.7K images of four crops containing 14 species of pests by our designed image collection equipment in the field environment. All of the images are manually annotated by agricultural experts with up to 264.7K bounding boxes of locating pests. This paper also offers a detailed analysis of AgriPest where the validation set is split into four types of scenes that are common in practical pest monitoring applications. We explore and evaluate the performance of state-of-the-art deep learning techniques over AgriPest. We believe that the scale, accuracy, and diversity of AgriPest can offer great opportunities to researchers in computer vision as well as pest monitoring applications.

In agriculture, pest always causes the major damage in fields and results in significant crop yield losses. Currently, manual pest classification and counting are very time-consuming and many subjective factors can affect the population counting accuracy. In addition, the existing pest localization and recognition methods based on Convolutional Neural Network (CNN) are not satisfactory for practical pest prevention in fields because of pests’ different scales and attitudes. In order to address these problems, an effective data augmentation strategy for CNN-based method is proposed in this paper. In training phase, we adopt data augmentation through rotating images by various degrees followed by cropping into different grids. In this way, we could obtain a large number of extra multi-scale examples that could be adopted to train a multi-scale pest detection model. In terms of test phase, we utilize the test time augmentation (TTA) strategy that separately inferences input images with various resolutions using the trained multi-scale model. Finally, we fuse these detection results from different image scales by non-maximum suppression (NMS) for the final result. Experimental results on wheat sawfly, wheat aphid, wheat mite and rice planthopper in our domain specific dataset, show that our proposed data augmentation strategy achieves the pest detection performance of 81.4 mean Average Precision (mAP), which improves 11.63%, 7.93%, 4.73% compared to three state-of-the-art approaches.