In recent years, deep-learning models have resulted in significant progress in insect recognition. However, training deep neural networks requires a large amount of data, and data collection and labeling are time consuming and labor intensive. This study proposes a method for establishing a synthetic image dataset of stored-product insects to provide well-labelled image data for insect detection tasks. Proxy virtual worlds are leveraged to obtain synthetic data with annotations. A dynamic generation approach was presented to generate synthetic images with diverse insect targets, various backgrounds, and changing lighting conditions by using a camera module in the constructed virtual scene. The coordinates of the bounding boxes and the category labels of insect targets in each synthetic image were obtained by calculating the geometrical relationships between the insect targets and the camera module. A texture translation network was developed to conduct image-to-image translation and launch to enhance the verisimilitude of the synthetic images. A synthetic image dataset was established for three insect species, Cyptolestes ferrugineus (Stephens), Sitophilus oryzae (Linnaeus), and Tribolium castaneum (Herbst).A set of assessments was introduced to evaluate the synthetic image dataset, including the statistical characteristics and experimental verification. The experimental results demonstrated that the use of synthetic data reduces the demand for real data. The proposed method may provide a novel solution for providing training data with correct annotations for insect detection, without tedious image collection and manual labeling.

Detecting stored-grain pests on the surface of the grain pile plays an important role in integrated pest management (IPM), which is crucial for grain security. Recently, numerous deep learning-based pest detection methods have been proposed. However, a critical limitation of existing methods is their inability to detect out-of-distribution (OOD) categories that are unseen during training. When encountering such objects, these methods often misclassify them as in-distribution (ID) categories. To address this challenge, we propose a one-stage framework named PestOOD for out-of-distribution stored-grain pest detection via flow-based feature reconstruction. Specifically, we propose a novel Flow-Based OOD Feature Generation (FOFG) module that generates OOD features for detector training via feature reconstruction. This helps the detector learn to recognize OOD objects more effectively. Additionally, to prevent network overfitting that may lead to an excessive focus on ID feature extraction, we propose a Noisy DropBlock (NDB) module and integrate it into the backbone network. Finally, to ensure effective network convergence, a Stage-Wise Training Strategy (STS) is proposed. We conducted extensive experiments on our previously established multi-class stored-grain pest dataset. The results show that our proposed PestOOD demonstrates superior performance over state-of-the-art methods, providing an effective AI-enabled solution to ensure grain security.