Parametric image-based identification for classification of five stored product coleopteran species using AI techniques

Accurate and early detection of insect pests is crucial in stored product protection to reduce economic losses and ensure food security. This study introduces a parametric image-based identification system designed to classify five significant beetle species commonly found in stored products: Sitophilus oryzae, Cryptolestes ferrugineus, Oryzaephilus surinamensis, Tribolium casteneum, and Rhyzopertha dominica. Utilizing artificial intelligence algorithms, this research aims to enhance the effectiveness and precision of pest identification in various storage settings. High-resolution images of the beetle species were obtained, followed by pre-processing steps such as noise reduction, normalization, and feature extraction. Key parameters, including morphological characteristics and color patterns, are meticulously analyzed for feature extraction. These features are fed into a Transformer Neural Network, specifically tailored for the unique challenges of insect classification. The system demonstrates high performance, achieving accuracy rates exceeding 95% in both training and testing phases. The model's robustness is validated through cross-validation and extensive testing with diverse datasets. This approach offers a scalable solution for real-time monitoring and management of insect pests in stored commodities. The findings have significant implications for integrated pest management programs, providing a reliable tool for early detection and intervention. Future work will expand the system to include a broader range of insect species and integrate it with IoT-powered storage management systems for automated, continuous monitoring. Overall, the parametric image-based identification system represents a significant advancement in entomology and stored product safety, leveraging AI to safeguard food supplies and improve pest control techniques.