Image-based Pest Identification Using Support Vector Machine for Agricultural Crop Protection

The identification of pests is critical for the protection of crops and the guaranteeing of stable food supplies. Outbreaks of pests in Bangladesh result in significant losses, and the traditional techniques of pest control are frequently delayed and inefficient. This research proposes an automated approach that makes use of a Support Vector Machine (SVM), which was selected due to its excellent performance on complex datasets that are relatively modest. For the purpose of accurate classification, the model makes use of fused features such as texture (LBP), color (RGB/HSV), and shapes descriptors. It obtained 99% accuracy, precision, recall, and F1-score after being trained on 3,000 images of pests from ten different categories under a variety of situations. This approach provides farmers with a solution that is both more accurate and more expedient than the techniques that are currently in use. In the future, there are plans to implement mobile deployment and expand the model to include pests that are found in more regions.