CNN-Driven Ensemble Model For Rice Pests, Diseases And Weeds Classification And Detection From Whole Rice Plant Images.

Abstract

Early and accurate detection of rice insect pests, diseases and weeds is a major challenge in precision agriculture due to symptom similarity, variability in environmental conditions and limited annotated data. This paper proposes a hybrid deep learning framework that combines convolutional neural networks (VGG16, VGG19, InceptionV3, ResNet50) with a Support Vector Machine (SVM) classifier to detect and classify multiple biotic stressors from full-plant rice images. The approach addresses key limitations of traditional models by leveraging transfer learning, data augmentation and ensemble strategies to improve generalization across diverse field conditions. Experimental results demonstrate that the proposed system achieves a classification accuracy of up to 98.50%. The model is lightweight and optimized for deployment on mobile and edge devices, enabling real-time field applications. Additionally, it incorporates decision support based on integrated pest management and economic thresholds. This work advances the state-of-the-art in agricultural AI by providing a scalable, accurate and practical solution for biotic stress detection in rice crops.