A Deep Learning Assisted Crop Pest Classification Model by Using Enhanced Light Gradient Boost Machine (ELGBM) Logics

Many different kinds of pests and insects impede the development of vital field crops including wheat, sugarcane, and rice, and they also diminish agricultural yields. Due to their similar appearance in the early stages of crop growth, accurately classifying and identifying all varieties of crop insects is a challenging task for farmers. Automatic feature extraction and handling of complicated high-level features in image classification applications are two requirements for a new model that can solve this problem with the help of deep learning aided architectures. This study investigates the use of an Enhanced Light Gradient Boost Machine (ELGBM), a deep learning algorithm, for the purpose of crop pest classification and pesticide recommendation. The suggested model's efficacy is assessed by cross-validation with a traditional learning technique known as Convolution Neural Network (CNN). Furthermore, the system incorporates a recommendation component that proposes the most efficient pesticide options for crop protection depending on the discovered pests. The suggested methodology outperforms more traditional approaches in terms of classification accuracy, which is around 98%. The outcomes proved that the suggested ELGBM model outperformed pre-trained models in labeling different kinds of field crop insects, suggesting its potential for use in agricultural crop protection systems.