Weed Classification Using Particle Swarm Optimization and Deep Learning Models

Weed is a plant that grows along with nearly all field crops, including rice, wheat, cotton, millets and sugar cane, affecting crop yield quality. Classification accurate identification of types weeds challenging task for farmers in earlier stage growth because similarity. To address this issue, an efficient weed classification model proposed the Deep Convolutional Neural Network (CNN) implements automatic feature extraction performs complex learning image classification. Throughout work, images were trained using CNN evolutionary computing approach to classify based on two publicly available datasets. The Tamil Nadu Agricultural University (TNAU) dataset used as first consists 40 classes other from Indian Council Agriculture Research – Directorate (ICAR-DWR) which contains 50 images. An effective Particle Swarm Optimization (PSO) technique applied automatically evolve improve its accuracy. was evaluated compared pre-trained transfer models such GoogLeNet, AlexNet, Residual neural (ResNet) Visual Geometry Group (VGGNet) This work shows performance PSO assisted significantly improved success rate by 98.58% TNAU 97.79% ICAR-DWR