Bell pepper leaf disease classification with LBP and VGG-16 based fused features and RF classifier

The present work focuses on detecting disease in big fields of bell peppers using a hybrid feature fusion and machine learning approach. For the most part, farmers of bell peppers don’t know if their plants are contaminated with bacterial spot disease until it's too late. Reduced harvests are a common result of the spread of the disease. The key idea is to determine early on if the bell pepper plant is infected with bacterial spot disease. For classification random forest (RF) is used. The proposed method is divided into three stages namely-image pre-processing, feature extraction and classification. For feature extraction, we considered three features such as local binary pattern (LBP) features, visual geometry group network (VGG-16) features and fused LBP & VGG-16 features. The classification accuracy obtained by the proposed model is compared with classification accuracy obtained by some researchers in their articles. The accuracy obtained with RF classifier for pepper bell dataset with LBP feature, VGG-16 feature and LBP + VGG-16 fused feature are 78.11%, 92.28% and 99.75% respectively. With the help of the proposed model, farmers can quickly detect the onset of plant diseases and take preventative measures to limit their impact. The goal of this work is to devise a means of diagnosing bacterial spot disease in bell pepper plants using only images captured in the farm. © 2022, The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of Computer Applications and Management.