An effective feature extraction method for olive peacock eye leaf disease classification

Abstract

Early diagnosis of plant diseases is one of the key elements determining plant productivity. The productivity and quality of plants are significantly reduced when plant diseases are not identified and prevented in a timely manner, which results in major financial losses for producers. Olive is a plant with high added value. While the fruit and oil of olive are consumed as food, its oil is used in cosmetics, medicine, etc. It is also used in industries. In addition, active substances such as oleuropein, triterpene, maslinic acid, and flavonoid found in olive leaves are also used in the pharmaceutical industry. Considering all these valuable uses of olive, the importance of productivity is understood. Plant diseases are one of the most significant factors affecting the yield of olives. Among these diseases, fungal disease called peacock eye can spread to the whole tree through the leaves. This disease causes reduced crop production, defoliation, and rot of tree branches. In this study, an efficient method was developed to detect peacock eye disease from olive leaves. In the first stage, an original dataset of healthy and diseased leaves was created. Then, by extracting deep features from this dataset with CNN models, diseased and healthy leaf classification was performed with the transfer learning approach. As a result of the experiments, very satisfactory results were obtained around 98.63%.