A CNN-based Decision Support System for Pests and Disease Control in Cucumber Plant

Résumé

Cucumber is a well-known vegetable because of its benefits and nutritional composition such as low-calorie, high water content, essential vitamins and minerals. The relatively short cultivation period and low requirement for specialized care also make cucumber farming an accessible option for small-scale farmers, contributing to improved livelihoods and food security. Cucumber cultivation contributes to sustainable agricultural practices; in that they serve as effective cover crops, they prevent soil erosion and suppress weed growth. However, pests such as insects, mites, and rodents, as well as diseases caused by fungi, bacteria, and viruses rapidly spread through cucumber, causing widespread infections, reduced quality, and market value depreciation. It has been found that pests reduce the average yield of vegetables (i.e. cucumber) by up to 50%. Cucumber is easily infested by pests and diseases which in turn affect their growth and consequently reduce crop yields. Aphids, spider mites, and cucumber beetles are among the pests that feed on the cucumber. Several efforts have been made to control pests and disease in cucumber cultivation, but because of their apparent challenges such as problem of  efficiency, accuracy, and  real-time analysis capabilities, with resulting inadequacy in management, delay in identifying and addressing pest and disease outbreaks, there is a need for a more accurate and effective system. In this paper, a smart system for the control of pest and disease in cucumber cultivation was presented. Dataset from standard repository was used for the model training. Opencv was used for feature extraction image data. The trained model was evaluated, and the evaluation results showed that precision of 92.7%, recall of  91.4%, and accuracy of 98%.