Classification Of Diseases In Sawi Pakcoy Using The Convolutional Neural Network (CNN) Method

Abstract

The Brassisca Rapa L plant, commonly referred to as pakcoy, is a vegetable renowned for its economically significant leaves. Pakcoy thrives in both highland and lowland regions, characterized by its rapid harvest cycle and straightforward cultivation process. However, the marked increase in pakcoy cultivation has rendered the crop susceptible to diseases caused by fungi, viruses, pests, and other microbes, highlighting the necessity for effective management strategies to mitigate crop failure. This study explores the application of Convolutional Neural Networks (CNN) in the identification of pakcoy diseases through advanced pattern recognition and image analysis techniques. Utilizing a dataset comprising 1000 images of pakcoy leaves—500 depicting diseased specimens and 500 healthy ones—sourced from greenhouse plants, the images are processed using CNN with RGB configurations at a resolution of 512x512 pixels. The data training, conducted with the Adam optimizer, achieved an accuracy rate of 89.12% and a loss value of 0.240. The findings demonstrate that the CNN methodology is highly effective in accurately classifying diseases in pakcoy, thereby providing a robust framework for informed decision-making in disease prevention and management for pakcoy crops