Classification of Corn Leaf Disease Using Convolutional Neural Network

Corn is a crop that plays a major role in food supply worldwide. Known as a cereal crop with high economic value, corn is one of the most important raw materials in the agricultural industry in many parts of the world. Leaf blight is characterized by small spots that gradually enlarge and turn brown. It is a decay of foliage caused by the fungus or species Rhizoctonia solani. Leaf spot is caused by the fungus Hel-minthoporium maydis, while stem rot is caused by Fusarium granearum. From these problems, a machine learning-based solution is given to classify corn leaf diseases using the Convolutional Neural Network (CNN) algorithm. CNN are used to classify corn leaf diseases. The selection of CNN is based on its ability to extract local attributes from image data and combine them for a more detailed and abstract representation, which is better. Classification was performed using 2145 datasets for leaf blight and 1574 datasets for leaf spot. The accuracy results obtained from this study reached 99% with the last training accuracy value of 99.06% and the last validation accuracy result of 98.50%. For future research may use more modern architectures such as classification using EfficientNet B3 architecture with transfer learning or MobileNet to improve accuracy results.