Smart Farming: Potato Leaf Disease Detection Using Deep Learning

Abstract

Due to a variety of crop species, crop diseases, and environmental conditions, early disease detection is the most difficult task. Several machine learning approaches have been developed to make this challenge easier. Data was primarily gathered by researchers to build their model. They gathered data in a variety of ways, including manually gathering data, downloading data from Google, and obtaining ready-made data from third parties. Because they used a variety of strategies, they received varying accuracy percentages. Even though everyone tried their best to reach the utmost accuracy, no one could come up with the same outcome. In order to construct my model, I employed CNN architecture. I gathered information for this from the Kaggle dataset. Since Kaggle is open source, researchers may quickly gather the precise data they need. After putting my model to the test, I got 99% accuracy. Disease detection from the leaves is very difficult. Early Blight and Late Blight are prevalent diseases in potato leaves. Those who are identified too late harm the crop. For this farmer must deal for both money loss and potato waste.