Machine Learning based Potato Leaves Disease Detection

The primary source of food, money, and employment for rural residents in economically developing nations is agriculture. Crop loss caused by plant diseases, which reduces production by 20 to 30%, is the main factor affecting agriculture productivity. Conventional methods have been used to diagnose the diseases in an attempt to prevent such losses, but they are inaccurate. To avoid losses brought on by such illnesses, accurate and prompt detection of plant diseases is vital. But occasionally those harvests and grains suffer a significant amount of damage, if not complete destruction, due to a lack of suitable cultivating knowledge, expertise, and sense of disease prediction. So, in order to lessen the loss caused by infections of plant leaves, this research attempts to integrate a portion of agriculture with the use of artificial intelligence. We used CNN transfer learning models such as VGG16, VGG19, and InceptionV3, to overcome this issue. To determine which strategy performs best at identifying potato leaf illnesses, we conducted trials using all three approaches on the standard dataset of potato leaves.