Tomato Leaf Disease Detection Using TensorFlow and Keras Libraries

In India around 58 percent of India's population depend on agriculture for their livelihood, still our Indian farmers do not earn enough to live their livelihood, they suffer more than other occupations. To improve their livelihood we must improve their income, to increase income we must increase crop production. To increase crop production we must prevent disease, but Indian farmers use traditional process to cultivate the crops, they must use smart technology to improve the crop production. Modern technology helps to monitor every stage of their crop growth, health, moisture, nutrition and control sun-light. In this paper, we introduce our tomato plant disease detection model to detect diseases like septoria spot, late blight, mosaic virus, bacterial spot and so many. To prevent spreading in the whole crop field we must identify at the initial stage of disease. To detect the disease use modern technology. Here in this paper to train our model we used 10,000 tomato leaf. After that we divide into 3 part, 80 % training, 20 % validation and 20 % for testing. After training the model achieved around 95.68 % accuracy rates, testing accuracy 88.9 % and validation accuracy 88.81 %. The total size of the train model is around 254 MB.