Tomato plant leaf diseases detection and classification using an advanced self-developed CNN model

Better disease detection and classification for tomato leaves at early sage provide finest productivity results. It is a natural occurrence for tomato plants to get sick, and if the appropriate attention and corrective action are not given in a timely manner. Moreover, it negatively impacts the productivity, quality, and quantity of the corresponding product. It is crucial to check the health and look for disease in tomato crop leaves since, if handled, these issues can affect the plant, leading to significant losses at marketing fields. Thus this article proposed an advanced Self-developed Convolution Neural Network (CNN) with fewer layers than the other standard model and its accuracy was tested by varying different parameters. Consequently, results shows the highest accuracy obtained at the learning rate of 0.0003, 20 epochs, and on using the three different optimizer. All the experiments are performed on 4 classes such as yellow curl diseases, early blight, bacterial spot and healthy leaf of the plant village public dataset the highest accuracy of 99.67% achieved on 12 layers CNN model.