Anthracnose Detection from Chili Leaf Images with Deep Learning

The main objective in plant breeding is to obtain quality crops by growing healthy plants from planting to harvest. Anthracnose is a common fungal disease that causes yield losses in plants. Anthracnose disease causes lesions on various organs and fruits of the plant and impairs the overall health of the plant. In this study, we propose a system that detects anthracnose on images of chili leaves. In the study, six different convolutional neural networks with different characteristics were analyzed using transfer learning techniques. As a result of the experiments, an average accuracy of approximately 97% was obtained. The results obtained using different models and transfer learning techniques were compared with state-of-the-art convolutional neural network architectures.