Utilizing Convolutional Neural Networks for Accurate Detection of Leaf Diseases in Fava Beans

Background: Worldwide, people appreciate the variety of faba beans, also referred to as broad beans. They have many health advantages and are rich in protein and other vital nutrients. One can enhance the immune system, help manage your weight and improve your general well-being by including fava beans in your diet. This crop has been grown in several countries, including northern Europe, the Mediterranean region, central Asia, East Asia and Latin America, in addition to China, Ethiopia and Egypt. It is also grown in India as a minor crop. Many diseases, such as fusarium wilt, ascochyta blight, chocolate spot, bean rust, alternaria leaf blight, powdery mildew and root rot, can affect the production of faba beans and create a global threat. Methods: In this paper, A Sequential Convolutional Neural Network is trained for the detection of diseases present in the leaves of faba bean. Jupiter notebook with Anaconda environment is used for extraction and classification of healthy and diseased plant leaves. The dataset is collected with the help of agriculture experts. Before training and testing, the dataset is preprocessed to enhance the performance of the model. Result: The outcomes achieved after training and testing of datasets show remarkable performance matrices. The overall accuracy of the model is 98.92%. The confusion matrix, precision, recall and F1-score show the effectiveness of the proposed model. This work can be utilized as a reference for the enhancement of faba bean production.