ResNet50: Deep Learning Method for Automated Detection of Banana Leaf Spot Diseases

Globally, bananas are a staple food crop mostly in tropical and subtropical areas where they greatly support food security and the economy. But different leaf spot diseases present significant difficulties for banana cultivators, which can severely lower quality and productivity. Traditional disease identification techniques depend on expert visual inspections, sometimes Costly, random and susceptible to human mistake. This work intends to classify banana leaf spot diseases using a strong Convolutional Neural Network (CNN) model derived on the ResNet50 architecture. The approach emphasises on differentiating among several illness groups, including those caused by common banana leaf spot diseases and those of healthy leaves. Using an extensive data set and cutting-edge computing methods, the model is taught to identify complex trends linked with various diseases. The results show that the ResNet50 model greatly improves disease management techniques by obtaining 96% in the identification of banana leaf spot illnesses. This method supports sustainable banana output, decreases the need for hand inspections, and finally helps to provide food security. This work underlines the need of multidisciplinary cooperation in developing creative solutions for pragmatic issues in crop management and contributes to the rising body of studies on deep learning applications in agriculture.