Citrus Fruit and Leaf Disease Detection Using DenseNet

Abstract

Citrus plants are high in Vitamin <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$C$</tex> , which has several health benefits and is used as a raw material in a wide range of agro-industrial processes. The ripple effects of citrus fruit diseases are recognized to have a considerable impact on citrus fruit production. This is why it is critical to build a detection method for these diseases. Deep learning techniques have made it possible to execute a variety of tasks linked to the recognition of leaf and citrus diseases. In this paper we have proposed DenseNet-121 model, which tries to compare healthy leaves and fruits with those infected with citrus diseases like black spot, greening, scab, and canker. This model may extract many attributes from its various layers. The results showed that the model performs well on a variety of criteria. In 50 epochs and five classes, we achieved 96% accuracy reducing degradation and vanishing gradient problems using the ImageNet pre-trained DenseNet-121 model.