Citrus Fruits–Leaves Diseases Detection and Classification with Optimized Deep CNN

Plant ailments are the biggest issue, affecting productivity in agriculture, causing substantial losses in revenue and instability in nutritional supply. About 140 nations manufacture and normally cultivate citrus fruit harvests, which are quite significant for plants economically. Sadly, a variety of factors, including parasites and viruses, have a significant influence on citrus growing and have caused significant crop quality and yield reductions. A fast and precise diagnosis is crucial to halt the spreading of diseases of plants and reduce crop damage. The effectiveness of several optimizers including Adam, SGD, and RMSprop for identifying and categorizing citrus plant leaves and fruits illnesses is examined in this research along with a novel AI-based deep convolutional neural network (CNN) model. This model is trained with seven different classes of plant leaves and fruits images. The performance matrix provides a visual representation of optimizers’ effectiveness. We discovered that using data enhancement can enhance the model’s accuracy. To train the suggested model, various training epochs, batch sizes, and dropouts were employed. The recommended model outperforms the other one while using the Adam optimizer. The proposed model successfully classifies cropped images with a precision of 98.6%, proving that the algorithm may be utilized as an outline for categorizing images.