Crop Disease Prediction and Pesticide Recommendation Using Deep Learning Techniques

In agriculture, crop diseases represent a meaningful threat to food security, greatly affecting farmers’ incomes in addition to causing large crop yield reductions. The proposed system presents both machine learning approach and deep learning approach for crop disease identification, plus pesticide suggestions. The system uses data from Ghanaian farms. It trains machine learning models like Random Forest, Support Vector Machine and K-Nearest Neighbors, along with deep learning models such as Customized VGG16 and Fine-tuned Customized Vision Transformer, by employing image preprocessing and augmentation. The fine-tuned customized Vision Transformer model performed best and achieved good accuracy. The model provides fast disease diagnoses and pesticide recommendations. This approach uses technology to greatly improve how crops are managed which aims to considerably increase how much food farms produce, make farming more sustainable and greatly decrease food shortages caused by severe crop diseases.