About this document

Agriculture is the backbone of most economies in Africa,including Uganda,and groundnut is one of the five most important oil seeds produced in the world and in Uganda.The crop plays a critical role in the food and nutrition security of both animals and humans. The optimal production of the crop is challenged by several factors,including limited access to quality seed, limited access to quality extension service, poor management of pests and diseases, and weak farm records management,among others. The crop is susceptible to many diseases, causing a decline in productivity and quality, all of which affect the agricultural economy. Traditional detection methods by farmers and researchers are knowledge-intensive, time-consuming, costly,and less accurate,necessitating the need to develop newer approaches which are more reliable and usable by farmers who have limited expert knowledge.Five major foliar diseases of groundnuts include groundnut rosette, early and late leaf spots, Bacterial wilt,and rust.Therefore, the main purpose of this study was to investigate the effectiveness of Artificial Intelligence to enable mobile models to detect groundnut rosette disease. The model described in this paper was developed with and without stepwise resizing and simultaneously validated using cross-entropy loss. The dataset used for training and validation purposes was manually created. To evaluate the performance of the model, various performance metrics such as accuracy, sensitivity, F1 score, and precision were applied and achieved a perfect precision value of 100% at a high confidence threshold of 0.964 and F1 score of 0.8 at a high confidence threshold of 0.454 demonstrating the model’s balanced effectiveness in identifying groundnut rosette disease. The Groundnut Rosette Disease Diagnosis Using Plant Physiological Feature Image Processing Technique model achieved reliable accuracy with a limited dataset of 9 groundnut rosette scale rates.