Published Paper Details:

Crop wastage as a result of diseases is a critical issue, increasing food costs and leading to enormous losses for farmers, hence making agriculture a high-risk profession. Therefore, PLD detection is relevant to the enhancement of food production and security. PLD detection is one of the growing interdisciplinary fields combining artificial intelligence and agricultural science; it has relatively been improved with the advent of transfer learning and deep learning. At the heart of each of these advancements has been access to meaningful datasets. It reviews some of the popular datasets used in PLD detection such as PlantVillage, PDDB, XDB, NLB, LWDCD2020, PDD271, and PlantDoc. This review has identified strengths and weaknesses associated with these datasets. Highlighting issues such as class imbalance, unrealistic laboratory conditions, and variability in data collection methods. Key papers on this topic are identified to show how machine learning and deep learning models have evolved in detecting PLDs. Future research may include data augmentation methods using GANs and the creation of a standardized data collection procedure. It will therefore help in raising the research community's awareness of the status of PLD datasets, underline continuous improvement for better accuracy and reliability in AI-based PLD detection models, and finally aid in sustainable agriculture and food security.