Plantech – Android-based Mobile Application for Plant Disease Detection (PDD) using EfficientNet-B3

Nowadays, technological evolution is foreseen in the agricultural field but it is not as rapid as industrial evolution. Plant disease remains a crucial issue affecting food security, plant health, farmer’s income and so on. Therefore, an AI-integrated mobile application named Plantech is developed to detect plant diseases either through real-time image capturing under natural environment or through image uploading. Besides, Plantech tends to provide the greatest advantages in the automation of the entire manual plant disease detection processes and concurrently minimizing its required duration. Apart from this core functionalities, there are some additional features implemented which are real-time location detection, weather forecasting, related news suggestion and others. Most importantly, the users must be registered before experiencing Plantech and this information will be stored within a database for further retrieval purposes. In Plantech, EfficientNet is the algorithm used for training purposes. Regarding its datasets, a leaf or non-leaf dataset will be responsible for leaf or non-leaf classification while plant disease dataset will be used to classify the plant diseases. Regarding the performance, EfficientNet managed to achieve at least 99% of accuracy and an average of 99% of F1-score.