State-of-the-art AI-enabled mobile device for real-time water stress detection of field crops

Abstract

An AI-enabled mobile device was developed for real time identification of water stress in field crops with the aim to assist crop breeding and precision irrigation management programs. The device assembles a Raspberry Pi single board computer and a digital RGB camera. A validated, state-of-the-art, transfer-learning based deep learning model GoogLeNet was deployed on-board to process the captured images and classify those into stressed or non-stressed categories in real time. The captured images and results are displayed on a custom graphic user interface. The device is also capable of predicting stress in-house from historically collected images fed as inputs to the device. The results are produced in 200 ms post-capture/feeding of the images. The device was field-tested in wheat and maize crops and pertinent stress identification accuracies were 92.9% and 97.9%, respectively. The developed device offers a user-friendly mode to identify water stress in real time for the crop breeders, researchers, and ultimately growers for timely decision making on plant health and crop yield improvement. The device framework capabilities can be extended in the future for identifying multiple problems of various biotic and abiotic stresses in multiple crops, and their cultivars grown in various agroclimatic conditions.