Novel Data Analytics Approach for Crop Health Assessment in Precision Agriculture
The project proposes a novel method for image-processing of aerial images of crop field collected with unmanned aerial systems. The objective of the project is to develop a methodology for automated assessment of crop health in images by employing deep neural networks. The specific aims of the project are: (1) collect field images by using a multispectral camera carried by an unmanned aerial system, and annotate the data; and (2) design a novel neural network architecture for segmentation of crop leaves in images, and subsequently, for discrimination of healthy and diseased plants. The significance of the application is in the potential to advance the area of automated crop disease detection. Early diagnosis of crop health symptoms can reduce the volume of chemical substances applied at later phases, and contribute to increased crop yield. The long-term goal of the research is to expand the proposed work and develop image-processing methods for the related problems for detection of weed, pests, and crop stress from high-resolution field images, and for biomass estimation from low-resolution field images. The expected outcome is a novel image-processing algorithm for crop heath detection. In addition, one journal publication and preliminary results for an NSF grant application are expected.