Detecting and Mapping Crop Stress using Sentinel Remote Sensing Imagery

The Australian agriculture industry is extremely important for economy. One common issue that grain farmers have to both our food supply and the deal with is crop stress which is caused by typical factors such as topography, pests and disease, water content and soil nutrients. Because these crop stress factors have the potential to affect the quality of grain, this has an impact on the grain grower’s profits. This research project aims to detect and map potential areas of crop stress within winter crops such as barley on the Darling Downs region of Southern Queensland, using Sentinel remote sensing imagery. The intention of this project is to determine the suitability of using the Sentinel 2 remote sensing imagery for crop stress detection, as both the data and software used are freely available. Specifically, the objectives of the project include acquisition of remote sensing data, utilising freely available software to process the images to create NDVI and classified images to highlight areas of crop stress, acquiring field data such as elevation maps and investigating potential factors causing stress using the data collected. The project methodology included acquiring remote sensing imagery, acquiring field and crop data from the landholder, pre-processing remote sensing imager, processing field data to create an elevation model, creating NDVI and classified images using unsupervised classification in order to depict areas of crop stress and compare field and crop data to the processed images to determine potential causes of crop stress. The results achieved using the project methodology within all field s allowed for detection of areas of crop stress investigated. It was concluded that the topography of the fields was not a cause of the stress, instead the areas were a result of either varying soil types or nutrients and moisture stress. However, there are limitations to this project such as the coarse resolution of the imagery, lack of certain field data such as yield and soil maps and limitations with the free processing software, had some minor impacts on the project. Despite this, conclusions made suggest that the methodology used to detect areas of crop stress were suitable for mainly qualitative analysis. Recommendations for future research included a comparison between Sentinel remote sensing imagery and other higher resolution platforms and a comparison between irrigated and dryland cropping