Iowa State University scientists are developing a machine learning system that could automate farmers’ ability to diagnose a range of major stressors in soybeans. The technology will use cameras attached to unmanned aerial vehicles (UAVs) to gather bird’s-eye images of soybean fields. A computer application will automatically analyze the images and alert the farmer of trouble spots. The project began with an app created at the university by an interdisciplinary team of engineers and plant scientists. The app allowed smartphone users to take pictures of soybean plants to determine if the plants suffered from iron deficiency. It required photos to be taken manually and was able to diagnose only one stressor. The algorithms for the original app are now being scaled up to develop a system that is able of identify a range of stressors from images taken from UAVs. The researchers have assembled and labeled a large image data set of healthy plants and plants undergoing stress and disease. They are using a computer program to assemble algorithms that will recognize stressors in new images that are received by the system. Unmanned aerial vehicles could be equipped with hyperspectral technology capable of detecting wavelength ranges beyond those detectable by the human eye. Such technology could combine with machine learning techniques under development at Iowa State to help farmers anticipate stress among their crops before symptoms appear. Courtesy of Arti Singh. The machine learning program could be capable of identifying a range of common soybean stressors, including fungal, bacterial, and viral diseases, as well as nutrient deficiency and herbicide injury. The use of hyperspectral imaging could allow the technology to predict the presence of stressors before symptoms even appear, giving farmers additional time to manage and control the problem, professor Arti Singh said. By the end of its three-year, roughly $500,000 grant from the U.S. Department of Agriculture, Singh said, the team intends to have completed a framework of best practices for data collection using UAVs, including optimal image resolutions and optimal heights and speeds for the UAVs. The researchers hope to seamlessly integrate data collection, curation, and analysis, leading to a system that can be applied in farm fields to detect and mitigate plant stressors in a timely manner. The approach has the potential for application in many other crops as well, Singh said. The team will make all its findings publicly available at the conclusion of the project.