This award is for the development of software for feature recognition in minirhizotrons, which are tubes for the observation of living roots. Fine root turnover is a significant component of ecosystem carbon and nutrient cycles, but methodological challenges have limited our ability to measure root turnover under natural conditions. Recently, the use of minirhizotrons and miniaturized camera equipment has emerged as the technique of choice for studying fine root dynamics in ecological and agronomic settings. However, the time required to process minirhizotron images is excessive. In this project, we will develop open-source, feature-recognition software to assist in the capture and analysis of minirhizotron images. By improving user interfaces and incorporating algorithms used to identify blood vessels in medical images, we will automate many aspects of minirhizotron data collection.
The production and mortality of plant roots represents a significant flux of carbon and nutrients through ecosystems. Measuring rates of root turnover is therefore crucial to predicting ecosystem responses to global change. This software-development research will aid in measuring root production and mortality, and it will help train students in an interdisciplinary research area.
