OCEANIT LABORATORIES INC — Department of Energy SBIR Phase I: 26b

Description

The characteristics of fine roots are vital for understanding the health of an ecosystem. Researchers monitor roots with “mini-rhizotrons”, which are tubes installed in the soil into which digital imagers can be inserted. The resulting imagery can provide a wealth of information about the presence, extent, morphology and progression of root systems and other phenomena, such as fungal hyphae. The current bottleneck for analyzing such imagery is the laborious process of manually marking roots in images. Even when low cost labor e.g. graduate students) is available, the time required to label root images precludes high volume data analysis. Moreover, human labeling is subject to bias. While software is available for automatically detecting roots in mini-rhizotron imagery, in practice it is unreliable in all but the most ideal conditions. The most widely used example is the open source tool Rootfly. While this tool has been a valuable work-horse for researchers, its automatic root detection capabilities is insufficient for this task. This problem has remained unsolved, because it requires a cross-disciplinary mix of software development, image processing, plant biology, and machine learning. Research will leverage a custom designed software suite for image processing and machine learning and incorporate, expert knowledge of plant biology, and a large repository of previously collected root imagery to develop Rhizomatic, a software solution for the production of rapid, automated assessments and quantification of in situ fine root measurements. These assessments will include root length, density, root diameter, color, and branching order. The software solution will be capable of identifying and digitizing individual roots, tracking changes in the roots/images over time. It will be able to handle multiple image formats, geo-reference image boundaries and control location points, and scale 2-dimensional mini-rhizotron images to extract a three-dimensional approximation. Rhizomatic will be a commercially supported software product marketed through a license agreement with a mini-rhizotron supplier that has already expressed their support for the project and interest in the results of the research. Agriculture is a major component of the US economy, with farm output alone contributing about $100 billion annually and derivative products accounting for ten times that amount. While it is hard to evaluate the economic impact of more efficient root studies, even a very small piece of such a large market represents a significant business opportunity.