The Department of Interior US Geological Survey (USGS) Western Ecological Research Center (WERC) has issued an RFQ for high-volume digital image processing work-flow that employs an iterative “active-learning pipeline” to obtain accurate image target detection within high-resolution aerial imagery. Ultimately, this articulated pipeline and final model architecture will allow USGS to detect objects of interest and label these to general taxonomic categories.
The USGS-WERC requires an articulated, well-documented, pipeline and final model architecture (in appropriate programming language such as Python) to detect objects of interest within our dataset and a method to effectively and efficiently use expert knowledge to label these to generalized taxonomic/behavioral/phenotypic categories.
The period of performance will start the day of award through June 2020. This is a total small-business set-aside.