Automated delineation of vegetation and vegetation species in Glen Canyon Natio

Notice of Intent to award noncompetitive for the development of a machine vision system for automated delineation of vegetation and vegetation species in Glen Canyon National Recreation Area.

The Department of the Interior, U.S. Geological Survey (USGS), Southwest Biological Science Center (SBSC), requires expedient analysis of vegetation imagery collected in association with a Spring Disturbance Flow (March 9-11, 2021) and from post-flow monitoring trips (March 29-31, 2021; July 21-23, 2021). The work will include the development of a machine learning model to automatically delineate regions of pixels in aquatic vegetation with the goal of identifying species and percent cover. The vendor will optimize this model using labeled image classes and then create vegetation maps that will form the basis of future monitoring efforts in Glen Canyon. This work must be completed by 5/1/2022 in order to inform a decision by the US Bureau of Reclamation on whether to implement a summer disturbance flow.

Marda Science is the only known vendor that possesses the capability to perform the work required. Marda Science created a python-based vegetation classification application that allows the user to input files, annotate vegetation classes, assign confidence levels, and generate Conditional Random Field predictions. No other commercial vendor has the capability for supplying these units in accordance with Federal Acquisition Regulation (FAR) 13.106-1(b) and the procedures of FAR Part 12, entitled, Acquisition of Commercial Items.
Other interested concerns having the ability to provide these supplies may submit a capabilities statement to the Contracting Officer via email at cjfejarang@usgs.gov no later than the closing date of this announcement on February 14, 2022. All submissions will be evaluated to determine if competitive opportunities exist.