You are here

Automated Camera Orientation Recovery Software

Award Information
Agency: Department of Defense
Branch: National Geospatial-Intelligence Agency
Contract: HM047620C0083
Agency Tracking Number: NGA-P1-20-23
Amount: $99,997.00
Phase: Phase I
Program: SBIR
Solicitation Topic Code: NGA201-006
Solicitation Number: 00.1
Timeline
Solicitation Year: 1900
Award Year: 2021
Award Start Date (Proposal Award Date): 2020-10-05
Award End Date (Contract End Date): 2021-07-04
Small Business Information
1845 West 205th Street
Torrance, CA 90501-1510
United States
DUNS: 153865951
HUBZone Owned: No
Woman Owned: No
Socially and Economically Disadvantaged: No
Principal Investigator
 Gregory Waligorski
 (310) 320-3088
 isproposals@poc.com
Business Contact
 Keith Baker
Phone: (310) 320-3088
Email: contracts@poc.com
Research Institution
N/A
Abstract

To address the NGA’s need to fully automate recovery of camera orientation parameters from ground-level imagery, Physical Optics Corporation (POC) proposes to develop new Automated Camera Orientation Recovery Software (ACORS). It is based on a new, multicue combination of algorithms for finding true horizon lines in images. Specifically, the innovation in locating occluded true horizon lines below visible skylines enables ACORS to perform fully automated extraction of camera pitch and roll angles and their uncertainties from individual images or video frames in which the skylines are already labeled. ACORS will also optionally annotate this imagery with the horizon lines it found and their error bounds. As a result, ACORS offers fully automated and accelerated recovery of camera orientations from batches of ground-based imagery, which directly addresses the NGA’s requirements. In Phase I, POC will demonstrate the feasibility of ACORS by identifying, implementing, optimizing, and evaluating a suitable combination of algorithms, reaching technology readiness level (TRL)-3. In Phase II, POC plans to develop a TRL-6 ACORS prototype supporting complete automation and rigorous error tracking. We will also assemble diverse training and testing datasets and develop performance metrics for varying operating conditions and sensor types, taking into account possible confounding factors.

* Information listed above is at the time of submission. *

US Flag An Official Website of the United States Government