---------- NIMA ----------

4 Phase I Selections from the 02.1 Solicitation

(In Topic Number Order)
CHI SYSTEMS, INC.
Gwynedd Office Park, 716 N. Bethlehem Pike, Ste 300
Lower Gwynedd, PA 19002
Phone:
PI:
Topic#:
(215) 542-1400
Dr. James Eilbert
NIMA 02-001      Selected for Award
Title:Imagery Exploitation Applications of Neuroscience
Abstract:The primary objective of this effort is to develop a robust object matching system that can be applied to a variety of image exploitation needs. The analytical decisions by our proposed context aware, neural column-based image expert (CANCIE) will utilize context knowledge or expectations about the current "mental space," as well as imagery having an object matching component and an analysis component. CANCIE will extract and compare information about objects or a group of objects in imagery using innovative neuromorphic techniques. In particular, it will emulate the static organization of initial visual field response within cortical columns, as well as the evolution of the response pattern as it is modified by feedback interactions. The principles encoded in the columnar organization that will be utilized in CANCIE include retinotopic organization, the multiplexing of information in individual neurons, and the trend toward larger visual fields with various types of spatial invariance farther along the visual pathway. CANCIE will support both a change detection mode and a context-sensitive, object-matching mode. An interface will be provided that will allow an image analyst to specify a set of context parameters that will impacts it processing. The need for systems that can reduce the workload of image and geospatial analysts is growing, and presents a variety of market opportunities. While the approach we propose is general, a CANCIE system would require tailoring to the specific domains, tasks, and imagery involved in a particular application. A product based on our approach could either be a stand alone system or a component of a more comprehensive image processing or GIS environment. Our market strategy is to initially concentrate on small military markets with a need for a specialized version of the CANCIE system to be developed in Phase II. This first application would provide an excellent proof of concept. Our long-term strategy is to work with a GIS vendor who is open to including third party components with their system, such as Autodesk or ESRI. Our long term goal is to incorporate CANCIE into a automated associate for a human imagery analyst that can improve both the quality and the speed of the analyst's work. CHI Systems is committed to Phase III commercialization of all Phase II SBIR products which it develops. In order to pursue commercialization of past products (which has been quite successful as indicated in the attached Company Commercialization Report), CHI Systems has engaged in many different types of relevant business development activities including applying for patent protection, hiring marketing consultants, arranging alliances with other companies, and forming a subsidiary company. We intend in the course of the proposed Phase I development of the CANCIE system to develop an appropriate, detailed commercialization plan to guide Phase II and III efforts toward production of a commercial product.

VERIDICAL RESEARCH AND DESIGN
922 South Third Avenue
Bozeman, MT 59715
Phone:
PI:
Topic#:
(406) 522-9045
Dr. Frank M. Marchak
NIMA 02-001      Selected for Award
Title:Neurophysiological Based Methods of Guided Image Search
Abstract:Complex analysis of intelligence imagery is crucial to the missions of intelligence organizations, yet remains constrained by labor-intensive, time-consuming visual search of large volumes of imagery. Many algorithms have been developed to automatically identify regions of interest in large, complex sets of imagery, yet the utility of such algorithms is limited by the fact that human analysts detect features in imagery with higher accuracy than existing methods. We propose to develop a new model of visual feature detection, Neuronal Synchrony Model, based on neurophysiological models of temporal neuronal processing, to improve the accuracy of automatic detection of features of interest in complex natural imagery. The Neuronal Synchrony Model of image feature detection will be applied to accurately identify and highlight regions of images that contain target features, thus automating the labor-intensive, "scanning" portion of imagery analysis. The accuracy of the Neuronal Synchrony Model will be tested with natural images containing visually controlled, synthetic targets as well as with natural targets using a variety of overhead imagery background and target types. The output of this effort will be a proof-of-concept demonstration of the effectiveness of this model in enhancing the speed and accuracy of interactive, guided visual search of representative imagery. Anticipated benefits of this effort are increased accuracy and speed of processing of large volumes of imagery for the purposes of identifying objects of interest, including intelligence targets. To the extent that the accuracy of the Neuronal Synchrony Model can be tuned to meet or exceed the level of human visual feature detection, labor intensive "image screening" tasks can be automated, and the imagery analysis expert can be guided to focus on only those regions of likely interest. Automated screening of visual imagery for target features of interest can also be applied, with the same benefits, to domains such as radiology, land management, and industrial applications (e.g., quality control).

PHYSICAL OPTICS CORPORATION
20600 Gramercy Place, Bldg. 100
Torrance, CA 90501
Phone:
PI:
Topic#:
(310) 320-3088
Dr. Eugene Levin
NIMA 02-002      Selected for Award
Title:Multi-level Frame Technology for Digital Elevation Data Generation
Abstract:The National Imagery and Mapping Agency is seeking new and innovative technology for generating Digital Terrain Elevation Data Level 2 (DTED-2) for territories above 60 degrees N. Important criteria include vertical accuracy of 16 m and horizontal accuracy of 20 m, and identification and selection of available cost effective imagery sources that will require a minimal number of ground control points. In response, Physical Optics Corporation (POC) proposes to develop the novel Multilevel Frame Technology for Digital Elevation Data Generation (MF-TEC). The key element of MF-TEC is Russian satellite imagery, which is extremely accurate because of its unique metrology, already covers the territory of interest, and is exceptionally cost effective. The proposed technology includes frame imagery phototriangulation from a minimal number of ground control points, and accurate DTED generation based on frame-and-fill imagery. Frame-and-fill imagery upgrades high-resolution imagery with weak geometry to ideal geometry by frame imagery. Innovative methods for quality assurance at each stage of MF-TEC are also proposed. To integrate MF-TEC into NIMA GIS and photogrammetric environments, POC proposes to develop a system of uniform metadata formats, and to ensure compatibility at the software level by means of modern software engineering. Successful integration of MF-TEC with NIMA technologies for DTED generation above 60 degrees N will increase productivity and accuracy. The proposed technology can be customized to run on GIS, photogrammetric, and remote sensing workstations, and will be of value to many government agencies for terrain model generation.

VEXCEL CORPORATION
4909 Nautilus Court
Boulder, CO 80301
Phone:
PI:
Topic#:
(303) 444-0094
Mr. Marty Marra
NIMA 02-002      Selected for Award
Title:Innovative Technology for Generating Digital Terrain Elevation Data Level 2 (DTED-2) Above 60 degrees N
Abstract:Vexcel Corporation proposes to address challenge of providing DTED-2 elevation data in the "north of sixty" latitude region of the earth using civilian spaceborne synthetic aperture radar. SAR data is well suited to complement the already collected Shuttle Radar Topographic Mission (SRTM) data and does not suffer from weather conditions that limit conventional EO systems. Based on prior work, and knowledge of the SRTM processing approach, we feel that civilian SAR data can produce the desired products from civilian SAR sources. Under this SBIR we will determine the inherent accuracy possible using archived ERS Tandem data and compare this to the accuracy possible from new data acquisitions from future systems and missions. This analysis will include consideration of new and innovative approaches for combing data for improving accuracy and reducing the need for ground control. We will demonstrate the performance over two NIMA test site. Development of new technology for improved mapping of elevation will benefit Vexcel in several areas. The first will be in its ability to serve a growing government demand for such information. The second will be in its ability to bring similar products to the private and civilian sector using existing and future SAR sensors. Vexcel will be in a much improved competitive situation with respect to these two areas given a successful SBIR program. Furthermore, we feel that a program for carrying out the topic of this SBIR, i.e. producing DTED north of sixty degrees, will become a funded effort from the US Government. Through completion of this SBIR, Vexcel will improve its ability to perform a portion of this work.