Defense Mapping Agency, PCA, MS D-88, 4600 Sangamore Road, Bethesda, Maryland 20816-5003

A -- INTELLIGENT MAP UNDERSTANDING BROAD AGENCY ANNOUNCEMENT DUE 053196 POC Point of Contact: Contracting Officer, Leslie Vandivere, MS D-88, Voice (301) 227-2158, Fax (301) 227-2218, E-Mail vandiverel@dma.gov The U.S. Defense Mapping Agency (DMA) is soliciting proposals for research projects on the subject of intelligent map understanding/vectorization to improve methods of automated spatial data collection. The main objective of this research is the development of algorithms, techniques, and processes in automated vectorization from maps and imagery. Focused research and development efforts are solicited to advance the state of the art in the following five (5) topical areas: automated feature recognition and delineation from imagery; automated feature recognition, delineation, and attribution from map and imagery sources; automated text recognition and association from digital raster maps; automated text and label placement; and pioneering processes and technologies. A single proposal may address a portion of a topic, a single topic, or multiple topics. If the proposal spans multiple topics, however, the topic of primary interest should be clearly annotated in the proposal. Proposals from industry, independent research centers, academia, and teams of the same are encouraged. Only proposals from U.S. offerors will be considered for award. All software developed as a result of this effort must be open and interoperable to comply with DMA's Technical Architecture Framework and Implementation Guide (TAFIG), Version 1.0, October 28, 1994. The government shall have Unlimited Rights to all technical data and software which result from this research. Results of research of awarded proposals shall be written as a paper in a manner suitable for publication in a peer reviewed journal. TOPIC 1. AUTOMATED FEATURE RECOGNITION AND DELINEATION. Topic 1A. Automatic Feature Recognition, Delineation, and Attribution Using Commercial Imagery: New algorithms and advanced data compositing techniques are sought which will allow the super impositioning of a variety of image sources and enhanced computer vision tools to allow automatic feature identification, delineation, and attribution. There are two main objectives of this topic area: 1) Efficient automatic feature recognition, delineation, and attribution using the supervised classification of multispectral imagery in conjunction with SPOT panchromatic imagery over large geographic areas and 2) Generation and population of vector data and geospatial information by maximizing the use of global rules to automatically merge similar areas and detect and resolve slivers and gaps in topology thereby limiting analyst interaction and reducing extraction time. Features and attributes of interest as well as imagery sources to be used are discussed in the Government Furnished Property (GFP) section of this announcement. Topic 1b. Automatic Feature Recognition, Delineation and Attribution Using Multispectral and National Technical Means Image Sources. New algorithms and advanced data compositing techniques are sought which will allow the super impositioning of a variety of image sources and enhanced computer vision tools to allow automatic feature identification, delineation, and attribution. There are two main objectives of this topic area: 1) Efficient automatic feature recognition, delineation, and attribution using the supervised classification of multispectral imagery and radar imagery in conjunction with National Technical Means (NTM) over large geographic areas and 2) Generation and population of vector data and geospatial information by maximizing the use of global rules to automatically merge similar areas and resolve slivers and gaps in topology thereby limiting analyst interaction and reducing extraction time. Features and attributes of interest as well as imagery sources to be used are discussed in the Government Furnished Property (GFP) section of this announcement. All offerors for Topic 1B shall have approved facilities and personnel for storage and processing of classified data up to DoD SECRET level. TOPIC 2. AUTOMATED FEATURE RECOGNITION, DELINEATION, AND ATTRIBUTION FROM MAP AND IMAGERY SOURCES: New techniques are sought for automated identification and attribution of features from a scanned, color litho (in raster format) and from the integration of the color litho with high resolution, panchromatic, digital imagery. The goal of this effort are new, automated feature extraction techniques with a focus on recognition, delineation, and attribution from the rasterized, color litho and from the rasterized, color litho composited with the digital imagery. Desired output are digital, vector feature data at the maximum level of attribution possible within severe time constraints and with minimal operator intervention. Features and attributes of interest as well as imagery to be used are discussed in the Government Furnished Property (GFP) section of this announcement. TOPIC 3. AUTOMATED TEXT RECOGNITION AND ASSOCIATION FROM DIGITAL RASTER MAPS: New research is needed to develop automated techniques for the recognition of text and the correlation of the text with its associated feature. The relationship between a feature and its associated text is based on the placement of the text relative to the feature, type size and style, and interference with type for neighboring features. Areas without definitive edges or outlines are particularly difficult to capture from printed maps. Text placement plays an important part in the identification and attribution of such areas. Techniques to be developed include but are not limited to: automated character recognition; construction of words from individual characters; translation of words into semantic meaning; and association of words on the map to its corresponding feature/symbol. The GFP Section discusses the data to be used in research activities. TOPIC 4. AUTOMATED TEXT AND LABEL PLACEMENT: New techniques are sought in the field of automated text and label placement. Automated name and label placement, particularly on a product with much detail (e.g. topographic maps or hydrographic charts), has not matched the quality of manual text placement achieved by trained cartographers. The goal of this effort is the development of algorithms that provide automated name and label placement on densely detailed map or chart products which accomplish the quality of placement achieved by professional cartographers using manual placement methods. The developed algorithms shall automatically place names and labels to accomplish the same, or very near the same, quality of placement as achieved in manual name placement as depicted on the hardcopy product. The GFP Section discusses the data to be used in research activities. TOPIC 5. PIONEERING PROCESSES AND TECHNOLOGIES: Research is sought for pioneering new technology for extracting geo-coded features in the development of geospatial information. Research should pioneer new ideas with little or no boundaries and be viable. Proposed research may use any of the data provided as Government Furnished Property (GFP) or offeror may provide a unique data source for research. GOVERNMENT FURNISHED PROPERTY: With the exception of Focus Area 5, all awarded research shall use the designated GFP as test data. No other data may be substituted. Except as noted below, the GFP will be provided upon award of a successful proposal. All imagery sources will be orthorectified. Requests and technical questions regarding GFP may be submitted in writing to Ms. Michele Limoges Motsko who is the POC for GFP (address information is below). GFP is only to be used for activities and research directly in support of this announcement. At the completion of the contract, the GFP shall be returned at the contractor's expense to the address from which it was shipped. If offeror wants an additional type of data, the type of data and rationale for data shall be included in the proposal. A. The government furnished property (GFP) for Focus Area 1A is: SPOT panchromatic imagery in Raster Product Format (RPF); SPOT MSS; LandSat MSS; scanned, color map (scanned at 1000 dpi) in Group 4 compressed Tagged Image File Format (TIFF); hardcopy of the topographic map; and a list of selected features and attributes. The source data will cover approximately a 15' x 15' geographic area. The list of selected features and attributes may be requested during proposal preparation. The request should be directed to the GFP POC. B. The GFP for Focus Area 1B is: SPOT MSS; LandSat MSS; Radar imagery in National Image Transmission Format 2.0 (NITF)format ; National Technical Means Imagery in Raster Product Format (RPF); scanned, color map (scanned at 1000 dpi) in Group 4 compressed Tagged Image File Format (TIFF); hardcopy of the topographic map; and, a list of selected features and attributes. The source data will cover approximately a 15' x 15' geographic area. The list of selected features and attributes may be requested during proposal preparation. The request should be directed to the GFP POC. C. The GFP for Focus Area 2 includes a scanned 7.5 minute quadrangle (scanned at 1000 dpi) in Group 4 compressed Tagged Image File Format (TIFF); Digital Orthophoto Quadrangle (DOQ) (the distribution format of the DOQ file contains four ASCII header records, followed by a series of 8-bit, binary image data records); corresponding hardcopy quad sheet; and a list of selected features and attributes. The area covered by the provided source material is Washington D.C. The list of selected features and attributes may be requested during proposal preparation. The request should be directed to the GFP POC. D. The GFP for Focus Area 3 is the hardcopy and scanned city map as described in paragraph C above. E. The GFP for Focus Area 4 will be a digital data set in Vector Product Format (VPF), the corresponding hardcopy litho, VPFView software, basic guidance for names placement, and a list of attributes which require text or label placement. F. For Focus Area 5, the offeror may request any of the listed GFP for use in research or the offeror may provide a unique data set along with rationale for using the data set. PROPOSAL SUBMISSION: Eight copies of the offeror's proposals shall be submitted via standard mail to Ms. Leslie Vandivere (phone 301-227-2158), Contracting Officer, 4600 Sangamore Road, Attn.: PCA, Mail Stop D-88, Bethesda, MD 20816-5003 by COB at DMA by 31 May 1996. Proposals shall consist of two volumes, a technical volume and a cost volume. The cost volume shall be specific and separate from the technical volume. Cost figures shall not be shown in the technical volume. The technical volume shall present a clear and concise summary of the research approach and development objectives (including schedules and milestones), provide information on prior experience of offeror in the targeted technical area(s), and supply resume(s) of principle investigator(s). The cost volume shall contain cost information sufficient to determine reasonableness of the price or cost realism. The cost information shall include a complete cost breakdown of the proposed effort including labor rates and hours, as well as the rationale for the basis of the estimate. The cost breakdown shall be directly traceable through section/paragraph numbers to the technical volume. EVALUATION PROCESS: Evaluation of proposals will be accomplished through a peer or scientific review process. Scientific and/or peer evaluators from DMA and other Government organizations will evaluate proposals. Proposals will be evaluated using the following criteria which is listed in descending order of relative importance. (A) Overall Scientific and Technical Merit. (B) Degree to which techniques proposed are new and advance the body of research and current applications. (C) Adequacy of principle investigator(s)' qualifications, capabilities, and experience to provide technical direction for offeror's proposal. (D) Adequacy of offeror's resources (including hardware and software), capabilities, technical credentials and relevant experience to support the proposed effort. (E) The Reasonableness and Realism of Proposed Costs and Fees will be considered, but not scored or otherwise combined with the technical factors. OTHER INFORMATION: Upon completion of proposal evaluation, the offeror will be notified of selectability or non-selectability. Selectable proposals will be considered for funding; non-selectable proposals will be destroyed. Not all proposals deemed selectable will be funded. Decisions to fund selectable proposals will be based on scientific and technical merit, potential contribution to the advancement of the targeted technical topic(s), and availability of funds. For Topics 1A, 2, 3, 4 and 5, initial funding is projected to be $500K for Fiscal Year (FY) 1996 for the award of several proposals. Additional funding may be made available to fund additional work so proposals should be valid for twelve (12) months from the date of the release of this announcement. For Topic 1B, initial funding is $250K (distinct and separate from the $500K for all other topics) for FY96 for the award of two or more proposals. The Government reserves the right to select for award all, some, or none of the proposals received in response to this announcement. No request for proposal (RFP) or other solicitation will be made. Technical questions may be submitted in writing to Ms. Michele Limoges Motsko (fax 301-227-2988; e-mail motskom@dma.gov), 4600 Sangamore Road, Attn.: ATPE, Mail Stop D-85, Bethesda, MD 20816-5003. All remaining questions, other than technical questions or questions concerning the government furnished property, may be submitted in writing to the Contracting Officer, Ms. Leslie Vandivere, at 301-227-2218 (fax) or e-mail vandiverel@dma.gov. (0072)

Loren Data Corp. http://www.ld.com (SYN# 0001 19960313\A-0001.SOL)