Foundation Digital Twin Auto Feature Extraction (FDT AFE)

This RFI Q&A document clarifies various aspects of a government Request for Information (RFI), primarily focusing on submission guidelines and technical requirements. Key clarifications include confirmation that font sizes smaller than 12pt (but not less than 10pt) are acceptable for tables and graphics, and that the maximum page count for responses addressing all six use cases is 27 pages. Respondents are permitted to include CUI if appropriately marked. The document also specifies that individual use case responses should focus on Attachment 4 content, while the 9-page RFI response covers Attachment 3 questions and Attachment 2 content. The RFI will not be extended, and responses to each use case will be evaluated independently. Vendors are not ineligible if all personnel lack TS/SCI clearance or facilities are not TS-cleared, as most work is anticipated at the unclassified level. The government also clarified that a non-open source API is not a disqualifying factor, though its interaction with other open APIs needs to be understood.

The National Geospatial-Intelligence Agency (NGA) is issuing a Request for Information (RFI) for an innovative solution to automate the identification, extraction, and attribution of geospatial features, termed Automated Feature Extraction (AFE). This RFI seeks industry and academia input on existing AFE capabilities that can be provided as a service via APIs to modernize NGA's Foundation Digital Twin (FDT) processes. The NGA aims to leverage AI/ML to reduce manual feature extraction, improve data accuracy, and accelerate intelligence delivery. The envisioned acquisition would involve a 12-month base period and four 12-month option periods, with performance primarily at the contractor's facilities. Solutions must support various data sources (imagery, raster maps), perform monoscopic/stereoscopic extraction, handle batch processing, and comply with NGA's security and data standards. Responders are asked to provide administrative details, relevant commercial and government experience, capabilities, and recommendations for the acquisition. This RFI is for planning purposes only and does not constitute a commitment to a future RFP.

The Automated Feature Extraction (AFE) Request for Information (RFI) outlines key objectives for developing and integrating a high-performing AFE capability within the National Geospatial-Intelligence Agency's (NGA) GEOINT ecosystem. The RFI seeks information on current AFE automation processes, data sources, methodologies for creating real-world feature representations, and data type combinations used. It emphasizes ensuring accurate positioning, confidence assessment, and temporal association with extracted features, as well as seamless integration with existing geospatial tools and systems. A critical objective is maintaining high-quality data through analyst-in-the-loop validation, including change detection, source format ingestion, GEOINT assurance processes, and robust notification systems. The RFI also focuses on implementing automated change detection to maintain data currency by comparing new source data against existing feature databases and identifying modifications or discrepancies. Finally, it stresses seamless integration with NGA's GEOINT ecosystem through open and modular capabilities, API utilization, standardized data formats, and adherence to stringent performance, scalability, reliability, and security standards, including processing speed and availability.

This document addresses questions regarding an RFI for the NGA, specifically focusing on formatting, page limits, and content requirements. It confirms that font sizes smaller than 12pt are acceptable for graphics and tables, and clarifies the maximum page counts for responses based on the number of use cases addressed (e.g., 27 pages for all six use cases). The RFI also permits the inclusion of CUI if appropriately marked. It distinguishes between the content expected in the 3-page individual use case responses (focusing on Attachment 4) and the 9-page RFI response to questions (including Attachments 2 and 3). Additionally, the document corrects a reference in Section 2.0 of the RFI from Section 4.0 to Section 3.0 of the Statement of Objectives (SOO).

The National Geospatial-Intelligence Agency (NGA) has issued a Request for Information (RFI) for its Foundation Digital Twin Automated Feature Extraction (FDT AFE) initiative. The RFI seeks innovative solutions from industry and academia to automate the identification, extraction, and attribution of geospatial features, utilizing artificial intelligence and machine learning (AI/ML) to enhance efficiency and accuracy in mission data collection. The AFE capability will be delivered as a service via APIs and aims to reduce manual processing time, increase data accuracy, and enable rapid intelligence delivery. The RFI outlines a potential 12-month base period with four 12-month option periods, with performance primarily at contractor facilities and remote job submission. Solutions must handle various data sources, support monoscopic and stereoscopic extraction, attribute features according to NGA standards, and ensure compatibility with NGA systems and security directives. The document requests administrative information, details on commercial and government experience, capabilities, and recommendations for the proposed acquisition. Responses are due by April 6, 2026, and must be unclassified.

This RFI details objective questions for an Automated Feature Extraction (AFE) capability or service, aimed at enhancing geospatial data production for government use. It seeks information on current AFE automation processes, methodologies for creating real-world feature representations from various data types, and methods for ensuring accurate position, confidence, and time association. The RFI also probes how attribution and source data identification are handled throughout the AFE chain, and how solutions integrate with geospatial tools and systems. Key objectives include ensuring high-quality data with analyst-in-the-loop validation, enabling seamless utilization via open APIs, and meeting stringent performance and security standards. Specific questions address change detection, GEOINT assurance, analyst notification systems, accuracy thresholds, GIS integration, and AI/ML models for quality assessment. Additionally, the RFI requests details on API types, authentication, data formats for interoperability, and infrastructure needs, as well as scalability, reliability, resiliency, and security standards, including operational benchmarks.

The National Geospatial-Intelligence Agency (NGA) requires an Automated Feature Extraction (AFE) capability for its Foundation Digital Twin (FDT) program. This initiative aims to modernize geospatial intelligence production by leveraging AI/ML to automate manual feature extraction, improve data accuracy, and integrate seamlessly with NGA's systems. The NGA, a Department of War combat support agency, seeks contractors to provide robust AFE solutions for identifying and cataloging critical infrastructure and objects of interest from various datasets. The scope of work includes provisioning, configuring, deploying, and sustaining commercial AFE capabilities across NGA's maritime, aeronautical, topographic, geographic, and geomatics mission domains. Key objectives include delivering a high-performing AFE capability as a service, ensuring high-quality data with analyst-in-the-loop validation (80% accuracy for topographic/maritime, 90% for aeronautical, with a 99% objective), enabling seamless utilization via open APIs, and meeting stringent performance and security standards (processing 1GB in under 10 minutes, 99% availability, deployable to unclassified and classified domains). A prioritized list of features for AFE, including bridges, buildings, roads, and rivers, is provided, with detailed definitions for each.

The National Geospatial-Intelligence Agency (NGA) has issued a Request for Information (RFI) for an innovative solution to automate the identification, extraction, and attribution of geospatial features. This Automated Feature Extraction (AFE) capability will be a critical enabler for modernizing Foundation Geospatial (FG) processes, provided as a service via APIs to key NGA systems like the Foundation Digital Twin (FDT). The RFI seeks industry and academia input on current practices and capabilities, particularly focusing on challenging features for FG. This is solely for information and planning, not a Request for Proposal (RFP). The NGA aims to leverage artificial intelligence and machine learning (AI/ML) to reduce manual processing, increase data accuracy, and accelerate intelligence delivery, aligning with the NGA's 2035 GEOINT Concept of Operations.

This RFI outlines objective questions for an Automated Feature Extraction (AFE) capability or service, targeting its deployment, data quality, integration, and performance. Objective 1 focuses on the AFE's automation processes, methodology for creating real-world representations, data type combinations (e.g., image-to-vector), accuracy assurance, source data handling, integration with geospatial tools, and Technical Readiness Level. Objective 2 emphasizes ensuring high-quality data and enabling analyst-in-the-loop validation, covering change detection, ingestible source formats, GEOINT assurance, analyst notification systems, threshold metrics for AFE outputs, objective accuracy goals (90-99%), and interfaces with GIS systems. It also inquires about AI/ML models for quality assessments. Objective 3 addresses seamless utilization within NGA’s GEOINT Ecosystem via open APIs, requesting details on API types (e.g., REST, gRPC), authentication, data formats, recommended standardized formats for interoperability, and compute/infrastructure needs. Objective 4 focuses on meeting and maintaining stringent performance and security standards, covering scalability, reliability, resiliency, security standards, and baseline operating benchmarks/infrastructure specifications.

The document outlines the National Geospatial-Intelligence Agency's (NGA) requirements for an Automated Feature Extraction (AFE) solution, focusing on specific use cases to assess vendor responses. Vendors must describe their solution's capabilities across three primary AFE functions: Object Detection, Geometry Extraction, and Attribution, in the context of selected features from a provided list. Key use case categories include Buildings, Linear Transportation Features, Discrete Transportation Features, Utilities, Built Water Infrastructure, and Aeronautical Infrastructure. Each category details a rationale, example use case, and specific objectives, emphasizing the need for accurate vector representations, characterization of physical and functional attributes, and topological correctness. The AFE functions are defined as identifying objects, creating real-world dimension vector representations, and adding informative characteristics to these geometries. This RFI seeks advanced AFE solutions that can address complex geospatial data challenges for various government operational requirements.

The National Geospatial-Intelligence Agency (NGA) Data Strategy, published in September 2021, outlines a comprehensive plan to manage data as a strategic asset, enhance data discoverability, and improve analytical capabilities. This strategy is critical for NGA's mission to provide GEOINT and support decision-making for warfighters and policymakers. Key goals include managing data as a strategic asset through a federated governance framework, delivering shared data services by breaking down data silos, scaling data and analytics capabilities by coordinating initiatives, and bolstering data literacy across the workforce. The strategy emphasizes modernizing data architecture, improving data sharing via APIs, and leveraging data for AI and machine learning initiatives. It aims to transform NGA into a data-driven organization, ensuring data is trusted, accelerated, and shared to maintain a competitive advantage.

The NGA Software Way is an implementation guide that supplements NGA's Technology Strategy, providing specific guidance for teams building and operating software. Its main purpose is to accelerate and standardize software delivery within the NGA. The document outlines 13 elements and 4 key metrics: availability, lead time for changes, deployment frequency, and product-specific metrics, collectively referred to as the “3+1 metrics.” These metrics are based on DORA research and aim to balance speed and stability in software development. The 13 elements are categorized into phases: Before Writing Code, When Beginning and Throughout Development, By the First Deliverable that Provides Value to a User, and As You Iterate. Key elements include identifying a product manager, understanding user needs, prioritizing rapid value delivery, using version control, automating testing and deployments, documenting APIs, providing support, automating monitoring and alerting, using data for decisions, and continual iteration. The NGA Software Way applies to all software products, new or existing, and emphasizes the use of NGA’s enterprise-managed tools like GitLab, Jenkins/GitLab CI, Jira, Confluence, and Matomo. The CTO's office is accountable for its adoption and success.

The document, titled "FDT AFE RFI Questions," is an unclassified government file likely associated with a Request for Information (RFI) process. It appears to be a structured template or a record of questions and corresponding responses related to a specific requirement, possibly from the National Geospatial-Intelligence Agency (NGA). The file's main purpose is to document inquiries and answers, indicating an information-gathering phase for a federal acquisition, grant, or similar government procurement initiative. It consists of a table format with columns for item number, document, requirement, section, comment/question, and NGA response, suggesting a formal process for clarifying details and addressing concerns related to a project or solicitation.