Foundation Digital Twin Auto Feature Extraction (FDT AFE)

Request for Information (RFI)

This RFI seeks information on acquiring an innovative solution for Automated Feature Extraction (AFE) to automate the identification, extraction, and attribution of geospatial features. The goal is to gain insights into current industry and academic practices and capabilities that can meet NGA's requirements as described in the AFE Statement of Objectives (SOO). This RFI is for information and planning purposes only and does not constitute a Request for Proposal (RFP).

• Solutions must automatically detect and extract geospatial features from various sources (imagery, raster maps) using monoscopic and/or stereoscopic extraction.

• Capabilities must include individual and batch processing over multiple geographical areas.

• Solutions must attribute extracted features based on NGA domain standards and preserve source data geometry, spatial relationships, and metadata.

• Must provide confidence metrics for data validation.

• Must utilize standardized data formats (e.g., GeoDatabase, Shapefile, TDS 7.1).

• Must be compliant with applicable security directives (DNI ICD 503) and implement robust access controls.

• Familiarity with NGA policies for software development is required.

• If a future RFP is released, the envisioned period of performance is a 12-month base period with four (4) 12-month option periods.

• The primary place of performance will be the contractor's facilities.

• Responses must contain only UNCLASSIFIED information.

• Responses are due by 5:00 PM ET on April 6, 2026, submitted via email.

• Page limits apply: up to 3 pages for individual use cases and 9 pages for RFI response to questions, totaling 12 pages (or up to 27 pages total if all use cases are addressed).

• Font type: Times New Roman, 12 point.

• The document references FAR 52.219-14 (Limitation of Subcontracting) for small business set-asides.

• Responders are asked to provide information on their business type (Small Business, Small Disadvantaged, Service Disabled, 8(a), HUBZone, Woman-Owned) and mentor/protégé program involvement.

• Responders may recommend alternative NAICS codes.

• The RFI may lead to a future RFP.

• All submissions become Government property and will not be returned.

• Proprietary information must be minimized and clearly marked.

• Questions should be submitted via email to Dan Fadely and Dee Hill.

Questions & Answers / Clarifications

This document presents objective questions related to the Automated Feature Extraction (AFE) Request for Information (RFI), referencing Sections 3.0, 4.0, and Appendix A of the FDT AFE Statement of Objectives (SOO).

• Objective 1: Provide and Deploy AFE Capability: Inquiries focus on current AFE processes, methodologies for creating real-world feature representations, data types used (vector, image, PDF), ensuring accurate position and confidence, handling attribution and source data, integration with existing geospatial tools, Technical Readiness Level (TRL), and separating conflated features.
• Objective 2: Ensure High-Quality Data and Analyst-in-the-Loop Validation: Questions address ensuring relevance of extracted features, source formats, GEOINT assurance processes, analyst notification systems, prioritization of critical processes, defined threshold metrics for accuracy (80% objective, 99% goal), and interface with GIS systems, including AI/ML models for quality assessment.
• Objective 3: Enable Seamless Utilization via Open APIs: Focuses on the types of APIs (REST, gRPC), authentication methods, data formats, standardized formats for interoperability (e.g., GDB, Shapefile), and compute/infrastructure needs for the solution.
• Objective 4: Meet and Maintain Performance and Security Standards: Seeks information on scalability, reliability, resiliency, security standards, nominal operating benchmarks, and infrastructure specifications for AFE products.

This document serves as a set of detailed questions to elicit specific technical and operational information from potential vendors regarding their AFE capabilities. Responses will likely inform the development of a future solicitation or contract.

Statement of Work (SOW) / Performance Work Statement (PWS)

To acquire a contractor who can provide robust Automated Feature Extraction (AFE) solutions to modernize NGA's geospatial intelligence production by leveraging AI/ML to automate manual processes, improve data accuracy, and ensure seamless integration with critical NGA production systems. This will enhance the timeliness, relevance, and accuracy of geospatial intelligence to support national security missions.

• Acquire services or capabilities for AFE focused on features identified by NGA as important but challenging.

• Provisioning, configuration, deployment, and sustainment of a commercial/out-of-the-box AFE capability.

• Automate the identification and extraction of geospatial features from imagery and other source materials.

• Integrate AFE capabilities with key NGA systems to support maritime, aeronautical, topographic, geographic, and geomatics mission domains.

• Objective 1: Provide and deploy a high-performing AFE capability as a service, capable of detecting and extracting geospatial features from various sources (imagery, raster maps) with individual and batch processing. Features must be attributed according to NGA domain standards.

• Objective 2: Ensure high-quality data with a threshold accuracy of 80% for topographic/maritime features and 90% for aeronautical features (objective 99%). Implement automated quality assessment and an analyst-in-the-loop interface for validation and feedback to improve AI/ML models.

• Objective 3: Enable seamless utilization via open APIs, ensuring interoperability with NGA systems and data repositories. Utilize NGA standardized formats for extracted features.

• Objective 4: Meet and maintain stringent performance and security standards, including scalability, reliability, and security. The system must process a standard 1GB source file in under 10 minutes and achieve 99% availability. Deployable to Unclassified and Classified domains, complying with DNI ICD 503 and DD Form 254.

Includes features such as Apron, Bridge, Building, Cable, Canal, Cart track, Cistern, Culvert, Dam, Ditch, Electric power station, Elevation contour, Extraction mine, Fence, Ford, Gate, Helipad, Hut, Inland waterbody, Island, Military installation, Non-water well, Pipeline, Power substation, Pylons, Railway, River, Road, Runway, Shoreline, Sports ground, Storage Tank, Taxiway, Tower, Trail, Tunnel, Wall, Water tower, and Water well. Detailed definitions are provided.

Not explicitly stated in this document, but implies production environments across all required security domains.

Request for Information (RFI)

The National Geospatial-Intelligence Agency (NGA) is seeking information on innovative solutions to automate the identification, extraction, and attribution of geospatial features. This capability, termed Automated Feature Extraction (AFE), aims to modernize NGA's GEOINT production processes by leveraging Artificial Intelligence and Machine Learning (AI/ML) to reduce manual processing time, increase data accuracy, and enable rapid intelligence delivery. This RFI is focused on features NGA considers valuable but challenging.

• Solutions must automatically detect and extract geospatial features from various sources (EO, SAR, HS, MS imagery, and raster maps).

• Capability for monoscopic and/or stereoscopic extraction, handling individual and batch processing.

• Automatic attribution of extracted features based on NGA domain standards, preserving source data geometry, spatial relationships, and metadata.

• Leverage existing interface solutions for analysts to review, validate, and correct features.

• Provide confidence metrics for data validation.

• Utilize standardized data formats (GeoDatabase, Shapefile, TDS 7.1).

• Full compliance with security directives (DNI ICD 503) and robust access controls.

• Familiarity with NGA policies (Data Strategy, Software Way).

• This RFI may lead to a future Request for Proposal (RFP).

• If an RFP is issued, the envisioned period of performance is a 12-month base period with four (4) 12-month option periods.

• Primary place of performance will be the contractor's facilities.

• Responses must be UNCLASSIFIED.

• Maximum page count for RFI response to questions is nine (9) pages.

• Responses are due by 5:00 pm ET on March 30, 2026.

• Submit responses via email to Dan Fadely and Dee Hill.

• The document requests information on business type, including small business categories (Small Business, Small Disadvantaged, Service Disabled, 8(a), HUBZone, Woman-Owned).

• Responders are asked to recommend alternative NAICS codes and discuss potential for small business set-asides.

• This RFI is for information and planning purposes only and does not constitute an RFP.

• The Government will not pay for any information or administrative costs incurred in response.

• Responses may be reviewed by NGA non-Government consultants under NDA.

• Proprietary information must be clearly marked and segregated.

Questions & Answers / Clarifications

This document presents objective questions related to the Automated Feature Extraction (AFE) Request for Information (RFI), referencing Sections 3.0, 4.0, and Appendix A of the FDT AFE Statement of Objectives (SOO).

• Objective 1: AFE Capability or Service: Vendors are asked to describe their current AFE processes, methodologies for creating real-world representations of features, data type combinations used (e.g., vector-to-vector, image-to-vector), how they ensure accurate position and confidence assessment, attribution of source data, integration with geospatial production tools, and their Technical Readiness Level (TRL).
• Objective 2: High-Quality Data and Analyst-in-the-Loop Validation: Questions focus on ensuring detected changes are relevant, source formats ingested, processes for GEOINT assurance, analyst notification systems, prioritization of critical processes, defined threshold metrics for AFE outputs, achieving objective accuracy targets (90% and 99%), and integration with GIS systems.
• Objective 3: Seamless Utilization via open APIs: Vendors must describe their proposed APIs (e.g., REST, gRPC), authentication methods, data formats, recommended standardized formats for interoperability (e.g., GDB, Shapefile), and their compute and infrastructure needs.
• Objective 4: Performance and Security Standards: Vendors need to describe their solution's scalability, reliability, resiliency, and security standards, along with nominal operating benchmarks and infrastructure specifications.

This document is a set of questions designed to elicit detailed technical and operational information from potential vendors regarding their Automated Feature Extraction capabilities. The responses will inform the government's understanding and potential future requirements.

Questions & Answers / Clarifications

This document provides responses to specific use cases that exemplify the National Geospatial-Intelligence Agency's (NGA) operational requirements for an Automated Feature Extraction (AFE) solution. Vendors are expected to describe their solution's capabilities in relation to these use cases.

• Use Case Categories: The document outlines six categories of features for which AFE solutions are being evaluated:
  • Buildings (including agglomerations like facilities, military installations, or built-up areas)
  • Linear Transportation Features (Road, Railway, Cart Track, Trail)
  • Discrete Transportation Features (Bridge, Ford, Tunnel, Gate)
  • Utilities (Cistern, Pylons, Cables, Electric Power Station, Pipeline, Power substation, Water Tower, Water Well)
  • Built Water Infrastructure (Ditch, Culvert, Inland Waterbodies, Dam, Shoreline)
  • Aeronautical Infrastructure (Helipad, Runway, taxiway, apron)
• AFE Functions: Vendors must detail their solution's proficiency across three primary AFE functions:
  • Object Detection: Identifying the existence of an object in an image.
  • Geometry Extraction: Creating a vector representation of the object with real-world dimensions.
  • Attribution: Characterizing and providing informative attributes about the detected feature (e.g., categorization, measurement).
• Vendor Response Requirement: For each feature selected from Appendix A of the SOO, vendors must describe their solution's capabilities within the context of the most relevant use case category, detailing proficiency in the three AFE functions.

This document clarifies the specific technical requirements and evaluation criteria related to Automated Feature Extraction (AFE) capabilities, directly impacting how vendors should structure their responses and demonstrate their solution's suitability for the stated use cases.

Form / Attachment / Pricing Sheet

This document, the "NGA Data Strategy 2021," outlines the National Geospatial-Intelligence Agency's (NGA) plan for managing and leveraging data as a strategic asset to support its mission.

• Vision for Data: To create, manage, and securely share trusted data with speed, accuracy, and precision.
• Data as a Strategic Asset: Emphasizes data's role in fueling decision support and mission advantage.
• Focus Areas:
  • Easily Discoverable and Accessible Data
  • Improved Data Reuse and Integration
  • Cross-Domain Efficiencies
  • Next-Generation GEOINT
• Key Goals:
  • Goal 1: Manage Data as a Strategic Asset
  • Goal 2: Deliver Shared Data Services
  • Goal 3: Scale Data and Analytics Capabilities
  • Goal 4: Bolster Data Literacy in the Workforce
• Implementation Steps: Outlines specific actions for each goal, including launching governance committees, enabling data policy development, operationalizing data stewardship, and refining data literacy programs.
• NGA Data Terminology: Defines "NGA GEOINT Mission Data" and "Corporate Data."

While this document is a strategic plan and not a solicitation, it provides insight into NGA's future direction regarding data management, analytics, and technology. Understanding this strategy may help potential bidders align their offerings with NGA's evolving needs and priorities, particularly in areas like data governance, AI/ML capabilities, and data accessibility. It signals NGA's commitment to modernizing its data infrastructure and capabilities.

Other / Unclassified

This document, "The NGA Software Way," outlines the National Geospatial-Intelligence Agency's (NGA) approach to building, iterating on, and operating software products. It details key metrics and elements designed to ensure software teams have the necessary people, processes, and tools to deliver useful software faster and more consistently.

While not a solicitation, this document provides insight into NGA's internal software development and operational standards. Potential bidders or contractors working with NGA should understand these principles, as they likely influence how NGA expects its software products to be developed, deployed, and maintained. Key areas include:

• Software Metrics: Focus on Availability, Lead Time for Changes, Deployment Frequency, and Product-Specific Metrics.
• The NGA Software Way Elements: A 13-element framework covering stages from "Before Writing Code" to "As You Iterate," emphasizing user needs, rapid delivery, automation, and continuous improvement.
• Tools and Processes: Mentions NGA's enterprise-managed tools for version control (GitLab), CI/CD pipelines (Jenkins, GitLab CI), work tracking (Jira, Confluence), and monitoring (Matomo).

Form / Attachment / Pricing Sheet

This document is an Excel spreadsheet containing a template for Request for Information (RFI) questions. It is intended to guide potential bidders in formulating their questions regarding the FDT AFE opportunity.

• The template includes columns for 'Item #', 'Document', 'Requirement', 'Section', 'Comment/Question', and 'NGA Response'.
• It appears to be a structured format for submitting inquiries to the National Geospatial-Intelligence Agency (NGA).

This template provides a standardized method for bidders to ask clarifying questions about the FDT AFE opportunity. Using this template ensures that questions are organized and contain the necessary context for NGA to provide a clear and relevant response, aiding bidders in understanding the requirements and preparing their proposals.