Promise of Spectral and Signatures Understanding Todd Hawley

Promise of Spectral and Signatures Understanding Todd Hawley Sean Acklam (Spec. TIR) Technical Director Signatures Technology Fellow National Signatures Program

Promise of Spectral n n n New sensor systems Novel spectral analysis Geo-database population Imagery fusion Information visualization Real Life Applications 03 April 2007 Todd Hawley 2

Biomass Overlay Relative Biomass High Low 03 April 2007 Todd Hawley 3

Relative Turbidity High Low 03 April 2007 Todd Hawley 4

Bio-Mass / Material Discrimination Study Hyperspectral data enables automated identification of roof types including discrimination of red asphalt shingles from terra cotta roofs. Natural RGB Density Map Combined Wildland Vegetation Density Analysis (determine fuel availability for wildfire and vegetation types) 03 April 2007 Todd Hawley 5

Urban/Vegetative Boundary Vegetation, permeability, and soil moisture mapping - Colorado Springs, CO 03 April 2007 Todd Hawley 6

Wetlands Principal Component - Unsupervised Classification Forested Wetlands – MD Eastern Shore – February 2006 03 April 2007 Todd Hawley 7

Geologic/Energy n n Mineralogy Geothermal Oil/gas exploration Mining remediation 03 April 2007 Todd Hawley 8

Superfund Site 03 April 2007 Todd Hawley 9

Land Use n Forest fire projections q n n Fuel abundance mapping Invasive species Land cover/land use q q Nutrient value Agriculture/stock Urban mapping Impervious surfaces 03 April 2007 Todd Hawley 10

AG Site in Mid-West 0. 5 meter spatial 5 nm spectral Mosaic of two lines 03 April 2007 Todd Hawley 11

Road Analysis Results Correct: half of road is degraded Correct: Recently paved Correct: Moderate road with slight cracks Correct: 03 April 2007 Spec. TIR Proprietary Severely degraded road Todd Hawley 12

HSI & LIDAR Integration Adding unique elements of hyperspectral imaging and material classification… …to LIDAR-derived topographic, very high resolution topographic information, yields unprecedented level of terrain information 03 April 2007 Todd Hawley 13

HSI & LIDAR Integration Paved Asphalt / Gravel (Lots) Paved Asphalt (Streets) Tar Roof Vegetation Sandy Soils Metallic Roofs 03 April 2007 Todd Hawley 14

Natural Disaster Preliminary analysis of spectral anomalies associated with hurricane damage. 03 April 2007 Todd Hawley 15

Paper Industry Near Infra. Red (NIR) spectral camera together with multiple fiber optics is used to acquire snaphot moisture profiles across paper web in paper machine. 03 April 2007 Todd Hawley 16

Textile Dyeing A four-point fiber optic spectrometer measures dyed color at a resolution of <0. 2 DE. Pictures: Coltex system by Iris DP 03 April 2007 Todd Hawley 17

Pharmaceutical Industry n NIR spectral imaging expands the capabilities of single point near infrared spectrometry to fully cover the material and product streams under inspection q q Inspection of chemical composition and its homogenity Detection of foreign pills 03 April 2007 Todd Hawley 18

GDB Product Examples Airfield Products • • • Airfield threat assessments Airfield line diagrams Airfield graphics Airfield image maps Landing zones Helicopter landing zones 03 April 2007 34 Dimensional Products • • • 3 D visualization 3 D flythroughs Elevation tints Line of sight/intervisibility Anaglyphs Viewsheds Interferomograms Foliage penetration Time sequencing Video feeds Todd Hawley Hydrological Products • • Littoral studies Threat predictions Predictive modeling Damage assessments Turbidity Ports of entry Beach landing zones 19

GDB Product Examples • • • Engineering Products Power distribution Water distribution Sewage distribution Petroleum distribution Utility isolation Damage assessments Threat prediction Lines of communication Communication studies City construction/public works 03 April 2007 • • • Domestic Security R&S reporting graphics Key infrastructure Trends & tactics Predictive modeling Threat predictions Damage assessments Crowd control Event security Dignitary security detail Raid graphics Todd Hawley • • • Earth Sciences Geological studies Vegetation studies Environmental hazards/impacts Terrain categorization Littoral studies Geophysical studies 20

GDB Product Examples • • 03 April 2007 Mobility Products Orientation Products Route analysis Road isolation Choke points Bridges/ ford/ tunnel studies Trafficability Traffic Rate MCOO/ COO Route studies • • Todd Hawley Urban – orientation City graphic City image map Targeting map Gridded matrix Geo – orientation Change detection 21

GDB Visualization 03 April 2007 Todd Hawley 22

Network Analysis 03 April 2007 Todd Hawley 23

Data Fusion Wavelet Fusion Tool: The wavelet fusion tool developed for the STF transforms any geographically linked data into wavelet space (sparse transformation) thereby decorrelating their coefficients, applies a fusion rule to the transformed data sets (dependent on internal geometry), and performs an inverse wavelet transformation on the newly fused datasets. The outcome is a fused dataset independent of wavelength and platform. Input Data Wavelet Transform Fusion Rule Inverse Wavelet Transform 03 April 2007 Todd Hawley 24

Data Fusion Independent Component Analysis Tool: The ICA tool uses fused data and transforms into a space where components within the datasets can be isolated for statistical independence from the rest of the dataset. Non independent data like Gaussian noise is “ignored” through the use of negentropy approximations as opposed to kurtosis during the transformation process. These independent components (or vectors) act as unique and accurate signatures for any future classification and feature extraction. 03 April 2007 Todd Hawley 25

Data Fusion Generalized Relevance Learning Vector Quantization Tool: The GRLVQ tool is a hybrid classification driven feature extraction that uses the input independent components to classify, discriminate andor identify features of interest and extract the features into a geo-database as a geographic information system. The final geo-database format, containing all inclusive signatures of the urban area of interest, can be used for a wide variety of analyses and products. 03 April 2007 Todd Hawley 26

Information Visualization Axis: Unit: Temporal Time Spatial Meters Wavelength Energy Nanometers JGHz Three Modules for Information Visualization: Three modules define the signatures visualization tool. The first module is data input from the absolute geometric database. Data is represented in the form of clouds and is categorized via six different relationship types. These data clouds are projected using the second module made up of four axes depicted below. Finally, collection gaps and complete signature coverage are visualized by depicting collection asset coverage over the data clouds projected using the four axes. 03 April 2007 Todd Hawley 27

Signature Visualization Tool (Exemplar) Axis 1. Temporal 2. Spatial 3. Wavelength 4. Energy Data Signatures Spectral Signatures Behavior Signatures RADAR Signatures TTP signatures IR Signatures Data relating to any sensorobservableevent can be loaded into the SVT through XML. 03 April 2007 Todd Hawley 28

National Signatures Program Broad-based program to improve signature management & application Multiple Web-Based Operation Providers What is NSP? Diverse Signatures Infrared Users Secure Networks Spectral Infrared Radar Data Acoustic Data What Does NSP Provide? Who are Key Players? 03 April 2007 Platform for analysis & decisions Multi-community participation Unified access to diverse, distributed signatures Operation on classified networks Defense Intelligence Agency National Ground Intelligence Center Data providers Senior steering group Todd Hawley 29

Signatures n n n Features characterize targets, threats, … q Unique, consistently reoccurring Multiple signature domains q Traditional (radar, radio frequency, electro-optical, geophysical, nuclear, materials) q New domains (e. g. , chemicals) Multiple data types q Measurements, computational predictions q Spectral, time series, images, etc. Spectral Infrared Image Radar Data Acoustic Data 03 April 2007 Todd Hawley 30

Objective Improve signature management and application by balancing data users’ and providers’ needs. n User perspective q q n Simple to use one stop shopping Common view of the nation’s signature pool data Clear, definitive search results Downloadable real data Provider perspective q q Provide data quickly & efficiently to many users Maintain visibility as the source for hosted data Control data content and quality Define and control data access 03 April 2007 Todd Hawley 31

National Signature Pool Data Need s on ti era Op st Te User Communities ce en g i ell ion Int lat u Sim & t g en lin m e d ern Mo ov G S r. U e h Ot NSP Multiple Providers Common View National Signature Pool 03 April 2007 v &E n tio a alu IR Domain Acoustic Chem Bio etc. Ta r ge t Aircraft Vehicles Ships Facilities Radar Provider Acoustic IR Provider Radar Seismic. MS/HSI Chemical Biologicaletc. Vehicles Ships Aircraft Facilities Mat’ls Chem’s Todd Hawley etc. 32

Operational Overview NSP Signature Providers Local Database NSP Common DC Tgt Spct etc. A B C C xxx yyy 8 -12 3 -5 xyz zzz xxx X 3 -5 yxz xyz zyx Community Wide Metadata Local Metadata Data Metadata Files Data Target Spect. etc. Set X 12 R 1 Y 5 S 2 3 4 Z X 3 Find Results Data Summary Target XYZ Data IR Location A Date 6/1/01 John Doe POC 123 -1234 Users File Download Dynamically Generated XML Data Summaries Retrieve Descriptions Thumbnail Images File Downloads (Metadata) (when available) (Signature data) T U Signature Files Data Loading 03 April 2007 X NSP Web-Based Application Data Location/Retrieval Todd Hawley 33

Key Components n Dynamic signature operations: immediate on-demand access to all sources of quality assured standardized signatures and related data maintained with DOD, IC, and OGA to support sensor reprogramming in a highly fluid environment Signature Based Tip-off/Cross Cueing n n Signature Based Direct Reporting Machine-to-Machine Signature Exchanges Signature support plan (SSP): Potential observable signature types associated with each critical element of an activity, event, or equipment withing a specific mission area Operational signature package (OSP): End user defined selection of operational signatures, their specific ordered integration, and the desired time sequencing required to support a specific mission area 03 April 2007 Todd Hawley 34

SSP/OSP Development n Gather information q Identify friendly aircraft signature requirements n n q n Check existing signature holdings Draft q Sensor specific, enemy aircraft OSP n n q n Sensor and technique used Required signature fidelity and format Add required signatures to NSP holdings, or Generate requirement for needed signatures Incorporate into air engagement SSP Finalize q q User validation/review Assess signature support 03 April 2007 Todd Hawley 35

Estimation Environment Modeling & simulation (M&S) tools to estimate signatures for environments and collectors not specifically available in NSP measured signature holdings NSP Customer Communities NSP Distributed Signature Modeling Centers NSP Distributed Signature Data Centers Modeled Signature Holdings 03 April 2007 Measured Signature Data Holdings Todd Hawley 36

Visualization Tool Gaps Platform Sensor Location Coverage Phenomenology Signatures data relating to any collections asset can be loaded into the signatures visualization tool 03 April 2007 Todd Hawley 37

Gap Identification Optimized Signatures Gaps Energy Current Collection Architecture o Temp ral Spectral Lower weighted Relevant Background al ti a Sp 03 April 2007 Todd Hawley 38

Summary n n n Target and background signature understanding is vital to achieve the promise of spectral technologies NSP is a one-stop federated signatures and signature data source To be versatile, signature data must be q q q n Measured Integrated Accessible by analysts and developers NSP is ready to accept, integrate, and provide relevant signature data 03 April 2007 Todd Hawley 39