SpecTIR Hyperspectral Airborne Experiment 2012

Overview

The SpecTIR Hyperspectral Airborne Experiment data campaign was a successful collaboration between DIRS Lab and industry partners, aiming to provide a ground truthed dataset to the remote sensing community.

On September 20th, 2012, the SpecTIR Hyperspectral Airborne Experiment (SHARE) 2012 data campaign was conducted. This initiative was led by the Digital Imaging and Remote Sensing (DIRS) Lab in collaboration with industry partners, and was carried out in and around Avon, NY. The main objective of this campaign was to gather and distribute a highly reliable dataset to the remote sensing community.

The documentation for this dataset can be accessed here.

  • RIT WASP (Vis, SWIR, MWIR, LWIR)
  • Kucera Leica ALS-60 LIDAR
  • SpecTIR HSI
  • MITRE polarmetric HSI video oblique
  • 2 Radiosonde launches
  • Weather station
  • Spectral Ground Truth
  • Ground Based LIDAR
  • GPS & Photo Documentation

The Digital Imaging and Remote Sensing (DIRS) Lab is a part of the Chester F. Carlson Center for Imaging Science at the Rochester Institute of Technology.

Sensors

Wildfire Airborne Sensor Program (WASP)
WASP's mission is the detection and monitoring of wildfires from an aircraft at an altitude of up to 12,000 feet. WASP is adaptable for other Infrared or visible remote sensing applications as well. It consists of three infrared cameras and one visible camera, and has the capability to geo-reference and generate data products and then send them down to the ground station while the aircraft is still flying over the target area.

twin-engine Piper Navajo airplane

WASP Aircraft (Piper Navajo):
A twin-engine Piper Navajo airplane used to fly WASP Instruments

WASP has three Indigo Phoenix infrared imagers and one Geospatial Systems KCM-11 high-resolution visible camera mounted in the sensor head that looks down through a hole in the aircraft. It also has positioning devices to determine the exact position and orientation of the imagers and aircraft when each image is captured. The platform has two holes allowing for simultaneous collection from a second instrument, such as the Leica LIDAR system.

infrared cameras

Infrared spectral bands:
WASP's infrared cameras cover three different ranges of the infrared spectrum

The IR images cover three bands in the infrared; short-, mid- and long-wave. They acquire 640x512 14 bit images. The KCM-11 acquires 11-megapixel color images over roughly the same space on the ground.

WASP sensor head
WASP Sensor Head (Installed):
The sensor head containing the cameras an IMU installed in the aircraft just before flight.

Speed and Resolution
The entire assembly can acquire one stack of images every 2 seconds. At a notional ground speed of 130 knots, approximately 4,600 acres/hr can be imaged depending on flight line configurations and ground resolution requirements.

The image data acquired can also be passed through a geo-rectification process in real time. All four cameras are also aligned and correlated together so that the 6 bands of image data acquired (3 IR, Red, Green, Blue) can be processed together. This multi spectral image stack allows the end user many options for the observation of ground based events.

The images are associated to each other using positioning measurements taken from IMU and GPS sensors processed by an Applanix position and orientation system. Geo-rectification software uses this data, along with a DEM of the local terrain, to correlate the imagery to precise locations on the ground.

Specification Table

The ProSpecTIR VS system was flown by SpecTIR, LLC. The system is a hyperspectral pushbroom spectrometer with 356 bands, coving the 400 - 240 nm range. The sensor looked from nadir, and had a 24 degree field of view. It flew all sites, and covered the Avon ground truth site in both the morning and afternoon. The flight altitude and gain varied only over the Conesus Lake site; the altitude was lowered and the gain increased to increase the image fidelity over water. The quantization of this system varies according to the spectral range, with the VNIR having quantization of 12 bits and the SWIR having quantization of 14 bits.

overhead image of Conesus Lake site

Project Report
Specification Table

Kucera International flew their Leica ALS60 LIDAR instrument. This sensor, pulsing at 1064 nm was flown over all sites in the same aircraft as the WASP sensor. The ALS60 is a descrete LIDAR system using a pulsed oscillating mirror configuration. This system has 8 bit of quantization in the intensity channel. It was flown looking at nadir, and had a nominal field of view of 20 degrees, except at theHemlock site, where high topology created a field of view of 12 degrees. This site also had a higher effective ground coverage, with 12 points per meter squared, in contrast to the other sites, which had coverage of 8 points per meter squared.

overhead view using LIDAR sensor

Specification Table

The Polarimetric sensor was flown over the Avon site only and was flown from around 2-4 pm to capture polarimetric signatures form cars and other objects that typically manifest in higher sun angles. This was the first flight for the sensor, and the data should be treated as highly experimental. In addition, the geolocation unit failed during the flight, so the data location should be treated as a best guess where available. This system is a full motion, hyperspectral, polarimetric imager tested by the MITRE Corporation. The system was flown with a look angle of approximately 45 degrees and a field of view of 20 degrees. It was flown in a racetrack configuration over the site at both 2000 ft and 4000 feet above ground level. It has a quantization level of 8 bits, and its GSD was between 103 and 206 cm, depending on the portion of the frame.

Specification Table

Through the cooperation of the Naval Research Lab, GeoEye and DigitalGlobe, three satellites were able to capture scenes of the Avon ground truth site throughout the day of the experiment. The three satellites: WorldView 2, GeoEye-1, and HICO all collected a single image of the site during the day. These images have much larger GSDs and were taken while people were working in the field, offering some interesting unknown targets. Download instructions may be obtained by contacting Ms. Nina Raqueno.

Specification Table

Ground Truth

The Analytical Spectral Devices FieldSpec Pro has a spectral range of 350- 2500 nanometers with a sampling interval of 1.4 nanometer from 350 - 1000 nm and 2 nanometers from 1000 -2500 nm. The Full width, Half max (FWHM) varies over the range from 3 nanometers at 700 nm to 10 nm at 1400 nm to 12 nm at 2100 nm. The fieldSpec Pro contains 3 detectors. One is a 512 element VNIR silicon photodiode array covering the 350 - 1000 nm range. The other two are separate TE cooled, graded index SWIR InGaAs photodiodes with complete the 1000 to 2500 nm range. The input to this system is a 1.4 m fiber optic light guide upon which can be fastened a variety of foreoptics. This experiment made extensive use of the 3 degree foreoptic, as well as the cosine diffuser and the high intensity contact probe was utilized in the lab for post collect measurements. This instrument, under perfect lighting conditions was capable of taking a measurement in just under 3 seconds, and was run off of wall power or external NiMH rechargeable cells over the course of the data campaign.

In-Scene Target Materials
Analytical Spectral Devices FieldSpec Pro

The HR-1024 from the Spectra Vista Corp. (SVC) is their newest high performance single-beam field spectroradiometer measuring over the visible to short-wave infrared wavelength range (350-2500nm). The HR-1024 features include low noise indium gallium arsenide (InGaAs) photodiode arrays for the SWIR spectrum, higher resolution throughout, a robust compact light weight housing and wireless communication to laptop or PDA computers. The HR-1024 can be operated through a netbook computer in the field with full graphic, data storage and functional control. Three interchangeable fore optic lenses are included with the system together with a fibre optic light guide giving a wide range of fields of view.

Spectra Vista Corporation HR1024 Spectrometer

Global Positioning System (GPS) information was collected for each target of interest at the Avon site by a dedicated team. This team carried a Trimble GeoXT GPS receiver and a camera with a fish eye lens to collect the locations of each measurement taken during the collect as well as context imagery in 6 directions: North, South, East, West, Up, and Down. The data, once post-processed, has an accuracy of less than 1 meter.

GPS

Trimble GeoXT GPS Receiver

The radiosonde system used during this experiment was on loan from The College at Brockport. The system was an InterMet Systems iMET 3050 system consisting of a conventional Laptop computer for control, a signal decoding unit, and an omnidirectional RF antenna. The system was launched twice during the project: first before the start of flights in the morning and second after the completion of all morning flight lines over the Avon site. Both launches were successful, reporting temperature, pressure, wind speed, and wind direction to about 30 mb. The image below shows the three dimensional tracks of both launches, with the morning launch in yellow and the afternoon launch in red. Both launches tracked north east from the original location due to north and north easterly winds being prevalent through virtually the entire atmosphere that day, with little speed shear and virtually no directional shear.

google earth image

Ground observations were collected over the course of the experiment in Avon at one location on site. The observation station, on loan from the Department of Energy, was an automated station with 6 sensors. The 6 sensors collected temperature, pressure, humidity, wind speed, wind direction, broad spectrum solar irradiance, and broad spectrum thermal irradiance at 5 minute intervals from approximately 8 am local time until approximately 4 pm local time. The data was processed using the calibration constants for each instrument after the experiment. The irradiance thermal irradiance measurements required that the instrument casing temperature stabilize before accurate measurements could be taken, and as a result the thermal measurements are nonexistent for much of the morning. This information does exist as part of the spectral measurements taken during the experiment, but at 10 minute intervals.

Publications

Rey Ducay, David W. Messinger, "Leveraging spatial content of images to enhance hyperspectral-multispectral fusion performance," Proc. SPIE 12519, Algorithms, Technologies, and Applications for Multispectral and Hyperspectral Imaging XXIX, 125190D (13 June 2023).

Rey Ducay, David W. Messinger, "Image fusion of hyperspectral and multispectral imagery using nearest-neighbor diffusion," J. Appl. Rem. Sens. 17(2) 024504 (10 April 2023).

Rey Ducay, David W. Messinger, "Hyperspectral-multispectral image fusion using nearest-neighbor diffusion-based sharpening algorithm," Proc. SPIE 12094, Algorithms, Technologies, and Applications for Multispectral and Hyperspectral Imaging XXVIII, 120940M (31 May 2022).

Rey Ducay, David W. Messinger, "Subpixel target implantation to assess pansharpening performance on hyperspectral datasets," Proc. SPIE 11727, Algorithms, Technologies, and Applications for Multispectral and Hyperspectral Imaging XXVII, 117270Z (12 April 2021).

Oduncu, Emrah & Yuksel, Seniha. (2021). "An in-depth analysis of hyperspectral target detection with shadow compensation via LiDAR," Signal Processing: Image Communication. 99. 116427. 

Rey Ducay, David W. Messinger, "Radiometric assessment of four pan-sharpening algorithms as applied to hyperspectral imagery," Proc. SPIE 11392, Algorithms, Technologies, and Applications for Multispectral and Hyperspectral Imagery XXVI, 1139203 (19 May 2020).

X. Zhang, Y. Liang and N. Cahill, "Using superpixels to improve the efficiency of Laplacian Eigenmap based methods for target detection in hyperspectral imagery," 2016 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Beijing, China, 2016, pp. 5876-5879.

Emmett J. Ientilucci, "New SHARE 2010 HSI-LiDAR dataset: re-calibration, detection assessment and delivery," Proc. SPIE 9976, Imaging Spectrometry XXI, 99760I (19 September 2016).

Emmett J. Ientilucci, "Target detection assessment of the SHARE 2010/2012 hyperspectral data collection campaign," Proc. SPIE 9472, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XXI, 94720H (21 May 2015).

S. Hagstrom and J. Broadwater, "Atmospheric and shadow compensation of hyperspectral imagery using voxelized LiDAR," 2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Milan, Italy, 2015, pp. 2959-2962.

Amanda K. Ziemann, David W. Messinger, "Hyperspectral target detection using graph theory models and manifold geometry via an adaptive implementation of locally linear embedding," Proc. SPIE 9088, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XX, 90880B (13 June 2014).

A. K. Ziemann and D. W. Messinger, "Manifold representations of single and multiple material classes in high resolution HSI," 2014 6th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS), Lausanne, Switzerland, 2014, pp. 1-4, doi.

Emmett J. Ientilucci, "Using a new GUI tool to leverage LiDAR data to aid in hyperspectral image material detection in the radiance domain on RIT SHARE LiDAR/HSI data," Proc. SPIE 8870, Imaging Spectrometry XVIII, 887009 (23 September 2013).

James A. Albano, David W. Messinger, Emmett Ientilucci, "Spectral target detection using a physical model and a manifold learning technique," Proc. SPIE 8743, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIX, 874318 (18 May 2013).

Emmett J. Ientilucci, "SHARE 2012: analysis of illumination differences on targets in hyperspectral imagery," Proc. SPIE 8743, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIX, 87430I (18 May 2013).

Kelly Canham, Daniel Goldberg, John Kerekes, Nina Raqueno, David Messinger, "SHARE 2012: large edge targets for hyperspectral imaging applications," Proc. SPIE 8743, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIX, 87430G (18 May 2013).

AnneMarie Giannandrea, Nina Raqueno, David W. Messinger, Jason Faulring, John P. Kerekes, Jan van Aardt, Kelly Canham, Shea Hagstrom, Erin Ontiveros, Aaron Gerace, Jason Kaufman, Karmon M. Vongsy, Heather Griffith, Brent D. Bartlett, Emmett Ientilucci, Joseph Meola, Lauwrence Scarff, Brian Daniel, "The SHARE 2012 data campaign," Proc. SPIE 8743, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIX, 87430F (18 May 2013).

Jared A. Herweg, John P. Kerekes, Oliver Weatherbee, David Messinger, Jan van Aardt, Emmett Ientilucci, Zoran Ninkov, Jason Faulring, Nina Raqueño, Joseph Meola, "SpecTIR hyperspectral airborne Rochester experiment data collection campaign," Proc. SPIE 8390, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XVIII, 839028 (24 May 2012).

David Gillis, Emmett Ientilucci, Jeffrey Bowles, "Results of GLMM-based target detection on the RIT data set," Proc. SPIE 7695, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XVI, 769523 (13 May 2010).

Graduate Theses
Ducay, Rey, "Exploitation of spatial content for enhancing pansharpening and image fusion performance," (2023). Thesis. Accessed from the Rochester Institute of Technology. 

Huang, Sihan, "Radiometrically-Accurate Hyperspectral Data Sharpening," (2022). Thesis. Accessed from the Rochester Institute of Technology. 

Carrock, Joseph A. II, "An Experimental Campaign to Collect Multiresolution Hyperspectral Data with Ground Truth," (2021). Thesis. Accessed from the Rochester Institute of Technology. 

Sun, Jiangqin, "Temporal Signature Modeling and Analysis," (2014). Thesis. Accessed from the Rochester Institute of Technology. 

Ding, Bo, "Hyperspectral Imaging System Model Implementation and Analysis," (2014). Thesis. Accessed from the Rochester Institute of Technology. 

Albano, James A., "Spectral Target Detection using Physics-Based Modeling and a Manifold Learning Technique," (2013). Thesis. Accessed from the Rochester Institute of Technology. 

2004 MegaCollect
Raqueno, R. & Raqueno, Nina & Weidemann, A. & Effler, Steven & Perkins, M. & Vodacek, Anthony & Schott, J. & Philpot, William & Kim, M. (2005). Megacollect 2004: Hyperspectral collection experiment over the waters of the Rochester Embayment. Proceedings of SPIE - The International Society for Optical Engineering. 5806

Raqueno, Nina & Smith, L. & Messinger, David & Salvaggio, C. & Raqueno, R. & Schott, J. (2005). Megacollect 2004: Hyperspectral collection experiment of terrestrial targets and backgrounds of the RIT Megascene and surrounding area (Rochester, New York). Proc SPIE. 

The SHARE 2012 Data Collection Campaign
AnneMarie Giannandrea, Nina Raqueno, David W. Messinger, Jason Faulring, John P. Kerekes, Jan van Aardt, Kelly Canham, Shea Hagstrom, Erin Ontiveros,
Aaron Gerace, Jason Kaufman, Karmon M. Vongsy, Heather Griffith, Brent D. Bartlett
SPIE Defense, Security, and Sensing. April 2013.

SHARE 2012: Subpixel Detection and Unmixing Experiments
John P. Kerekes, Kyle Ludgate, AnneMarie Giannandrea, Nina Raqueno
SPIE Defense, Security, and Sensing. April 2013.

For questions about this dataset, email us!

Sponsors

  • Rochester Institute of Technology
  • Chester F. Carlson Center for Imaging Science
  • Exelis
  • UTC Aerospace Systems
  • AFRL
  • MITRE
  • Naval Research Laboratory