IceSpec: An ice core hyperspectral imaging system

System design The IceSpec system, a modified Specim SisuSCS (Single Core Scanner). It is equipped with a Specim FX10e hyperspectral camera. To enable scanning in a −24 °C clean room, several modifications were made: Airways to the Data Acquisition Computer (DAC), Power Control Unit (PCU), and AC/DC power supply were blocked to prevent freezing of electrical components. The belt tension on the ROLLCO linear unit (RHL-80) was adjusted. The original system tray was replaced with a new tray designed for scanning ice cores (100 × 4 × 10 cm) in both transmissive and /or reflective light. The system is mounted on a custom leveled table. To minimize ambient light interference, the IceSpec system is placed inside a 1.5 × 2.7 m dark enclosure (Gorilla Grow Tent) with an opening for loading ice core samples. The enclosure is positioned near the freezer room window, with cables routed through the wall to a Dell workstation equipped with a serial port. All system components are painted with Rust-Oleum Specialty Flat BBQ Black Oil-Based High Heat Enamel. A Logitech C922 webcam monitors the data collection area, while a Dell WB7022 webcam displays an ice core tray overview feed on an Apple Mac Mini using OBS Studio (v. 31.0.3). Midland two-way radios facilitate communication between the computer operator and personnel inside the freezer during maintenance or sample loading. Camera Enclosure The freezer’s −24 °C temperature presents challenges for maintaining a stable camera temperature above 5 °C for reliable data collection. To prevent freezing, the camera is housed in an insulated enclosure with a removable side panel secured by small magnets for easy focusing or maintenance access. Inside the enclosure, a 35 × 21 mm positive temperature coefficient (PTC) ceramic heating element, controlled by a HiLetgo W1209 12V DC digital temperature controller board, maintains a stable internal temperature. As a safety measure, an emergency power shutdown is triggered if the internal temperature exceeds 18 °C, using an Emerson 3L11-60 1/2-inch snap disc thermostat (open on rise). The camera enclosure has a free-sitting N-BK7 optical window (75 × 75 mm, 4 mm thick, uncoated) from Edmund Optics. This window has transmission efficiency exceeding 91% across the 300–2000 nm spectral range, encompassing ultraviolet, visible, and near-infrared (NIR) wavelengths. Lighting, Optical Components, Spectral Standards The system works with two lighting configurations: halogen and LED. Initially, it employed halogen-based light sources for their broad, continuous spectrum. Four Schott DCR III light sources, each fitted with a custom OSL2BIR 150 W, 3200 K bulb from Thorlabs, deliver focused light via optical fibers. To enhance NIR performance, the built-in IR filter was removed, extending the emitted light spectrum from 750 to 1000 nm. This modification does not cause overheating, as the cold freezer environment effectively dissipates heat, ensuring safe operation. The system was also tested with LED lighting designed for hyperspectral applications. Four 200 mm wide bar lights from EFFILUX (Part Number: CCS, EFFI-FLEX-CPT-HSI-X2-200-910-970-TR-P3-LS-CYL), sealed to IP67 standards, cover a spectral range of 400–1000 nm. Two units are positioned above the ice sample, and two illuminate from below. The interchangeable design of halogen and LED lights, with compact dimensions, allows versatile lighting configurations. To control light direction and minimize instability during the first hour of operation, remotely operated light-blocking shades or simple black cardboard are used when only top or bottom illumination is required, avoiding startup time instability. The halogen light assembly includes a high-reflectance mirror: a 75 × 100 mm protected silver mirror from Edmund Optics, with average reflectance >98% from 450 to 2000 nm and surface flatness (peak-to-valley) of 4–6λ, ensuring reliable light redirection and focusing. Data collection and storage Before data collection, the focusing grid and reflection calibration standards are leveled using acrylic shims of varying thickness to align with a plane 5 mm below the ice core surface. The original white reference panel and focusing grid from Specim were modified to include a calibration block used with every scanned sample. This block comprises an acrylic bar and three reflectance panels (100%, 50% SRT-50–050, and 20% SRT-20–050 Spectralon Diffuse Reflectance Target from Labsphere) with certified calibration properties, ensuring accurate hyperspectral data collection and white reference values. Secondary spectral standards are prepared using a mixture of barium sulfate (97%) and Neodymium(III) oxide (REacton™, 99.99% REO from Thermo Scientific™), mixed with urethane and painted onto a wooden substrate. These standards, alongside Teflon PTFE, Paint White 2.0, and Paint Black 2.0 products, help verify spectral data collection and ensure consistent spectral accuracy. For top-only lighting, data are calibrated with reflectance panels. When only bottom lights are used, white reference calibration is performed using a 100 × 100 mm sandblasted glass panel (Part 43-725 from Edmund Optics) illuminated from below. For mixed lighting experiments, light intensities are recorded for top and bottom sources prior to data acquisition. The IceSpec system, utilizing a Specim FX10e camera, captures data in ENVI format and stores it initially on a local hard drive. Backups are maintained on a dedicated network-attached storage (NAS) system and external drives for transfer to the University of Maine’s processing server. SpecimINSIGHT (v.1.4) software enables image previewing, renaming, and organization. Data analysis is performed using custom Python libraries, PyHAT, Spectral Python (SPy), Napari with the napari-sediment plugin, pandas, NumPy, scikit-learn, and PyTorch. All data are archived in ENVI format to ensure compatibility and fidelity for paleoenvironmental research. ACKNOWLEDGEMENTS Developed as part of the US National Science Foundation Center for Oldest Ice Exploration project (COLDEX, 2025), the IceSpec system is designed for the US ice core community and currently operates at the NSF Ice Core Facility in Denver, Colorado. Integrated into a new ice core laboratory, this system supports long-term archiving, timescale development, and mapping of physical properties and impurities in ice cores. It is poised to enhance future ice core projects, advancing our understanding of paleoclimate dynamics.

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2025-08-14
DOI
https://doi.org/10.5281/zenodo.16875811
Primary Topic
Arctic and Antarctic ice dynamics
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
article

IceSpec: An ice core hyperspectral imaging system

Zenodo (CERN European Organization for Nuclear Research)
Arctic and Antarctic ice dynamics
article

IceSpec: An ice core hyperspectral imaging system

article en

Abstract

System design The IceSpec system, a modified Specim SisuSCS (Single Core Scanner). It is equipped with a Specim FX10e hyperspectral camera. To enable scanning in a −24 °C clean room, several modifications were made: Airways to the Data Acquisition Computer (DAC), Power Control Unit (PCU), and AC/DC power supply were blocked to prevent freezing of electrical components. The belt tension on the ROLLCO linear unit (RHL-80) was adjusted. The original system tray was replaced with a new tray designed for scanning ice cores (100 × 4 × 10 cm) in both transmissive and /or reflective light. The system is mounted on a custom leveled table. To minimize ambient light interference, the IceSpec system is placed inside a 1.5 × 2.7 m dark enclosure (Gorilla Grow Tent) with an opening for loading ice core samples. The enclosure is positioned near the freezer room window, with cables routed through the wall to a Dell workstation equipped with a serial port. All system components are painted with Rust-Oleum Specialty Flat BBQ Black Oil-Based High Heat Enamel. A Logitech C922 webcam monitors the data collection area, while a Dell WB7022 webcam displays an ice core tray overview feed on an Apple Mac Mini using OBS Studio (v. 31.0.3). Midland two-way radios facilitate communication between the computer operator and personnel inside the freezer during maintenance or sample loading. Camera Enclosure The freezer’s −24 °C temperature presents challenges for maintaining a stable camera temperature above 5 °C for reliable data collection. To prevent freezing, the camera is housed in an insulated enclosure with a removable side panel secured by small magnets for easy focusing or maintenance access. Inside the enclosure, a 35 × 21 mm positive temperature coefficient (PTC) ceramic heating element, controlled by a HiLetgo W1209 12V DC digital temperature controller board, maintains a stable internal temperature. As a safety measure, an emergency power shutdown is triggered if the internal temperature exceeds 18 °C, using an Emerson 3L11-60 1/2-inch snap disc thermostat (open on rise). The camera enclosure has a free-sitting N-BK7 optical window (75 × 75 mm, 4 mm thick, uncoated) from Edmund Optics. This window has transmission efficiency exceeding 91% across the 300–2000 nm spectral range, encompassing ultraviolet, visible, and near-infrared (NIR) wavelengths. Lighting, Optical Components, Spectral Standards The system works with two lighting configurations: halogen and LED. Initially, it employed halogen-based light sources for their broad, continuous spectrum. Four Schott DCR III light sources, each fitted with a custom OSL2BIR 150 W, 3200 K bulb from Thorlabs, deliver focused light via optical fibers. To enhance NIR performance, the built-in IR filter was removed, extending the emitted light spectrum from 750 to 1000 nm. This modification does not cause overheating, as the cold freezer environment effectively dissipates heat, ensuring safe operation. The system was also tested with LED lighting designed for hyperspectral applications. Four 200 mm wide bar lights from EFFILUX (Part Number: CCS, EFFI-FLEX-CPT-HSI-X2-200-910-970-TR-P3-LS-CYL), sealed to IP67 standards, cover a spectral range of 400–1000 nm. Two units are positioned above the ice sample, and two illuminate from below. The interchangeable design of halogen and LED lights, with compact dimensions, allows versatile lighting configurations. To control light direction and minimize instability during the first hour of operation, remotely operated light-blocking shades or simple black cardboard are used when only top or bottom illumination is required, avoiding startup time instability. The halogen light assembly includes a high-reflectance mirror: a 75 × 100 mm protected silver mirror from Edmund Optics, with average reflectance >98% from 450 to 2000 nm and surface flatness (peak-to-valley) of 4–6λ, ensuring reliable light redirection and focusing. Data collection and storage Before data collection, the focusing grid and reflection calibration standards are leveled using acrylic shims of varying thickness to align with a plane 5 mm below the ice core surface. The original white reference panel and focusing grid from Specim were modified to include a calibration block used with every scanned sample. This block comprises an acrylic bar and three reflectance panels (100%, 50% SRT-50–050, and 20% SRT-20–050 Spectralon Diffuse Reflectance Target from Labsphere) with certified calibration properties, ensuring accurate hyperspectral data collection and white reference values. Secondary spectral standards are prepared using a mixture of barium sulfate (97%) and Neodymium(III) oxide (REacton™, 99.99% REO from Thermo Scientific™), mixed with urethane and painted onto a wooden substrate. These standards, alongside Teflon PTFE, Paint White 2.0, and Paint Black 2.0 products, help verify spectral data collection and ensure consistent spectral accuracy. For top-only lighting, data are calibrated with reflectance panels. When only bottom lights are used, white reference calibration is performed using a 100 × 100 mm sandblasted glass panel (Part 43-725 from Edmund Optics) illuminated from below. For mixed lighting experiments, light intensities are recorded for top and bottom sources prior to data acquisition. The IceSpec system, utilizing a Specim FX10e camera, captures data in ENVI format and stores it initially on a local hard drive. Backups are maintained on a dedicated network-attached storage (NAS) system and external drives for transfer to the University of Maine’s processing server. SpecimINSIGHT (v.1.4) software enables image previewing, renaming, and organization. Data analysis is performed using custom Python libraries, PyHAT, Spectral Python (SPy), Napari with the napari-sediment plugin, pandas, NumPy, scikit-learn, and PyTorch. All data are archived in ENVI format to ensure compatibility and fidelity for paleoenvironmental research. ACKNOWLEDGEMENTS Developed as part of the US National Science Foundation Center for Oldest Ice Exploration project (COLDEX, 2025), the IceSpec system is designed for the US ice core community and currently operates at the NSF Ice Core Facility in Denver, Colorado. Integrated into a new ice core laboratory, this system supports long-term archiving, timescale development, and mapping of physical properties and impurities in ice cores. It is poised to enhance future ice core projects, advancing our understanding of paleoclimate dynamics.

Zenodo (CERN European Organization for Nuclear Research)
United States Geological Survey (US), U.S. National Science Foundation (US), Oregon State University (US), United States Department of the Interior (US), Colorado School of Mines (US), Cold Regions Research and Engineering Laboratory (US), Specim (Finland) (FI), University of Maine (US)
Industry, innovation and infrastructure
Openalex Percentile: Top 54%
Arctic and Antarctic ice dynamics
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.