SPECIM WEBINAR

INTRODUCTION TO HYPERSPECTRAL
IMAGING
15.9.2021
JUSSI SOUKKAMÄKI
SALES AND BUSINESS DEVELOPMENT MANAGER
 TODAY HEAD OFFICE
OULU, FINLAND
FOUNDED IN
1995
SPECTROGRAPHS
+
SPECTRAL CAMERAS

> 8 000
DELIVERED

70
PERSONNEL CIRCA

40% CAGR IN INDUSTRIAL SEGMENT

PART OF SPECIM LOCATIONS
KONICA MINOLTA BARCELONA, NEW JERSEY, SHANGHAI
SINCE 2020 + WORLDWIDE DISTRIBUTOR NETWORK
GLOBAL LEADER IN SPECTRAL IMAGING

R&D PRODUCTION SUPPORT

In-house design with superior 1000 m2 of scalable production A strong team to solve
know-how in optics, space, maintenance, calibration application
electronics, software and and repair services related challenges for pre-and
mechanics. Specim innovations with 160 m2 of cleanroom after-sales.
are protected by wide patent facilities.
portfolio.
 TIMELINE

Specim AFX
AISA Eagle HSI for unmanned
The first digital airborne airborne usage
hyperspectral camera.
SISU Series
Turnkey system SpecimONE
ImSpector for geology spectral imaging platform
Spectrographs for sorting industry
Specim IQ
AisaFENIX & FENIX 1k The first mobile full
Airborne VNIR/SWIR co- hyperspectral imaging Acquired by
AisaOWL
the world’s first thermal
registered turnkey solution system Konica Minolta
hyperspectral camera for
AISA Series UAVs
The first analogue airborne AisaIBIS Specim FX Specim FX50
hyperspectral camera. Airborne VNIR-SIF turnkey The first hyperspectral MWIR camera for
AisaDUAL solution cameras designed industrial use.
specifically for industrial
The first co-aligned
machine vision.
VNIR/SWIR airborne
turnkey solution.

1995 2000 2005 2015 2017 2019 2020 2021

What is spectral imaging ?

Different colors – Distinct spectra in the visible region

Every material has “spectral fingerprint”

Spectral imaging:
Combines spectral
measurement with
digital imaging
How to benefit from spectral imaging ?

Different chemical compounds – Distinct spectra in near infrared (NIR)

Spectral imaging:
The most powerful camera technique to
chemically identify and sort materials
 - for sorting applications

SpecimINSIGHT:
Off-line PC software
Specim FX camera for sorting model
training and validation

SpecimCUBE:
Real-time processing
software with WebUI

GigE
output
stream to
sorting
system
VISION
How SPECIM spectral camera works?

Line scan camera

(push-broom)

Records light spectrum

from each image pixel

Push-broom advantages
§ Full spectral data simultaneously
§ Full spatial 2D data over the scan area
§ Straightforward data processing
§ High light collection efficiency (F/1.7)
§ High spectral resolution (1-3 nm) achievable
§ Wide and pure spectral range achievable
§ VIS, VNIR, SWIR, MWIR and LWIR cameras
§ Line illumination only required
 Spectral Imaging
One of the quickest growing new imaging
technologies

AGRICULTURE & COLOR FOOD HEALTHCARE & GEOLOGY WASTE & RECYCLING ENVIRONMENT DEFENCE & SECURITY
VEGETATION MEASUREMENT PHARMA
§ Identify, sort and § Threat detection
§ Crop readiness, § Precise color § Food packaging § Production quality § Geological recycle waste § Environment § Target identification
ripeness, sugar identification under quality control control exploration § Textile sorting – monitoring § Man made
content different lighting § Identify § Package integrity § Mineral mapping of separate fiber § Water quality infrastructure
§ Yield forecasting conditions contamination of § Chemical drill cores materials inspection detection
via remote sensing § Display & light contents composition analysis § Detection of § Reliable polymer § Wildlife populations
§ Under the skin source quality / § Sizing and sorting § API distribution and minerals and ore detection § Leaks and
blemishes / bruises color accuracy § Chemical grading concentration in mine walls § Reliably identify contamination
§ Disease presence / § Color conformity in (moisture, fat, § Determining API and black plastic
phenotyping production protein) dosage § Material quality /
§ Vegetation and soil § Printing process § Identify foreign § Skin anomaly purity detection –
studies monitoring bodies detection driving up material
§ Art analysis and § Health & quality of re-use
restoration fish and fresh
produce
PRODUCT RANGE
VNIR NIR/SWIR MWIR LWIR
380 – 740 nm 900 – 2 500 nm 3 000 – 5 000 nm 8 000 – 12 500 nm

TYPICAL Colour, Fruit & Veg, Skin & Sorting, Recycling, Chemical Black Plastic, Surfaces, Mining, Detection,
APPLICATIONS Tissue, Plants, Film, Water Waste, Pharma, Quality Minerals, Detection Gas, Geothermal

IQ SERIES SPECIM IQ

FX SERIES FX10 FX17 FX50

AFX SERIES AFX10 AFX17

AisaFENIX / FENIX 1K
AISA SERIES AisaIBIS AisaOWL
AisaKESTREL

SisuCHEMA
SISU SERIES SisuROCK
SisuSCS

SPECTRAL PFD4K-65-V10E OWL

SWIR
CAMERAS sCMOS-50-V10E LWIR-HS
APPLICATIONS EXAMPLES
 Higher quality, safer food – Less waste, higher profit

Detection of foreign objects Meat grading

Motivation:
Estimated 20% of meat and
40% of fruit and vegetables
are wasted.

Sorting nut shells from the nut

Detection of stones in the fruit Detection of mold and bruises

 INTERNAL DEFECTS IN CUCUMBER

§ Transmittance between 740-

1000nm
§ 95% classification accuracy

Ariana, P. & Lu, R.: Hyperspectral

waveband selection for internal
defect detection of pickling cucumbers
And whole pickles. (2010). Univ.
FOOD INDUSTRY

§ Bruises in Kiwi
§ The VNIR camera
reveals defect on
fruits

Detection of bruises through PCA with selected wavelength

Jiangsu university and Henan university

FOOD INDUSTRY

§ Total soluble solid, water

content and firmess in
Mango
§ VNIR
§ High correlation
§ Other fruits?

McGill University Montreal

FOOD INDUSTRY

McGill University Montreal

BIOSAMPLE HEALTH IN MICROSCALE
SWIR detection of fungal infection in
maize kernels
Quick scanning of 1 to 2 seconds
over a kernel minimizes heat load

University of Stellenbosch, South Africa

 DISEASE DETECTION

§ Wheat monitoring
§ VNIR (also SWIR)
§ Infection
§ Accurate treatment
§ Specim comment: Field
work possible with IQ Yellow rust Septoria

Catholic University of Leuven

DISEASE DETECTION

§ Tobacco monitoring
§ VNIR
§ TMV pre-Infection
§ Tobacco Mosaic Virus

Photo: Wikipedia

DOI: https://doi.org/10.1038/s41598-017-04501-2
DISEASE DETECTION

§ Tobacco monitoring
§ VNIR
§ Black-Shank fungus

Pots 3 and 4 with disease

Photo: NC state university

DOI: 10.5897/AJAR11.1281
 FOREST HEALTH MAPPING

Sarawak Forest Department

Malaysia

Airborne HSI in VNIR provides

sensitive and high resolution
detection and mapping of
fungus disease in oil palm
trees

>50 km2/h
@0.5 m ground resolution
@50 m/s (100 knots)

www.specim.fi
AFX10 FIX MOUNT ON
ECODRONE EAS-8 PRO

www.specim.fi
 DETECTION OF NARCOTIC PLANTS

VNIR+
SWIR

Processed Hyperspectral Image

Marijuana • Known spectral characteristics
plants • Known background
• Geo-rectified

Courtesy of SpecTIR LLC, USA www.specim.fi

 Invasive/Target Species
• Invasive/Target Species
– Identify eucalyptus (in BLUE) in fire district
– Analysis: Utilize
spectral signature or
(fingerprint) of species

VNIR
www.specim.fi