Spectral Imaging Laboratory

The research of the Spectral Imaging Laboratory combines hyperspectral imaging and image data analysis with state-of-the-art methods in modeling, simulation and machine learning. Spectral imagery includes spatial and spectral data: we focus on solving theoretical and practical problems involving multidimensional data.

Hyperspectral imaging can be used, for example, in remote sensing applications in forestry, agriculture and marine biology. There are also applications in the paper industry and quality control in the pharmaceutical industry. Hyperspectral imaging can also be used to monitor food quality and safety. In biomedical applications, e.g. wounds, skin cancer tumors and breast cancer have been researched.

Hyperspectral imaging is applied in research in many fields. The application requires expertise in spectral imaging and data analysis, as well as knowledge of the research topic. For our part, we are trying to lower the adoption threshold of hyperspectral imaging in applications in various fields. In connection with this, we work in several cooperation projects with e.g. VTT's hardware developers, hospitals, doctors, biologists and other professionals.

Our laboratory consists of a group of experienced researchers with a strong motivation to develop hyperspectral imaging applications. The group has worked on hyperspectral imaging in various applications ranging from crime scene investigation and art authentication to medical imaging and remote sensing. In addition, we have developed programs and programming libraries related to hyperspectral data processing and hardware.

Our cameras

Our laboratory has several spectral cameras for different wavelength ranges (400 - 2500 nm). In addition to our permanent equipment, we have worked with various prototype cameras, and developed our own line scanner-type hyperspectral camera. In addition to the cameras, we have a compatible microscope with which we can photograph, for example, cell samples.

Our cameras are easy to move: you can shoot with them both in the laboratory and in the field.

Cameras	wavelengths	FWHM's	Imaging type
Specim IQ	400-1000 nm		pushbroom
Specim FX10	400-1000 nm		pushbroom
Specim FX17	1000-1700 nm		pushbroom
Specim SWIR	1000-2500 nm		pushbroom
VTT FPI "Luigi"	400-950 nm		framing
VTT FPI "Pamela"	400-950 nm		framing
VTT FPI-proto 1	400-950 nm		framing
DIY imager	400-950 nm		pushbroom
MV cameras + filters	400-950 nm		framing

You can find open spectral data on our lab's website on Zenodo: https://zenodo.org/communities/hsilab

Collaboration

The Spectral Imaging Laboratory has good networks with research institutes in Finland and abroad. We are happy to do interdisciplinary research cooperation. For example, we can participate in your project or create a new project together. You can find our ongoing and finished projects at the bottom of the page.

The spectral imaging laboratory has good networks with research institutes in Finland and abroad. We are happy to do research cooperation with companies. For example, we can participate in your project, produce commissioned research or apply for funding for a new joint research project. You can find our ongoing and finished projects at the bottom of the page.

If you want to develop your business through new products or services and have thought about how hyperspectral imaging could be applied in your industry, we can help. We specialize in hyperspectral imaging and analysis research, as well as the latest methods of mathematical modeling, simulation and machine learning.

After the needs survey, we will provide you with an order research offer that meets your needs. In addition to this, we can teach you the basics of hyperspectral imaging. Book an appointment with us!

Professor Ilkka Pölönen

Open positions and general PhD student and Postdoc inquiries

HSILab currently receives a large number of applications and inquiries for PhD and postdoctoral positions, often 5–10 per day. Many of these messages appear to be broadly generated and have only a limited connection to our actual research activities. Unfortunately, we do not have the capacity to review or respond individually to applications or inquiries submitted outside open recruitment calls.

If you are interested in joining HSILab, please first check the open positions announced through the University of Jyväskylä recruitment pages. When we are recruiting for a funded PhD, postdoctoral, or other paid research position, we most likely publish an official call.

If no relevant position is currently advertised, this generally means that we are not recruiting for a paid position at that time.

Please use professional university channels for any recruitment-related communication. Do not contact HSILab members regarding job opportunities through WhatsApp, personal social media accounts, or other non-official, work-related channels. Thank you for understanding and for taking this into consideration.

Head of Laboratory

Ilkka Pölönen

Professor

Anna-Maria Raita

Academy Research Fellow

Laboratory personnel

Postdoctoral Researcher

Pritish Naik

Doctoral Researcher

Research projects

FireMan: Unmanned aerial systems based solutions for real-time management of wildfires

Developing disruptive AI-based technology for the fast-paced detection of wildfires and for creating situational awareness during wildfire events using autonomous unmanned aerial systems (UASs, drones).

HSI-UNITED

ML4DRONE - Learning techniques for autonomous drone based hyperspectral analysis of forest vegetation

Research Council of Finland

Smart-HSI: Smart hyperspectral imaging solutions for new era in Earth and planetary observations

The Smart-HSI project develops hyperspectral instruments and data processing methods for Earth observation and planetary science missions that employ CubeSats.

Publications

Active Hyperspectral Imaging with a Fabry–Pérot Interferometer-Based Tunable Light Source

Convolutional Neural Networks in Chronic Wound Segmentation and Tissue Classification Using Real‐World Images

dMMR prediction from colorectal cancer histopathology: leveraging non-tumor and low-magnification regions

Dual selective gleason pattern-aware multiple instance learning with uncertainty regularization for grade group prediction in histopathology images

Early detection of metastatic risk in primary cutaneous melanoma using weakly supervised learning

ISAC-Enhanced Secure Symbiotic Radio with Passive Localization and Joint Beamforming Design

Machine Learning Soft Sensor for Predicting the Product Quality of a Hydrogen Reduction Furnace

MQTT-Enabled Federated Self-Learning Minimal Learning Machine Classifier for Real-Time Hyperspectral Processing

Raita, Anna-Maria

Louichon, Hadrien

Pölönen, Ilkka

SAM-HovNet: SAM-Enhanced Multi-Class Nucleus Segmentation and Classification in Hepatocellular Carcinoma Slides

Water Quality Estimation Using Enmap Satellite Data for Loch Leven

A Review: Wavelengths of Interest in Hyperspectral Imaging-Based Medical Skin Diagnosis Applications

Raita-Hakola, Anna-Maria

A variational autoencoder and neural network approach to generating synthetic data in well-being research

Spectral Imaging Laboratory

Cookie consent

Our cameras

Collaboration

Academic collaboration

Funding for collaboration projects

Theses and doctoral studies

Business collaboration

Commissioned research

Open positions and general PhD student and Postdoc inquiries

Head of Laboratory

Laboratory personnel

Research projects

FireMan: Unmanned aerial systems based solutions for real-time management of wildfires

HSI-UNITED

ML4DRONE - Learning techniques for autonomous drone based hyperspectral analysis of forest vegetation

Smart-HSI: Smart hyperspectral imaging solutions for new era in Earth and planetary observations

Publications

Active Hyperspectral Imaging with a Fabry–Pérot Interferometer-Based Tunable Light Source

Convolutional Neural Networks in Chronic Wound Segmentation and Tissue Classification Using Real‐World Images

dMMR prediction from colorectal cancer histopathology: leveraging non-tumor and low-magnification regions

Dual selective gleason pattern-aware multiple instance learning with uncertainty regularization for grade group prediction in histopathology images

Early detection of metastatic risk in primary cutaneous melanoma using weakly supervised learning

ISAC-Enhanced Secure Symbiotic Radio with Passive Localization and Joint Beamforming Design

Machine Learning Soft Sensor for Predicting the Product Quality of a Hydrogen Reduction Furnace

MQTT-Enabled Federated Self-Learning Minimal Learning Machine Classifier for Real-Time Hyperspectral Processing

SAM-HovNet: SAM-Enhanced Multi-Class Nucleus Segmentation and Classification in Hepatocellular Carcinoma Slides

Water Quality Estimation Using Enmap Satellite Data for Loch Leven

A Review: Wavelengths of Interest in Hyperspectral Imaging-Based Medical Skin Diagnosis Applications

A variational autoencoder and neural network approach to generating synthetic data in well-being research