MIRA develops multimodal imaging systems and image reconstruction algorithms spanning the complete imaging pipeline, from sensor acquisition and calibration to registration, restoration, fusion, and deployment. Particular emphasis is placed on spectral and hyperspectral imaging, multimodal sensor configurations, reconstruction under real acquisition conditions, and practical hybrid multi-camera systems.
Developing registration and fusion methods for RGB and hyperspectral cameras operating at different spatial resolutions. The research focuses on real cross-sensor misalignment, uncertainty in geometric alignment, and spectral-preserving super-resolution under imperfect registration.
Investigating restoration methods for hyperspectral imagery using spectral-domain models and transferable priors learned from conventional imagery. Current research includes denoising, model-based reconstruction, and learned priors for spectral inverse problems.
Designing reproducible multi-camera systems for synchronized multimodal acquisition, with emphasis on RGB, multispectral, and hyperspectral sensors. Research includes camera integration, acquisition software, and deployment on robotic and mobile platforms.
Studying the physical and radiometric properties of spectral imaging systems, including radiometric calibration, spectral response, noise and exposure characteristics, spatial degradation, and consistency across sensors.
Building multimodal spectral datasets and acquisition protocols for scientific and industrial applications. MIRA combines controlled imaging, sensor characterization, physical measurements, and reproducible processing for quantitative spectral analysis.
The MIRA platform contributes to ongoing international research programs and bilateral collaborations, combining expertise in spectral imaging, fusion, and machine vision.
Project: Spectral Analysis of Cocoa — A Physical and Chemical Laboratory Approach
Dataset: CoCoaSpec (Hugging Face)
Focus: Multisensor fusion and imaging for agricultural and robotic perception.
Multi-camera Visualizer: MultiCamGUI (GitHub)
For some description on the following topics, click a card or visit the Knowledge page.
