2021

Wolfmayr, Monika; Pölönen, Ilkka; Lind, Leevi; Kašpárek, Tomáš; Penttilä, Antti; Kohout, Tomáš

In October 2024, European Space Agency’s Hera mission will be launched, targeting the binary asteroid Didymos. Hera will host the Juventas and Milani CubeSats, the first CubeSats to orbit close to a small celestial body performing scientific and technological operations. The primary scientific payload of the Milani CubeSat is the SWIR, NIR, and VIS imaging spectrometer ASPECT. The Milani mission objectives include mapping the global composition and the characterization of the binary asteroid surface. Onboard data processing and evaluation steps will be applied due to the limited data budget for the downlink to Earth and to perform the technological demonstration of a novel semi-autonomous hyperspectral imaging mission. Before downloading, the image data is evaluated in terms of sharpness and coverage and processed by compression. The challenges and their proposed solutions for the data processing part of the mission are investigated through studies. Since most noise contributors are unknown until Milani is activated, different noises are studied based on previous missions and derived from hyperspectral images taken in a laboratory environment mimicking the real-life situation. The hyperspectral camera technology in the laboratory is similar to the one used in the ASPECT imager payload. Both ASPECT and the imagers utilized in our measurements are based on employing a Fabry-Pérot interferometer as an adjustable transmission filter. The imagers are also designed and built by the same party, the Technical Research Centre of Finland (VTT). Best performing denoising techniques for each noise type are discussed on the one hand for the entire datacubes and on the other hand for the spatial domain only since the mission includes images taken only at specific wavebands. The advantage of applying denoising for the whole datacube comes from the internal dependencies between the wavebands, allowing efficient processing. A trade-off study for several noise reduction algorithms is presented. The goal is to implement efficient image processing algorithms with low computational complexity, securing the successful execution of the mission.