As New Horizons flies-by Pluto today, at a speed of 16+ km/s there will be a short window of opportunity for the spacecraft to perform the most accurate images of this planet before it continues its journey to the Kuyper belt (the speed of the spacecraft makes it impossible to orbit Pluto).
Images like the one above are taken by LORRI and are black and white (panchromatic) but the instrument that will provide much of the science data for this Pluto encounter will be Ralph
More detailed on this camera can be found here....Ralph consists of three panchromatic (black-and-white) and four color imagers inside its Multispectral Visible Imaging Camera (MVIC), as well as an infrared compositional mapping spectrometer called the Linear Etalon Imaging Spectral Array (LEISA). LEISA is an advanced, miniaturized short-wavelength infrared (1.25-2.50 micron) spectrometer provided by scientists from NASA’s Goddard Space Flight Center. MVIC operates over the bandpass from 0.4 to 0.95 microns. Ralph’s suite of eight detectors – seven charge-coupled devices (CCDs) like those found in a digital camera, and a single infrared array detector – are fed by a single, sensitive magnifying telescope with a resolution more than 10 times better than the human eye can see. The entire package operates on less than half the wattage of an appliance light bulb.
All this to say, that any improvement on obtaining hyperspectral data, such as the one provided by Ralph during the fly-by, coupled with compression from cheap (powerwise) hardware could eventually be very useful to future space missions (please note the 6.3 watts power use of the camera). It so happens that in compressive sensing, we have the beginning of an answer as exemplified by the hardware in the CASSI imager (many of the blog entries relating to Hyperspectral imaging and ompressive sensing can be found under this tag.)
Today, Dror and colleagues show us how to reconstruct hyperspectral images when they are taken by these compressive imagers using AMP solvers. Here is the tutorial video made by Jin Tan and Yanting Ma followed by their preprint:
Today, Dror and colleagues show us how to reconstruct hyperspectral images when they are taken by these compressive imagers using AMP solvers. Here is the tutorial video made by Jin Tan and Yanting Ma followed by their preprint:
Compressive Hyperspectral Imaging via Approximate Message Passing by Jin Tan, Yanting Ma, Hoover Rueda, Dror Baron, Gonzalo Arce
We consider a compressive hyperspectral imaging reconstruction problem, where three-dimensional spatio-spectral information about a scene is sensed by a coded aperture snapshot spectral imager (CASSI). The CASSI imaging process can be modeled as suppressing three-dimensional coded and shifted voxels and projecting these onto a two-dimensional plane, such that the number of acquired measurements is greatly reduced. On the other hand, because the measurements are highly compressive, the reconstruction process becomes challenging. We previously proposed a compressive imaging reconstruction algorithm that is applied to two-dimensional images based on the approximate message passing (AMP) framework. AMP is an iterative algorithm that can be used in signal and image reconstruction by performing denoising at each iteration. We employed an adaptive Wiener filter as the image denoiser, and called our algorithm "AMP-Wiener." In this paper, we extend AMP-Wiener to three-dimensional hyperspectral image reconstruction. Applying the AMP framework to the CASSI system is challenging, because the matrix that models the CASSI system is highly sparse, and such a matrix is not suitable to AMP and makes it difficult for AMP to converge. Therefore, we modify the adaptive Wiener filter to fit the three-dimensional image denoising problem, and employ a technique called damping to solve for the divergence issue of AMP. Our simulation results show that AMP-Wiener in three-dimensional hyperspectral imaging problems outperforms existing widely-used algorithms such as gradient projection for sparse reconstruction (GPSR) and two-step iterative shrinkage/thresholding (TwIST) given the same amount of runtime. Moreover, in contrast to GPSR and TwIST, AMP-Wiener need not tune any parameters, which simplifies the reconstruction process.
Credit: NASA/Johns Hopkins University Applied Physics Laboratory/Southwest Research Institute
Liked this entry ? subscribe to Nuit Blanche's feed, there's more where that came from. You can also subscribe to Nuit Blanche by Email, explore the Big Picture in Compressive Sensing or the Matrix Factorization Jungle and join the conversations on compressive sensing, advanced matrix factorization and calibration issues on Linkedin.
Igor, thanks for the generous introduction! Indeed, my coauthors and I all hope that advances in hardware algorithms will bring advanced imaging techniques closer to practice.
ReplyDeleteDror
Igor, thanks for your generous introduction. My coauthors and I hope that indeed advances in hardware systems and algorithms will bring advanced imaging systems closer to practice.
ReplyDeleteDror