Machine learning for quality assessment of ground-based optical images of satellites

Kyono, Trent; Lucas, Jacob; Werth, M.; Calef, Brandoch; McQuaid, Ian; Fletcher, Justin

doi:10.1117/1.oe.59.5.051403

Cited by 9 publications

(2 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is anticipated that machine learning will help minimise the number of measurements required for indirect sensing methods without additional instrumentation, thus allowing simplification of the overall system while simultaneously widening its scope of applications. Machine learning for wavefront sensing has been successfully demonstrated in several AO fields using point objects and for retinal imaging [212][213][214][215] or for aberration prediction in astronomy 216 . However, significant work remains to be done before machine learning be applied in more complex specimens and implemented in distributable packages for universal usage.…”

Section: [H1] Outlookmentioning

confidence: 99%

Adaptive optics for high-resolution imaging

2021

Nat Rev Methods Primers

View full text Add to dashboard Cite

Adaptive optics (AO), a technique that corrects for optical aberrations, was originally proposed to correct for the blurring effect of atmospheric turbulence on images in ground-based telescopes; indeed, the technique was instrumental in the work that resulted in the discovery of a supermassive compact object at the centre of our galaxy, which was awarded the 2020 Nobel Prize in physics. When AO is used to correct for the eye's imperfect optics, retinal changes at the cellular level can be detected, allowing us to study the operation of the visual system and to assess ocular health in the microscopic domain. By correcting for sample-induced blur in microscopy, it has pushed the boundaries of imaging in thick tissue specimens, such as when observing neuronal processes in the brain. The focus of this primer is the application of AO for high resolution imaging in astronomy, vision science, and microscopy. It begins with an overview of the general principles of AO and its main components, which include methods to measure the aberrations and devices for their correction. These components are linked in operation via a control system. Results and applications from each field are presented, along with reproducibility considerations and limitations. Finally, future directions are discussed. [H1] IntroductionHigh resolution optical imaging relies upon the high fidelity focussing of light. Light can be described in terms of optical fields[G], and thus its properties are parametrized for a given wavelength at each point in space and time in terms of amplitude, phase and polarization. However, these fields can be perturbed (in amplitude, phase and polarization) as they propagate through optical systems and other media, and the performance of the imaging systems can be highly sensitive to those perturbations. For instance, astronomical image quality is limited by atmospheric turbulence; microscopes produce blurred images when samples have a non-uniform refractive index distribution; and ophthalmoscopes that image the back of the eye are detrimentally affected by the eye's imperfect optics. Adaptive optics (AO) is an ensemble of electro-optical and computational methods that aim at recovering the optimal performance of an optical system 1-6 . This has brought benefits to a range of applications. For example, by integrating AO in their telescopes, astronomers have been able to expand the observation of celestial bodies 7 . Implemented into microscopes, AO has enabled neuroscientists to monitor the activity of neurons embedded deep inside the living mammalian brain 8,9 . And integrated into ophthalmoscopes, vision scientists and ophthalmologists are able to visualize, quantify, and track in situ the many different types of cells that compose the retina, offering significant clinical potential 10-14 .There are many imaging applications in which AO has a significant impact, and the above list is far from exhaustive.Compensation [G] through modulation of the optical field is the fundamental working principle through which AO allevia...

show abstract

Section: [H1] Outlookmentioning

confidence: 99%

Adaptive optics for high-resolution imaging

2021

Nat Rev Methods Primers

View full text Add to dashboard Cite

show abstract

“…9 Kyono and Fletcher also studied resolved imagery, and partially resolved imagery to estimate the quality of optical images of satellites. 10 Peng proposed a method of using Gaussian processes to augment orbit determination methods by learning the offset from lower fidelity predictions to true observations, and is applicable to the orbit determination steps to be considered in the future. 11,12 Another study looked at the possibility of using imagery from star-trackers for SSA purposes and used a Recurrent-Convolutional Neural Network (R-CNN) to estimate RSO position and velocities.…”

Section: Introductionmentioning

confidence: 99%

Satellite Detection in Unresolved Space Imagery for Space Domain Awareness Using Neural Networks

Jordan¹,

Posada²,

Zuehlke³

et al. 2022

Preprint

View full text Add to dashboard Cite

This work utilizes a MobileNetV2 Convolutional Neural Network (CNN) for fast, mobile detection of satellites, and rejection of stars, in cluttered unresolved space imagery. First, a custom database is created using imagery from a synthetic satellite image program and labeled with bounding boxes over satellites for "satellitepositive" images. The CNN is then trained on this database and the inference is validated by checking the accuracy of the model on an external dataset constructed of real telescope imagery. In doing so, the trained CNN provides a method of rapid satellite identification for subsequent utilization in ground-based orbit estimation.

show abstract