Lobster eye telescopes are a type of innovative telescope design, which could observe celestial objects over a very wide field of view in x-ray band. Thanks to this property, lobster eye telescopes are widely used to detect x-ray transients in time-domain astronomy. However, images obtained by lobster eye telescopes are modified by their unique point spread functions, which would spread photons from point sources to large images with crucify structure. Therefore, it is hard to design an automatic source detection algorithm with high efficiency and fast speed. Manual interventions are always required to modify parameters of contemporary methods to fit data properties of each observed images. In this paper, we will review the classical method and several new methods proposed by our group to detect sources from images obtained by lobster eye telescopes. We have compared the performance of different methods and results show that we would require to integrate different methods to develop a pipeline to process images obtained by lobster eye telescopes.
Atmospheric turbulence phase screens are important for performance test of adaptive optics systems and image restoration algorithms. In recent years, more and more model based adaptive optics control algorithms, wave-front reconstruction algorithms and image restoration algorithms have been proposed. These algorithms have very strong prior information about properties of turbulence. Ordinary atmospheric turbulence phase screen generation methods would become inappropriate to test these methods, because these phase screen generation methods are also developed according to prior information of turbulence. In this paper, we will report our recent progress in building a digital twin of atmospheric turbulence phase screen, which automatically obtain models of atmospheric turbulence and generates phase screens directly. Our method has no prior assumptions about the statistical properties of turbulence phase screens, which make it adequate to test the performance of model based algorithms.
Complementary Metal Oxide Semiconductor (CMOS) detectors have attracted more and more interests as appropriate imaging instruments, because their performance is close to that of CCD with low power consumption and low noise level. Yangwang-1 is a commercial satellite in China that has two telescopes on board. An optical telescope and a ultra-violet telescope are installed in the satellite to observe near earth objects and bright stars in optical and ultra-violet band for space mining. Yangwang-1 uses CMOS as its camera for optical and UV-band. However, CMOS detectors have different response for different pixels and there is no independent cooling system for the CMOS in the Yangwang-1, which would make dark current control and modelling hard. In this paper, we propose an unsupervised dark current modelling and bad pixel recognition method for CMOS detectors in Yangwang-1 UV camera. Our method obtains several dark current calibration frames on the ground when the CMOS is in different temperatures. Then vectors of dark current in each pixel will be clustered with Gaussian Mixture Model to identify bad pixels and pixels that have the same trends in dark current and temperature relations. We would fit different dark-current and temperature relations for pixels that belong to the same cluster as templates for dark current estimation. Then we could quickly estimate dark current for real observation data with pixels that have no sources nearby. With estimated dark current and bad pixel mask, we find that our source detection pipeline could achieve higher accuracy for UV band observation images.
The blurred range of astronomical image data we observe is usually uncertain, Due to the complex space environment, random noise, unpredictable atmospheric turbulence and other external factors. We usually use ground-based large aperture optical telescopes to observe astronomical images, which are mainly affected by atmospheric turbulence. Therefore, the restoration of astronomical images under the influence of arbitrary atmospheric turbulence is of great significance for the theoretical development and technological progress of astronomy. In this paper, a novel astronomical image restoration algorithm is proposed, which connects the deep learning based image restoration algorithm with the data generation method. The algorithm could effectively restore images within predefined blur or noise levels. We use long exposure galaxy images and short exposure Solar images to test the algorithm. We find that a well trained algorithm can restore these images.
As more and more images are obtained by astronomical observations, a fast image quality evaluation algorithm is required for data processing pipelines. The image quality evaluation algorithm should be able to recognize blur or noise levels according to scientists’ requirements and further mask parts of images with low qualities. In this paper, we introduce a deep learning based image quality evaluation and fast masking algorithm. Our algorithm uses an auto-encoder neural network to obtain blur or noise levels and we further use blur or noise levels to generate mask maps for input images. Tested with simulated and real data, our algorithm could provide reliable results with small amount of images as the training set. Our algorithm could be used as a reliable image mask algorithm for different image processing pipelines.
Photo plates have been used to capture and store astronomical images for quite a long time. In recent years, several projects are carried out to digitize photo plates and these digitized photo plates are shared through the Internet. We could extract invaluable astronomical data from digitized photo plates to analyse astronomical targets with very long temporal variations (up to decades). Extracting positions of celestial objects from photo plates and calculating their positions in celestial coordinates would be the first step. However, since astronomers would use multiple exposures to obtain images in photo plates and there are some scratches and mildews during storage of these photo plates, it becomes hard for us to directly obtain necessary information from digitized photo plates. In this paper, we will discuss the data processing pipeline developed by us to process photo plates digital archives.
Meteorites are remains of space targets, which could provide direct evidence to analyse the history of the solar
system. Meteorites normally come from comets, asteroids or meteoroids. When these space targets travel through
the atmosphere, they would radiate energy and become fireballs. With appropriate instruments, we could build a
system to obtain meteorites. The system includes devices to observe these fireballs and unmanned aerial vehicles
to discovery these meteorites on the ground. In a meteorites detection system, an adequate observation device
and detection algorithm are important. In this paper, we would show the device developed by our group with
commercial camera and neural networks to detect fireballs efficiently. The device could achieve more than 97%
accuracy in detection of fireballs in real time. Besides, the device is cheap and small, which could be placed in
different places to form an array to locate meteorites.
In recent years, time domain astronomy has become an active research area. Thanks to its low cost and moderate observation ability, wide field small aperture telescopes are commonly used to observe celestial objects for time domain astronomy. We would use several wide field small aperture telescopes to form an array to observe celestial objects continuously. Because there are many celestial objects for telescope arrays to observe, such as obtaining positions or magnitudes of celestial objects or discovering new transients, it would be necessary to investigate an optimal control strategy to maximize their scientific outputs. To achieve this target, we need to make trade-offs between observations of different targets and define appropriate tasks for each telescope. In this paper, we propose a framework, which includes a simulator and a reinforcement learning based algorithm, to obtain optimal control strategy for wide field small aperture telescope arrays, according to predefined scientific requirements. Our method could achieve better performance than ordinary sky survey strategies and has good generalization ability after training.
KEYWORDS: Telescopes, Optical instrument design, Point spread functions, Cameras, Signal to noise ratio, Prototyping, Genetic algorithms, Optimization (mathematics), Monte Carlo methods, CCD cameras
Time domain astronomy requires continuous observations to capture series of images of celestial object. To satisfy observation requirements, scientists need to place several telescopes in different sites or in the same site to form a telescope array. In recent years, several telescope array projects have been proposed. Design of telescope arrays is quite different from that of a single telescope. We need to make trade off between the image quality, the aperture size, the field of view and the number of telescopes to satisfy the scientific requirement with minimal overall cost. In this paper, we will introduce our method to optimize the design of a telescope array which would consider detail design of telescopes as well as the overall cost for building and maintaining of these telescopes. This paper provides a useful tool for future telescope array projects.
Digital twin of optical telescopes could provide high reliable models to predict the performance of the overall system during the design or the manufacture stage and could further be used to analyse faults after we build these telescopes. Comparing with conventional simulation methods, the digital twin of optical telescopes would import real test or telemetry data with machine learning algorithms and integrate numerical models with these data. As there are a lot of data to be collected, such as test data for optical elements, telemetry data from the telescope and the outer environment, it would be necessary to develop an appropriate database for digital twin of the optical telescope. In this paper, we would discuss our design of the database and use a simple example to show applications of the database.
Wide field small aperture telescopes (WFSATs) are commonly used for fast sky survey. Telescope arrays composed by several WFSATs are capable to scan sky several times per night. Huge amount of data would be obtained by them and these data need to be processed immediately. In this paper, we propose ARGUS (Astronomical taRGets detection framework for Unified telescopes) for real-time transit detection. The ARGUS uses a deep learning based astronomical detection algorithm implemented in embedded devices in each WFSATs to detect astronomical targets. The position and probability of a detection being an astronomical targets will be sent to a trained ensemble learning algorithm to output information of celestial sources. After matching these sources with star catalog, ARGUS will directly output type and positions of transient candidates. We use simulated data to test the performance of ARGUS and find that ARGUS can increase the performance of WFSATs in transient detection tasks robustly.
The point spread function (PSF) is the impulse response of an optical system. PSFs of an adaptive optics system have very strong variations both in temporal and spatial domain and a stable PSF reconstruction algorithm is required to provide prior information for scientific data processing. In this paper, we report our recent progress in developing a framework for PSF modelling with non-parametric model. The non-parametric PSF model uses compressive wavefront sensing method to build PSFs from wavefront measurements. Then a PSF-NET is used to learn map between PSFs estimated from wavefront sensing and PSFs in different field of views in a ground layer adaptive optics system. We use simulated data to test performance of the non--parametric PSF model and the results show its effectiveness.
From ground-based extremely large telescopes to small telescope arrays used for time domain astronomy, point spread function plays an important role both for scientific data post-processing and instrument performance estimation. In this paper, we propose a new method which can restore astronomical images and obtain the point spread function of the whole optical system at the same time. Our method uses simulated high resolution astronomical images and real observed blurred images to train a deep neural network (Cycle-GAN). The Cycle- GAN contains a pair of generative adversarial neural networks and each generative adversarial neural network contains a generator and a discriminator. After training, one generator (PSF-Gen) can learn the point spread function and the other generator (Dec-Gen) can learn the deconvolution kernel. We test our method with real observation data from solar telescope and small aperture telescopes. We find that the Dec-Gen can give promising restoration results for solar images and can reduce the PSF spatial variation for images obtained by smaller telescopes. Besides, we also find that the PSF-Gen can provide a non-parametric PSF model for short exposure images, which would then be used as prior model for PSF reconstruction algorithms in adaptive optics systems.
MEMS deformable mirrors (DM) have many merits of low drive voltage, high response speed, small power consumption, low cost and small size. Its surface shape and displacement versus applied voltage are significant factors of MEMS DM. Phase-shifting interferometer (PSI) has many advantages such as non-contact, quickness and high precision. A phase-only liquid crystal spatial light modulator (LC-SLM), as a linear phase-shifter in PSI, is linear calibrated for its phase-shift characteristics. The PSI is set up to measure the static characteristic of MEMS DM. Five-step phase-shifting method is used to calculate the phase distribution from interference fringes, and Global phase unwrapping algorithm to solve the holes, noise and breakpoint of interfere images. Compared to the measurement results using Zygo instrument, these two experimental results are very close. The experiment results show, this measuring system is very reliable, convenient and cheap. Moreover, this test system need not stitch some fringe images to get the whole surface shape of the mirror like the Zygo instrument.
A modal control optimization method for adaptive optics on the tempo-spatial domain is presented. The spatial modes of
the adaptive optics system can be obtained by the singular value decomposition of the response function matrix of the
adaptive optics system. The number of correction modes is determined dynamically by the root mean square estimation
of the residual aberration after the correction with different number of modes. A Smith compensator is designed to
reduce the time delay effect on the closed-loop system. The modal optimization method is experimentally verified by
compensating phase distortion produced by artificial atmospheric turbulence in laboratory. Experimental results show
that the correction capability of the adaptive optics system can be greatly improved in comparison to that of the generic
modal gain integrator approach with the fixed number of correction modes. The modal control optimization method is an
attractive and practical alternative to adaptive optics control.
The Antarctic is an ideal place for optical and infrared astronomy observations. Chinese scientists are planning
to build a 2.5m telescope in Dome A. The telescope will be built in a tower about 15 meters high to avoid the
ground layer atmospheric turbulence. The Ground layer Adaptive Optics system (GLAO) will also be suggested
to be installed to further reduce the seeing. The GLAO system with one laser guide star, one deformable mirror
and one wide field Shack-Hartmann wavefront sensor is designed and simulated. The Strehl ratio has increased
2 to 3 times in visible and infrared band in 20 arc min field of view.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.