Leonid Shakun, Astronomical Observatory of Odesa I.I.Mechnikov National University; Krzysztof Kaminski, Adam Mickiewicz University; Oleksandr Briukhovetskyi, Western Radio Technical Surveillance Center, National Space Facilities Control and Test Center of State Space Agency of Ukraine; Justyna Golebiewska, Astronomical Observatory Institute, Faculty of Physics, A.Mickiewicz University,; Edwin Wnuk, Astronomical Observatory Institute, Faculty of Physics, A.Mickiewicz University,; Monika K. Kaminska, Astronomical Observatory Institute, Faculty of Physics, A.Mickiewicz University,; Mikolaj Kruzynski, Astronomical Observatory Institute, Faculty of Physics, A.Mickiewicz University,; Vadym Savanevych, Department of Systems Engineering Kharkiv National University of Radio Electronics; Oleksandr Kozhukhov, National Space Facilities Control and Test Center of State Space Agency of Ukraine; Vladimir Vlasenko, National Space Facilities Control and Test Center; Eugen Dikov, Research and Design Institute Of Micrography; Artem Dmytrenko, Institute of Astronomy of V.N.Karazin Kharkiv national university; Nikolay Koshkin, Astronomical Observatory of Odessa I.I. Mechnikov National University
Keywords: astrometry, image analysis, ILRS, LEO, optical observations
Abstract:
As more and more space debris accumulate in orbit, accurate measurements of their positions become increasingly crucial for ensuring the safety of satellites and spacecraft, especially in low Earth orbit (LEO) regime. One of the ways to observe space debris is through ground-based optical observations. Measurements of the position of resident space objects (RSOs) have both random and systematic errors. Currently, the magnitude of random measurement errors and systematic differences of measurements between stations is comparable. While the magnitude of random errors can easily be reduced by extensive means of increasing the number of measurements, reducing the influence of systematic differences between station measurements can only be achieved by increasing the number of observing stations, which is usually complex and sometimes impossible. Therefore, the possibility of detecting, measuring, and subsequently reducing the magnitude of systematic measurement errors is of great interest for improving the accuracy of orbit estimation of RSOs.
Today many observers use telescopes with small apertures and a wide field of view for space debris observations. They compensate for an image’s low angular resolution using special image processing pipelines that allow measurements of satellite positions in images with subpixel precision. A typical way of identifying systematic measurement errors is to compare measurements with well-known orbits of reference objects. Unfortunately, this method usually does not allow to identify the source of the origin of systematic errors. In this work, we supplement the traditional way of comparing measurements of reference objects with a known orbit by comparing measurements between two independent implementations of image processing pipelines. This allows us to identify systematic errors associated with the implementation of a specific image-processing pipeline. In this work, we present the comparison of the Ukrainian Lemur and Polish Pozna? Satellite Software Tools (PSST) image processing pipelines.
The goals of the work are:
to present a brief description of both pipelines;
to investigate differences in astrometric measurements produced by the Lemur and PSST image processing pipelines;
to study the residuals between observations and the smoothed or reference orbit for systematic variations in measurements;
to separate the systematic variations in measurements due to image processing methods from those due to equipment and observation process.
We used the same source observation set to compare the Lemur and PSST image processing pipelines. The source images were made using several telescopes equipped with CMOS cameras with global shutter and GNSS based time registration. We selected several reference objects on LEO, MEO, and GSO. Each reference object had independent orbit estimation provided by one or more services: ILRS, DORIS, or IGS. Extensive observational data allows not only to detect the local systematic shifts in residuals, but also to extract the systematic variations in them.
We decomposed the residuals into the along and across directions of the visible satellite path. This allows us to estimate the time bias of each measurement and avoid introducing prior assumptions about a unified time bias for aggregate sets of measurements.
The comparison shows that:
Both pipelines allow us to measure the object coordinates with a precision of about 0.1 pixels.
The magnitude of random measurement errors depends on the satellite elevation.
The pipelines produce very similar results, but occasionally can produce measurements that may have significant systematic differences.
In some cases, we have identified systematic variations in measurement residuals that are similar in both pipelines. We suppose that these systematic variations are due to equipment work peculiarities.
We found that the systematic differences between both pipelines are usually lower than systematic differences between measurements and reference orbits. Therefore, we conclude that the quality of results obtained by the pipelines is high and the remaining residuals are worth investigation. The observation, image processing, and residual analysis methods presented in this work allow us to study the systematic differences between predicted and observed positions of satellites with sub-arcsecond accuracy, including objects on LEO. We suppose that specific systematic variations of our observations presented in the work arent unique, and our description can help other researchers better understand their observation results.
The Lemur image processing pipeline has implementation as a service (https://colitec.space/). So, anybody can use this service for image processing of their observations.
Date of Conference: September 19-22, 2023
Track: SDA Systems & Instrumentation