{"title":"HIGH-PRECISION IN-FLIGHT CALIBRATION USING UNKNOWN LANDMARKS","authors":"A. Tkachenko","doi":"10.34229/1028-0979-2021-4-11","DOIUrl":null,"url":null,"abstract":"An in-flight geometric calibration (further — calibration) is interpreted here as a procedure of making more preceise mutual attitude parameters of the onboard imaging camera and the star tracker. The problem of calibration is solved with using of observations of the landmarks from the orbit. In this work, the landmarks are considered as unknown in the sense that they may be identified on the several snapshots, they may be associated with synchronous data of the star tracker and GPS, but their location in the Earth coordinate frame is unknown. While unknown markers are used, it is more complicated to provide high accuracy of calibration than when geo-referenced markers are observed. In such a situation, improvement of the onboard devices and gauges and increasing of their accuracy strenghtens advisability of agreement of attainable accuracy of calculations while in-flight geometric calibration with accessible measurings accuracy. It concerns properly calibration so as geo-referencing of space snaps using results of calibration. In particular, it is important to consider how accuracy of calibration depends on the accuracy of specific measurings and initial data. Actuality of the considered problem is indisputable. Without its solution, attraction of high-accurate measurings is senseless. A main means of investigation is computer simulanion and analysis of its results. The combined algorithm is proposed for the processing of the calibration measuring equations. It consists of two independent parts. The first one belongs to author of this work and is based on photogrammetric condition of collinearity The second part belongs to D.V. Lebedev and is based on photogrammetric condition of coplanarity. The method of state estimation with high convergence characteristics — fuzzy state observer — is used for resolving of measuring equations. The results of above-mentioned calibration are fully fit for the geo-referencing of the unknown ground objects with acceptable accuracy. Computer simulation had demonsrated good accuracy of the proposed method of the in-flight geometric calibration using unknown landmarks in a combination with high-precise characteristics of used technical means. The simulation had shown the calibration accuracy on the level of 5 arc sec and accuracy of the geo-referencing on the level of 10–20 m. It is fully comparable with accuracy when geo-referenced markers are observated.","PeriodicalId":54874,"journal":{"name":"Journal of Automation and Information Sciences","volume":" ","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2021-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Automation and Information Sciences","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.34229/1028-0979-2021-4-11","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"Engineering","Score":null,"Total":0}
引用次数: 0
Abstract
An in-flight geometric calibration (further — calibration) is interpreted here as a procedure of making more preceise mutual attitude parameters of the onboard imaging camera and the star tracker. The problem of calibration is solved with using of observations of the landmarks from the orbit. In this work, the landmarks are considered as unknown in the sense that they may be identified on the several snapshots, they may be associated with synchronous data of the star tracker and GPS, but their location in the Earth coordinate frame is unknown. While unknown markers are used, it is more complicated to provide high accuracy of calibration than when geo-referenced markers are observed. In such a situation, improvement of the onboard devices and gauges and increasing of their accuracy strenghtens advisability of agreement of attainable accuracy of calculations while in-flight geometric calibration with accessible measurings accuracy. It concerns properly calibration so as geo-referencing of space snaps using results of calibration. In particular, it is important to consider how accuracy of calibration depends on the accuracy of specific measurings and initial data. Actuality of the considered problem is indisputable. Without its solution, attraction of high-accurate measurings is senseless. A main means of investigation is computer simulanion and analysis of its results. The combined algorithm is proposed for the processing of the calibration measuring equations. It consists of two independent parts. The first one belongs to author of this work and is based on photogrammetric condition of collinearity The second part belongs to D.V. Lebedev and is based on photogrammetric condition of coplanarity. The method of state estimation with high convergence characteristics — fuzzy state observer — is used for resolving of measuring equations. The results of above-mentioned calibration are fully fit for the geo-referencing of the unknown ground objects with acceptable accuracy. Computer simulation had demonsrated good accuracy of the proposed method of the in-flight geometric calibration using unknown landmarks in a combination with high-precise characteristics of used technical means. The simulation had shown the calibration accuracy on the level of 5 arc sec and accuracy of the geo-referencing on the level of 10–20 m. It is fully comparable with accuracy when geo-referenced markers are observated.
期刊介绍:
This journal contains translations of papers from the Russian-language bimonthly "Mezhdunarodnyi nauchno-tekhnicheskiy zhurnal "Problemy upravleniya i informatiki". Subjects covered include information sciences such as pattern recognition, forecasting, identification and evaluation of complex systems, information security, fault diagnosis and reliability. In addition, the journal also deals with such automation subjects as adaptive, stochastic and optimal control, control and identification under uncertainty, robotics, and applications of user-friendly computers in management of economic, industrial, biological, and medical systems. The Journal of Automation and Information Sciences will appeal to professionals in control systems, communications, computers, engineering in biology and medicine, instrumentation and measurement, and those interested in the social implications of technology.