ABSTRACT:This paper will introduce the goals, concept and results of the project named CLOSE-SEARCH, which stands for 'Accurate and safe EGNOS-SoL Navigation for UAV-based low-cost Search-And-Rescue (SAR) operations'. The main goal is to integrate a medium-size, helicopter-type Unmanned Aerial Vehicle (UAV), a thermal imaging sensor and an EGNOS-based multi-sensor navigation system, including an Autonomous Integrity Monitoring (AIM) capability, to support search operations in difficult-to-access areas and/or night operations. The focus of the paper is three-fold. Firstly, the operational and technical challenges of the proposed approach are discussed, such as ultra-safe multi-sensor navigation system, the use of combined thermal and optical vision (infrared plus visible) for person recognition and Beyond-Line-Of-Sight communications among others. Secondly, the implementation of the integrity concept for UAV platforms is discussed herein through the AIM approach. Based on the potential of the geodetic quality analysis and on the use of the European EGNOS system as a navigation performance starting point, AIM approaches integrity from the precision standpoint; that is, the derivation of Horizontal and Vertical Protection Levels (HPLs, VPLs) from a realistic precision estimation of the position parameters is performed and compared to predefined Alert Limits (ALs). Finally, some results from the project test campaigns are described to report on particular project achievements. Together with actual Search-and-Rescue teams, the system was operated in realistic, user-chosen test scenarios. In this context, and specially focusing on the EGNOS-based UAV navigation, the AIM capability and also the RGB/thermal imaging subsystem, a summary of the results is presented.
Since the advent of the first large format digital aerial cameras, high expectations have been placed on their performance. The dream of obtaining aerial images virtually free of geometric errors and with greater radiometric quality is getting close. Nevertheless, systematic image residuals, unexpected height errors in aerial triangulation and the need for additional self‐calibration parameters have been reported since 2005. In this paper a preliminary analysis of the theoretical accuracies in aerial triangulation using the Zeiss/Intergraph (Z/I) Digital Mapping Camera (DMC) and an analogue camera is conducted, motivated by those recent reports. This analysis considers a mathematical model where the image has conical geometry and is free of systematic errors. The influence on the propagated block accuracy of the base‐to‐height ratio, image pointing precision (both manual and automatic), GPS observations for projection centres and of pass/tie point density is studied. Moreover, the expected accuracy in the aerial triangulation of analogue images using current procedures (having regard to the a priori accuracy for image pointing, ground control measurement and GPS and pass/tie point density) is computed. The goal of this theoretical study is to find the requirements for aerial triangulation with DMC data which would yield the same or an even higher level of accuracy than that obtained with analogue data under the same conditions.
The paper continues with a check on the conclusions of this theoretical analysis, using real data‐sets and aerial triangulation set‐up, which fit with the theoretical analysis. The results prove that the expected theoretical accuracy in aerial triangulation is only obtained if an appropriate self‐calibration parameter set is considered in the bundle block adjustment and/or if good GPS observations are available. These requirements result from the unfavourable propagation from unmodelled systematic error in the DMC image blocks. Some authors have detected systematic residuals in the order of one‐tenth of a pixel rms in DMC image space. For this reason, investigations are being carried out on systematic error characterisation, distribution in image space and stability over time and flying height, and systematic error modelling, using self‐calibration parameter sets and applying correction grids. Finally, conclusions are drawn from the investigations.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.