Prospects of simultaneous localization and mapping algorithms application in imagery intelligence

摘要

The article contains an analysis of potential prospects of simultaneous localization and mapping (SLAM) algorithmsapplication in imagery intelligence (IMINT). The first part of the paper presents a detailed description of the SLAMproblem. Diverse solutions to the simultaneous localization and mapping problem and related research over the years arepresented. The most promising of SLAM approaches are pointed out. To facilitate SLAM analysis, the problem ispartitioned into three parts. First, various SLAM estimation techniques are characterized. A mathematical theory behindthe usage of parametric filters, non-parametric filters, and least squares method is presented. Further, differences betweenSLAM algorithms are described in terms of various sensors used on-board SLAM platforms for the examination of theenvironment. The examination is commonly addressed as landmark extraction. A separate part of the paper discusses theimage processing in SLAM. The last part of the SLAM analysis is dedicated to various approaches to map presentation.Further, the properties of SLAM techniques are characterized in terms of their potential benefits to IMINT. Prospects ofincreased efficiency, accuracy and safety of intelligence gathering process are discussed.