Visual Simultaneous Localization and Mapping: A review

Abstract

The complexity of Simultaneous Localization and Mapping (SLAM) and drastic technological changes makes development of an autonomous robot system a challenge. As the market demand and awareness of autonomous robot systems are increased, intensive knowledge and information of this technology are essential. The fluidity of information has resulted in a mismatch between SLAM and Visual SLAM where the current approaches are having problems in visual odometry and sensor inputs that describe the common characteristics with different techniques due to the lack of formalism in the knowledge on the localization and mapping. This has caused confusion to the researchers leading to incorrect development of autonomous robot system. Therefore, this review aims to describe and assist researchers to identify the characteristics of SLAM and in the context of the visual odometry characteristics and perception techniques. This review will also discuss the attributes of the visual odometry and the techniques of autonomous robot system from various types of application based on the semantic structure. It is expected that the Visual Simultaneous Localization and Mapping (VSLAM) will give a positive impact to the development of autonomous robot system to meet the market demand.