American Journal of Mechanical Engineering, 2020, Vol. 8, No. 2, 76-87 Available online at http://pubs.sciepub.com/ajme/8/2/5 Published by Science and Education Publishing DOI:10.12691/ajme-8-2-5 Monocular Visual SLAM for Underwater Navigation in Turbid and Dynamic Environments Chinthaka Amarasinghe 1,* , Asanga Ratnaweera 2 , Sanjeeva Maitripala 2 1 Department of Science & Technology, Uva Wellassa University, Badulla, Sri Lanka 2 Department of Mechanical Engineering, University of Peradeniya, Peradeniya, Sri Lanka *Corresponding author: chinthakaa@uwu.ac.lk Received July 17, 2020; Revised August 19, 2020; Accepted August 28, 2020 Abstract Localization, navigation, and mapping using vision-based algorithms are an active topic in underwater robotic applications. Although many algorithms developed in recent years, especially in the ground and areal robotic communities, directly applying those methods in underwater navigation remain challenging due to the visual degradation induced by the medium. In this paper, we proposed UW-SLAM (Underwater SLAM), a new monocular visual SLAM algorithm focused on the underwater environment which addresses the turbidity and dynamism. The proposed method was evaluated with several underwater datasets with comparison to the state of the art monocular SLAM methods. Keywords: monocular visual navigation, underwater vision, visual SLAM Cite This Article: Chinthaka Amarasinghe, Asanga Ratnaweera, and Sanjeeva Maitripala, “Monocular Visual SLAM for Underwater Navigation in Turbid and Dynamic Environments.” American Journal of Mechanical Engineering, vol. 8, no. 2 (2020): 76-87. doi: 10.12691/ajme-8-2-5. 1. Introduction Exploration of the oceans and shallow waters is attracting the interest of many industries and institutions all over the world, because of the valuable resources, the knowledge that it houses for scientists, and also for rescue purposes. For the past decades, remotely operated vehicles (ROV) are the widely used method for the exploration of the underwater environment. ROVs operated using a wired connection between the operator and the vehicle which limits usability and maneuverability. Due to these limitations, ROVs are now replaced by Autonomous Underwater Vehicles (AUVs). Although AUVs offers unique advantages over ROV and also present a uniquely challenging navigational problem as they operate autonomously in a highly unstructured environment where satellite-based navigation isn’t directly available [1]. Navigation plays a significant role in the operation of AUVs and consists of two fundamental aspects localization and mapping [2]. Currently used methods for AUV navigation can be grouped into three categories [1]. 1. Inertial / Dead Reckoning 2. Acoustic transponders and modems techniques 3. Geophysical navigation An inertial navigation system (INS) is navigation that uses a processing unit, motion sensors (accelerometers), rotational sensors (gyroscopes), and magnetic sensors (magnetometers) to continuously calculate the dead reckoning the velocity, orientation, and the position of the moving object without any need for external references. All of the methods in this category have position error growing with time and need to be corrected by an external reference. Acoustic transponders and modems are used to measure the time-of-flight (TOF) of the sound signals underwater. Different types of acoustic-based sensors such as Doppler Velocity Log (DVL), Mechanically Scanning Imaging Sonar (MSIS), Underwater Acoustic Positioning System (UAPS), bathymetric sonars, Side scan, etc have been developed. On the other hand, geophysical navigation techniques need external environmental information as references for navigation. This is achieved by detecting, identifying, and classifying of environmental features by using various kinds of sensors such as Cameras, Laser range finders, magnetic sensors, pressure/depth sensors, and processing those sensor data with effective fusion algorithms [1]. Sonar-based (Acoustic Navigation) methods are the most extended approach in the underwater robotics community, because of the good properties of sound propagation in the water. However, these are more suited for long distances underwater missions and not for short-range missions (below 1 m) [1]. The high cost of the acoustic sensors and the infrastructure required for acoustic-based sensing limit their usability in small-scale AUVs and bio-inspired vehicles such as Robotic fish [2]. Further, most of the underwater missions are used for inspection purposes that require sub metric accuracy, which is difficult and expensive to achieve with acoustic type sensors. However, when navigating close to the seabed or the inspection structures, visual information becomes available and cameras