Integration of High-resolution Embedded Camera for Maritime Object Georeferencing

Abstract

In maritime security contexts, it is essential to have an accurate and timely geo-referencing system for maritime objects such as ships and vessels. This is crucial for enhancing situational awareness and ensuring efficient monitoring. Existing systems are often static, and when they are moved, they lose the reference for image transformation. To address this issue, this thesis proposes the integration of a high-resolution camera into a mobile housing unit with an embedded system for image processing and a smartphone for complementary data acquisition, which allows ship georeferencing in a mobile manner.

The ship recognition software used in this system is YOLOv8, and the Android smartphone collects sensor data. This configuration facilitates accurate conversion of pixel coordinates of the recognized ships into geographic locations using raycasting as georeferencing method, which enables display of latitudes and longitudes of ships on a map. The experimental results validate that the detection and georeferencing of ships is performed with an error of 16±7 meters using the integrated system. These results are comparable to those of static camera systems, however adding the potential of mobile embedded systems to further enhance maritime monitoring capabilities. This innovative solution is a step forward in improving the accuracy and efficiency of maritime monitoring systems.