Maximizing target detection under sunlight reflection on water surfaces with an autonomous unmanned aerial vehicle

Abstract

Reflected sunlight can significantly impact vision-based object detection and tracking algorithms, especially ones based on an aerial platform operating over a marine environment. Unmanned aerial systems above a water surface may be unable to detect objects on the water surface due to sunlight glitter. Although the area affected by sunlight reflection may be limited, rapid course corrections of unmanned aerial vehicles (UAVs) — especially fixed-wing UAVs — is also limited by aerodynamics, making it challenging to determine a reasonable path that avoids sunlight reflection while maximizing chances to capture a target. In this paper, we propose an approach for autonomous UAV path planning that maximizes the accuracy of the estimated target location by minimizing the sunlight reflection influences.