Object Detection and Location Tracking on Water Surface Using Edge Computing

Object detection and location tracking on water surfaces are critical to improve safety and automation in applications like small boats, autonomous cars, and environmental surveillance. This survey paper compares and examines existing methods for object detection and tracking in aquatic settings, with emphasis on those that use Convolutional Neural Networks (CNNs) and the NVIDIA Jetson Orin Nano Developer Kit. This study explores key phases of system development, such as data collection, preprocessing methods, CNN-based detection techniques, and approaches for determining object position and distance and provides a detailed review of current technologies that are being used in detecting and tracking objects on water surfaces. In addition, the study addresses obstacle avoidance strategies and real-time navigation adjustment mechanisms. Through a critical evaluation of current approaches, the review identifies core challenges such as fluctuating lighting conditions, dynamic backgrounds, and the need for dependable edge computing capabilities. The findings are intended to guide future developments in designing more accurate and efficient real-time systems for water-based environments.