Exploring deep neural networks for real-world ship detection using scaled model images and chroma key technology

Abstract This paper presents the development and evaluation of a deep neural network model for the detection of naval surface vessel using laboratory-generated datasets. By employing chroma-key technology, images of a scale model naval vessel were superimposed onto realistic maritime backgrounds to create a diverse training dataset. Fine-tuned with these datasets and evaluated using the YOLOv8 framework, the model achieved high precision and recall in identifying the naval surface vessel despite data limitations. This zero-shot learning approach, validated through extensive testing, supports visual navigation and target identification in GPS/RF-denied environments, advancing autonomous maritime operations and aligning with the United States Navy strategy to leverage AI/ML for military enhancement.