Synthetic Enclosed Echoes (SEE)
收藏资源简介:
SEE数据集是一个新型的声纳数据集,旨在提高机器在水下环境中的感知和3D重建能力。该数据集主要由高保真合成声纳数据组成,辅以少量真实世界的声纳数据。为了便于灵活的数据采集,开发了一个模拟环境,可以通过添加新结构或成像声纳配置等修改来生成额外的数据。这种混合方法利用了合成数据的优势,包括易于获得的地面实况和生成多样化数据集的能力,同时通过在类似环境中获取的真实世界数据来弥合模拟与现实的差距。SEE数据集全面评估了基于声学数据的方法,包括基于数学的声纳方法和深度学习算法。这些技术被用于验证数据集,证实了其在水下3D重建中的适用性。此外,本文还提出了对最先进的算法的一种新颖的改进,与现有方法相比,展示了更好的性能。SEE数据集能够在现实场景中评估基于声学数据的方法,从而提高了它们在真实水下应用中的可行性。
The SEE dataset is a novel sonar dataset designed to enhance machines' perception and 3D reconstruction capabilities in underwater environments. It primarily comprises high-fidelity synthetic sonar data, supplemented by a small amount of real-world sonar data. To facilitate flexible data collection, a simulated environment has been developed, which can generate additional data via modifications such as adding new structures or adjusting sonar imaging configurations. This hybrid approach leverages the advantages of synthetic data, including readily accessible ground truth and the ability to generate diverse datasets, while bridging the gap between simulation and reality using real-world data collected in similar environments. The SEE dataset comprehensively evaluates acoustic data-based methods, including mathematical sonar approaches and deep learning algorithms. These techniques were used to validate the dataset, confirming its applicability to underwater 3D reconstruction. Furthermore, this work proposes a novel improvement to state-of-the-art algorithms, which demonstrates superior performance compared to existing methods. The SEE dataset enables the evaluation of acoustic data-based methods in real-world scenarios, thereby enhancing their feasibility in practical underwater applications.

- 1Synthetic Enclosed Echoes: A New Dataset to Mitigate the Gap Between Simulated and Real-World Sonar Data巴西国家石油、天然气和生物燃料人力资源计划 - PRHANP, FINEP 和 CNPq, FURG 联邦大学里约热内卢分校, USP 圣保罗大学数学与计算机科学学院 · 2025年



