Calibration and Comparative Analysis of Forward-Looking Sonar and 3D Sonar for Enhanced Underwater Object Recognition
Computer Vision and Pattern RecognitionRobotics
Summary
The authors explain that sonar devices create a lot of noise, which makes it hard to identify shapes and features in 3D images used for tasks like navigation. They use two kinds of sonar data—2D images and 3D point clouds—and combine them with an automatic calibration method to remove noisy parts. Their tests show that this automatic method works better than manual calibration and helps find features much more clearly. This makes sonar data more useful for recognizing objects or places underwater.
Authors
Aditya Penumarti, Khanh Dong, Zi-Hao Zhang, Yongkyoon Park, Zhenqi Wu, Trung Dong, Shahriar Negahdaripour, Xiaomin Lin, Jane Shin
Abstract
Sonars generate a significant amount of noise. With the advent of new technology capable of producing full 3D point clouds, the noise is amplified in sparse point clouds, making it challenging to recognize features for navigation, recognition, or reconstruction. To address this challenge, we propose using two different sonar modalities: one that produces a 2D intensity image and another that generates a 3D point cloud. By implementing auto-calibration, we can filter out noisy features between the modalities to enhance feature extraction. Experiments demonstrate that auto-calibration improves performance over manual calibration by 5% and that filtering enhances feature extraction by more than 40% relative to the raw point cloud. Code and datasets are given at https://theaprilab.org/fls-3d-calibrator