Abstract
The object detection algorithm is a key component for the autonomous operation of unmanned surface vehicles (USVs). However, owing to complex marine conditions, it is difficult to obtain large-scale, fully labeled surface object datasets. Shipborne sensors are often susceptible to external interference and have unsatisfying performance, compromising the results of traditional object detection tasks. In this paper, a few-shot surface object detection method is proposed based on multimodal sensor systems for USVs. The multi-modal sensors were used for three-dimensional object detection, and the ability of USVs to detect moving objects was enhanced, realizing metric learning-based few-shot object detection for USVs. Compared with conventional methods, the proposed method enhanced the classification results of few-shot tasks. The proposed approach achieves relatively better performance in three sampled sets of well-known datasets, i.e., 2%, 10%, 5% on average precision (AP) and 28%, 24%, 24% on average orientation similarity (AOS). Therefore, this study can be potentially used for various applications where the number of labeled data is not enough to acquire a compromising result.
Highlights
Academic Editors: Richard J.In recent years, with growing global interest in commercial, scientific, and military issues associated with the marine environment, there has been a corresponding growth in demand for the development of unmanned surface vehicles (USVs) with advanced guidance, navigation, and control (GNC) capabilities [1]
Inspired by ResNet50 [12], we introduce a regularization term for background suppression into the feature extractor, which enhances the extraction of foreground object regions and improves the accuracy of the region proposal box; We propose a feature fusion region proposal network (RPN) that utilizes multiple modalities to produce region proposals for small classes; We propose a few-shot learning module based on metric learning, with a better label classification result and a more accurate localization; We propose a key object detection method on the water surface that utilizes multimodal data as the data source
Traditional object detection methods rely heavily on a large amount of labeled data; there is a lack of surface object data in complex marine conditions
Summary
The few-shot object detection network utilizes small-scale datasets (image) to perform the detection of objects, which consists of five components: feature extraction module, proposal generation network, distance compute module, bounding box regression and classification. Inspired by ResNet50 [12], we introduce a regularization term for background suppression into the feature extractor, which enhances the extraction of foreground object regions and improves the accuracy of the region proposal box; We propose a feature fusion region proposal network (RPN) that utilizes multiple modalities to produce region proposals for small classes; We propose a few-shot learning module based on metric learning, with a better label classification result and a more accurate localization; We propose a key object detection method on the water surface that utilizes multimodal data as the data source. Our approach outperformed the state-of-the-art approach by around 2%, 10%, 5% on AP and 28%, 24%, 24% on AOS in three sampled sets of well-known datasets; This study can be potentially used for various applications where the number of labeled data is not enough to acquire a compromising result
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
