Image feature extraction based on improved FCN for UUV side-scan sonar
- PDF / 4,266,458 Bytes
- 17 Pages / 595.276 x 790.866 pts Page_size
- 63 Downloads / 210 Views
ORIGINAL RESEARCH PAPER
Image feature extraction based on improved FCN for UUV side‑scan sonar Hongjian Wang1 · Na Gao2 · Yao Xiao1 · Yanghua Tang1 Received: 25 July 2020 / Accepted: 3 October 2020 © Springer Nature B.V. 2020
Abstract Current methods for edge contour feature extraction for Unmanned Underwater Vehicle (UUV) side-scan sonar images have yet to solve the problems of low accuracy, discontinuous edges, and loss of detail. This paper proposes a new feature extraction method for UUV side-scan sonar images. By adding a batch normalization layer, the skip structure of the fully convolutional network (FCN) is improved, the redistribution of parameters in the skip structure is realized, and the training of the network is more sufficient. And we design a positive sample weighted loss function (WPSL) to improve the problem that the performance of the classification algorithm is degraded due to the imbalance of sample distribution in the data set. In this paper, an initial dataset is expanded by turning, rotating, and adding noise. An improved feature extraction network is then constructed, and the training of the improved FCN is completed by using a mini-batch gradient descent method, thus realizing accurate extraction of edge contour features of seabed topography. The experimental results show that the proposed method is more suitable to reject speckle noise than the traditional Canny and Fuzzy C-Means algorithms. Compared with current deep learning methods, the proposed method improves the ability to fuse detailed information and make discontinuous edges continuous. The mean intersection over union (IU) reaches 83.05%, which is 5.48% higher than the 77.57% before improvement. Keywords Feature extraction · Fully convolutional network · Unmanned underwater vehicle · Side-scan sonar
Introduction As one of the most effective sensors for underwater detection, the side-scan sonar system is an important tool for UUV detecting seabed topography (Flemming 2015). The detection and feature extraction technology of UUV is the premise of all marine activities and plays an extremely important role in marine engineering, the development of marine resources and science, and the protection of marine rights and interests. However, the interpretation of sonar images has mostly been dependent on manual work. When the amount of data increases rapidly, manual recognition methods are inefficient and it is difficult to outline the target contour accurately. For unmanned operating UUV, sonar * Hongjian Wang [email protected] 1
College of Intelligent Science and Engineering, Harbin Engineering University, Harbin 150001, Heilongjiang, China
Hisense Electronic Information Research and Design Center, Qingdao 266000, Shandong, China
2
automatic image processing function is particularly important. It is therefore of urgent importance to explore an accurate and efficient contour feature extraction method. The key to feature extraction of edge contour is edge detection and segmentation; that is, the process of dividing images i
Data Loading...