Specific target recognition and segmentation algorithm for real-time side scan sonar images

Author(s):  
Lei Wang ◽  
Minhui Li ◽  
Xiufen Ye ◽  
Tian Wang
2021 ◽  
Vol 13 (18) ◽  
pp. 3555
Author(s):  
Yongcan Yu ◽  
Jianhu Zhao ◽  
Quanhua Gong ◽  
Chao Huang ◽  
Gen Zheng ◽  
...  

To overcome the shortcomings of the traditional manual detection of underwater targets in side-scan sonar (SSS) images, a real-time automatic target recognition (ATR) method is proposed in this paper. This method consists of image preprocessing, sampling, ATR by integration of the transformer module and YOLOv5s (that is, TR–YOLOv5s), and target localization. By considering the target-sparse and feature-barren characteristics of SSS images, a novel TR–YOLOv5s network and a down-sampling principle are put forward, and the attention mechanism is introduced in the method to meet the requirements of accuracy and efficiency for underwater target recognition. Experiments verified the proposed method achieved 85.6% mean average precision (mAP) and 87.8% macro-F2 score, and brought 12.5% and 10.6% gains compared with the YOLOv5s network trained from scratch, and had the real-time recognition speed of about 0.068 s per image.


Sensors ◽  
2019 ◽  
Vol 19 (9) ◽  
pp. 1985
Author(s):  
Qi Wang ◽  
Meihan Wu ◽  
Fei Yu ◽  
Chen Feng ◽  
Kaige Li ◽  
...  

Real-time processing of high-resolution sonar images is of great significance for the autonomy and intelligence of autonomous underwater vehicle (AUV) in complex marine environments. In this paper, we propose a real-time semantic segmentation network termed RT-Seg for Side-Scan Sonar (SSS) images. The proposed architecture is based on a novel encoder-decoder structure, in which the encoder blocks utilized Depth-Wise Separable Convolution and a 2-way branch for improving performance, and a corresponding decoder network is implemented to restore the details of the targets, followed by a pixel-wise classification layer. Moreover, we use patch-wise strategy for splitting the high-resolution image into local patches and applying them to network training. The well-trained model is used for testing high-resolution SSS images produced by sonar sensor in an onboard Graphic Processing Unit (GPU). The experimental results show that RT-Seg can greatly reduce the number of parameters and floating point operations compared to other networks. It runs at 25.67 frames per second on an NVIDIA Jetson AGX Xavier on 500*500 inputs with excellent segmentation result. Further insights on the speed and accuracy trade-off are discussed in this paper.


IEEE Access ◽  
2020 ◽  
Vol 8 ◽  
pp. 173450-173460
Author(s):  
Tang Yulin ◽  
Shaohua Jin ◽  
Gang Bian ◽  
Yonghou Zhang

IEEE Access ◽  
2021 ◽  
Vol 9 ◽  
pp. 29416-29428
Author(s):  
Xiaoming Qin ◽  
Xiaowen Luo ◽  
Ziyin Wu ◽  
Jihong Shang

2021 ◽  
Author(s):  
Théo Le Moigne ◽  
Libero Gurrieri ◽  
Pierre Crozet ◽  
Christophe H. Marchand ◽  
Mirko Zaffagnini ◽  
...  

1992 ◽  
Vol 14 (2) ◽  
pp. 125-136 ◽  
Author(s):  
D. C. Mason ◽  
T. P. LeBas ◽  
I. Sewell ◽  
C. Angelikaki
Keyword(s):  

2019 ◽  
Vol 294 (46) ◽  
pp. 17437-17450 ◽  
Author(s):  
Yuichi Yokochi ◽  
Kazunori Sugiura ◽  
Kazuhiro Takemura ◽  
Keisuke Yoshida ◽  
Satoshi Hara ◽  
...  

Thioredoxin (Trx) is a redox-responsive protein that modulates the activities of its target proteins mostly by reducing their disulfide bonds. In chloroplasts, five Trx isoforms (Trx-f, Trx-m, Trx-x, Trx-y, and Trx-z) regulate various photosynthesis-related enzymes with distinct target selectivity. To elucidate the determinants of the target selectivity of each Trx isoform, here we investigated the residues responsible for target recognition by Trx-f, the most well-studied chloroplast-resident Trx. As reported previously, we found that positively-charged residues on the Trx-f surface are involved in the interactions with its targets. Moreover, several residues that are specifically conserved in Trx-f (e.g. Cys-126 and Thr-158) were also involved in interactions with target proteins. The validity of these residues was examined by the molecular dynamics simulation. In addition, we validated the impact of these key residues on target protein reduction by studying (i) Trx-m variants into which we introduced the key residues for Trx-f and (ii) Trx-like proteins, named atypical Cys His-rich Trx 1 (ACHT1) and ACHT2a, that also contain these key residues. These artificial or natural protein variants could reduce Trx-f–specific targets, indicating that the key residues for Trx-f are critical for Trx-f–specific target recognition. Furthermore, we demonstrate that ACHT1 and ACHT2a efficiently oxidize some Trx-f–specific targets, suggesting that its target selectivity also contributes to the oxidative regulation process. Our results reveal the key residues for Trx-f–specific target recognition and uncover ACHT1 and ACHT2a as oxidation factors of their target proteins, providing critical insight into redox regulation of photosynthesis.


Sign in / Sign up

Export Citation Format

Share Document