Author Contributions
Conceptualization, J.H. and J.C.; methodology, J.H. and J.C.; validation, J.H. and H.X.; writing—original draft preparation, J.H.; writing—review and editing, J.H. and M.S.A.; visualization, J.H.; supervision, J.C.; project administration, J.C.; funding acquisition, J.C. All authors have read and agreed to the published version of the manuscript.
Figure 1.
Side-scan sonar images and their 3D visualization. (a–c) The first row shows three side-scan sonar images and the second row shows the 3D visualization results of the corresponding images with different SCR conditions, in which the red rectangles are small target samples from different areas of the seafloor.
Figure 1.
Side-scan sonar images and their 3D visualization. (a–c) The first row shows three side-scan sonar images and the second row shows the 3D visualization results of the corresponding images with different SCR conditions, in which the red rectangles are small target samples from different areas of the seafloor.
Figure 2.
The flowchart of our proposed algorithm based on low-rank sparse matrix factorization. The algorithm can be divided into three parts: image preprocessing, matrix factorization, and morphological operation. The red rectangular box is the detection result.
Figure 2.
The flowchart of our proposed algorithm based on low-rank sparse matrix factorization. The algorithm can be divided into three parts: image preprocessing, matrix factorization, and morphological operation. The red rectangular box is the detection result.
Figure 3.
Original SSS image and preprocessed image. (a) A representative SSS image. (b) The preprocessed image. (c) The histogram of the original SSS image (a). (d) The histogram of the preprocessed image (b).
Figure 3.
Original SSS image and preprocessed image. (a) A representative SSS image. (b) The preprocessed image. (c) The histogram of the original SSS image (a). (d) The histogram of the preprocessed image (b).
Figure 4.
Sonar background images and their singular value results. The first row (a–c) shows three sonar background images and the second row (d–f) shows the singular values of the corresponding preprocessed images. (a–c) The background images with a different submarine environment, and the black area is an acoustic shadow in subfigure (c).
Figure 4.
Sonar background images and their singular value results. The first row (a–c) shows three sonar background images and the second row (d–f) shows the singular values of the corresponding preprocessed images. (a–c) The background images with a different submarine environment, and the black area is an acoustic shadow in subfigure (c).
Figure 5.
The area proportion of the target in the whole image for three sequences. The proportion of small targets in SSS images occupies very few pixels compared to the total image.
Figure 5.
The area proportion of the target in the whole image for three sequences. The proportion of small targets in SSS images occupies very few pixels compared to the total image.
Figure 6.
Illustration of the experimental scheme. The experiment collects SSS images under various conditions including various scanning heights, scanning angles, and orientation angles.
Figure 6.
Illustration of the experimental scheme. The experiment collects SSS images under various conditions including various scanning heights, scanning angles, and orientation angles.
Figure 7.
Overall structure of AUV. The AUV consists of seven parts: head section, side-scan sonar, navigation and control unit, propulsion unit I, propulsion unit II, tail section, and towrope.
Figure 7.
Overall structure of AUV. The AUV consists of seven parts: head section, side-scan sonar, navigation and control unit, propulsion unit I, propulsion unit II, tail section, and towrope.
Figure 8.
The obtained SSS image containing the small targets. They were collected in different waters.
Figure 8.
The obtained SSS image containing the small targets. They were collected in different waters.
Figure 9.
The ROC curves of eight algorithms for the three real image sequences. (a) Real Seq. 1, (b) Real Seq. 2, (c) Real Seq. 3.
Figure 9.
The ROC curves of eight algorithms for the three real image sequences. (a) Real Seq. 1, (b) Real Seq. 2, (c) Real Seq. 3.
Figure 10.
3D visualization of the sonar image detection result. The first row shows three input sonar images (the red box in the image is the target), the second row shows the input SSS images plotted in 3D (across-track, along-track, and backscattering strength), and the third row shows the detection results plotted in 3D (across-track, along-track, and backscattering strength).
Figure 10.
3D visualization of the sonar image detection result. The first row shows three input sonar images (the red box in the image is the target), the second row shows the input SSS images plotted in 3D (across-track, along-track, and backscattering strength), and the third row shows the detection results plotted in 3D (across-track, along-track, and backscattering strength).
Figure 11.
Visualization result of different algorithms in three sequences. The red rectangular box is the ground truth and the yellow rectangular box is the detection result.
Figure 11.
Visualization result of different algorithms in three sequences. The red rectangular box is the ground truth and the yellow rectangular box is the detection result.
Figure 12.
ROC curves of (a) Sequence 1, (b) Sequence 2, (c) Sequence 3.
Figure 12.
ROC curves of (a) Sequence 1, (b) Sequence 2, (c) Sequence 3.
Figure 13.
Probability of detection for the (a) = 0.5, (b) = 1, (c) = 1.8.
Figure 13.
Probability of detection for the (a) = 0.5, (b) = 1, (c) = 1.8.
Figure 14.
Visualization results of the ablation experiment. The red rectangular box is the ground truth.
Figure 14.
Visualization results of the ablation experiment. The red rectangular box is the ground truth.
Table 1.
Specifications of the side-scan sonar.
Table 1.
Specifications of the side-scan sonar.
Parameter | Value |
---|
Operating frequency | 260 kHz/800 kHz |
Transducer beam width | 260 kHz: 2.2° × 75°/800 kHz: 0.7° × 30° |
Range resolution | 0.2% of range |
Maximum imaging range | 100 m |
Image size | 2000 × 500 pixels |
Frame rate | 6–20 fps |
Depth rating | 300 m |
Table 2.
Details of the three sequences.
Table 2.
Details of the three sequences.
Sequence | Number | The Target Size in the Whole Image | The Mean of SCR | The Variances of SCR |
---|
Seq1 (High SCR) | 100 | 0.011 | 6.36 | 0.16 |
Seq2 (Medium SCR) | 100 | 0.012 | 4.91 | 0.53 |
Seq3 (Low SCR) | 100 | 0.009 | 1.08 | 0.37 |
Table 3.
Probability of detection for different algorithms.
Table 3.
Probability of detection for different algorithms.
| Method | (Seq1) | (Seq2) | (Seq3) |
---|
0.5 | IPI | 0.850 | 0.852 | 0.711 |
LCM | 0.825 | 0.855 | 0.666 |
AMWLCM | 0.599 | 0.778 | 0.140 |
ILCM | 0.640 | 0.853 | 0.583 |
TOPHAT | 0.598 | 0.245 | 0.350 |
MAXMEAN | 0.622 | 0.511 | 0.347 |
ACA_CFAR | 0.661 | 0.781 | 0.525 |
OURS | 0.895 | 0.900 | 0.732 |
1.0 | IPI | 0.877 | 0.875 | 0.743 |
LCM | 0.861 | 0.887 | 0.754 |
AMWLCM | 0.625 | 0.823 | 0.210 |
ILCM | 0.733 | 0.877 | 0.665 |
TOPHAT | 0.702 | 0.315 | 0.379 |
MAXMEAN | 0.721 | 0.547 | 0.465 |
ACA_CFAR | 0.750 | 0.799 | 0.665 |
OURS | 0.920 | 0.902 | 0.783 |
Table 4.
Probability of detection for the ablation experiment.
Table 4.
Probability of detection for the ablation experiment.
| Image Preprocessing | RPCA | Improved RPCA | Morphological Operation | (Seq1) | (Seq2) | (Seq3) |
---|
| ✘ | ✔ | ✘ | ✘ | 0.474 | 0.368 | 0.301 |
0.5 | ✘ | ✘ | ✔ | ✘ | 0.526 | 0.672 | 0.579 |
| ✔ | ✘ | ✔ | ✘ | 0.862 | 0.840 | 0.706 |
| ✔ | ✘ | ✔ | ✔ | 0.895 | 0.900 | 0.732 |
| ✘ | ✔ | ✘ | ✘ | 0.737 | 0.605 | 0.403 |
1.0 | ✘ | ✘ | ✔ | ✘ | 0.837 | 0.805 | 0.753 |
| ✔ | ✘ | ✔ | ✘ | 0.899 | 0.881 | 0.737 |
| ✔ | ✘ | ✔ | ✔ | 0.920 | 0.902 | 0.783 |
| ✘ | ✔ | ✘ | ✘ | 0.868 | 0.642 | 0.516 |
1.8 | ✘ | ✘ | ✔ | ✘ | 0.842 | 0.811 | 0.758 |
| ✔ | ✘ | ✔ | ✘ | 0.921 | 0.887 | 0.780 |
| ✔ | ✘ | ✔ | ✔ | 0.947 | 0.903 | 0.815 |
Table 5.
Computational cost comparison among the proposed algorithm and other algorithms. (unit: s).
Table 5.
Computational cost comparison among the proposed algorithm and other algorithms. (unit: s).
Sequence | IPI | LCM | AMWLCM | ILCM | TOPHAT | MAXMEAN | ACA_CFAR | OURS |
---|
1 | 1.832 | 0.309 | 4.902 | 0.270 | 0.260 | 0.316 | 33.499 | 1.853 |
2 | 2.799 | 0.371 | 6.360 | 0.319 | 0.275 | 0.304 | 46.8 | 2.943 |
3 | 1.486 | 0.303 | 3.972 | 0.280 | 0.257 | 0.288 | 28.754 | 1.495 |