Reconstructing Seabed Topography from Side-Scan Sonar Images with Self-Constraint

Zhao, Jianhu; Shang, Xiaodong; Zhang, Hongmei

doi:10.3390/rs10020201

Cited by 11 publications

(16 citation statements)

References 32 publications

(48 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the past few years, extensive studies have shown that once a small amount of prior knowledge about the data is incorporated, the performance of clustering can be improved greatly [28], [43]. In order to use the background knowledge of data points reasonably, semi-supervised clustering is proposed.…”

Section: A Clustering and Semi-supervised Clustering Variantsmentioning

confidence: 99%

Robust Semi-Supervised Non-Negative Matrix Factorization With Structured Normalization

Wang

Guan²,

Shi³

et al. 2019

IEEE Access

View full text Add to dashboard Cite

Non-negative matrix factorization (NMF) approximates a non-negative data matrix with the product of two low-rank non-negative matrices by minimizing the cost of such approximation. However, traditional NMF models cannot be generalized in the cases when the dataset contains outliers and limited knowledge from domain experts. In this paper, we propose a robust semi-supervised NMF model (RSS-NMF) to overcome the aforementioned deficiency. RSS-NMF utilizes the L 2 /L 1-norm to encourage approximation and makes the model insensitive to outliers by prohibiting them from dominating the cost function. To incorporate the discriminative information, RSS-NMF utilizes the structured normalization method when learns a diagonal matrix to normalize the coefficients such that they get close to the label indicators of the given labeled examples. Although the multiplicative update rule (MUR) can be adopted to minimize RSS-NMF, it converges slowly. In this paper, we adopt a fast gradient descent algorithm (FGD) to optimize RSS-NMF and prove its convergence to a stationary point. FGD uses a Newton method to search the optimal step length and thus, FGD converges faster than MUR. The experimental results show the promise of RSS-NMF comparing with the representative clustering models on several face image datasets. INDEX TERMS Non-negative matrix factorization, semi-supervised learning, L 2 /L 1-norm, structured normalization.

show abstract

Section: A Clustering and Semi-supervised Clustering Variantsmentioning

confidence: 99%

Robust Semi-Supervised Non-Negative Matrix Factorization With Structured Normalization

Wang

Guan²,

Shi³

et al. 2019

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Intensity of acoustic return converted to grey-level pixel value (0-255) as seabed image of side scan sonar. (Cervenka and de Mustier, 1993;Daniel, et al, 1998;Moreno et al, 2014, Burguera andOliver, 2015;Zhao et al, 2017. ).…”

Section: Introductionmentioning

confidence: 99%

Seabed Characterization through Image Processing of Side Scan Sonar Case Study: Bontang and Batam

Subarsyah¹,

Arifin²

2019

BoMG

View full text Add to dashboard Cite

Acoustic waves propagate through a medium meet the Snell’s Law, its energy is reflected and some are scattered back at certain angle. The Side Scan Sonar (SSS) methods use this principle to identify seabed character. The intensity of the backscatter greatly depends on the morphology and sediments texture or rocks distributed on seabed.The intensity of backscatter waves is a representation of the morphology, sediments texture, and types of rock that distributed on the seabed, therefore it is possible to estimate sedimentary texture and identify the presence of rocks or coral reefs based on this information. In this publication authors estimate sediments texture, rocks or coral reefs based on backscatter intensity through the image processing on the Side Scan Sonar (SSS) image. Intensity will be converted into pixel values on the image with range value 1-255 (gray scale image) and entropy values which are statistical measures of randomness. Entropy value is maximum when most of pixel value image is in the middle of the colour spectrum range (between very dark to very bright), in contrast, it is minimum when pixel value is in the spectrum of very dark or very bright. Based on both parameters, classification is conducted. The classification is carried out on the SSS image at Bontang and Batam that have very different seabed characters.The classification results using an image processing shows that the distribution of sediment textures consist of 4 (four) classes for either Batam or Bontang. In the Bontang area, very fine sediments were identified which are associated with low value of both intensity and entropy - dark zones in gray scale images, and coarse sediments associated with high value of both intensity and entropy - bright zone in the gray scale image. Similar characteristic is observed in Batam area, which are identified fine sediment (associated to low intensity) - coarse sediments (high intensity). In contrast to Bontang, in Batam the entropy exhibit the opposite value, high value are correlated to fine sediment and vice versa. This might be due to the presence of rocks and sedimentary structures.Keywords: Side Scan Sonar, Intensity, Backscatter and entropy.Gelombang akustik sebagian besar energinya dipantulkan memenuhi prinsip snellius dan sebagian kecil dihamburkan balik dengan sudut. Metode Side Scan Sonar (SSS) memanfaatkan prinsip hambur-balik gelombang untuk mengidentifikasi permukaan dasar laut. Intensitas gelombang dari karakter hambur-balik akan sangat tergantung morfologi dan tekstur sedimen atau batuan dari permukaan dasar lautnya. Intensitas gelombang hambur-balik merupakan representasi dari morfologi, tekstur sedimen, dan jenis batuan yang tersebar di permukaan dasar laut, sehingga sangat memungkinkan untuk melakukan estimasi tekstur sedimen dan identifikasi keberadaan batuan maupun terumbu karang berdasarkan informasi tersebut. Pada publikasi ini akan dilakukan estimasi tekstur sedimen atau batuan berdasarkan intensitas hambur-balik melalui image yang dihasilkan oleh Metode Side Scan Sonar (SSS). Intensitas akan dikonversi ke dalam nilai pixel dalam image dengan rentang nilai 1-255 (gray scale image) dan nilai entropi yang merupakan ukuran statistik ketidakteraturan dari image. Entropi akan maksimum ketika nilai pixel kebanyakan di tengah dari rentang spektrum warna dan sebaliknya akan minimum ketika nilai pixelnya berada di spektrum warna sangat gelap atau sangat terang. Berdasarkan kedua parameter tersebut, kemudian dilakukan klasifikasi. Klasifikasi dilakukan pada data SSS Bontang dan Batam yang memiliki karakter permukaan dasar laut yang sangat berbeda.Hasil klasifikasi dengan image processing memperlihatkan pola sebaran tekstur sedimen masing-masing terdiri dari 4 (empat) kelas baik untuk Batam atau Bontang. Pada area Bontang teridentifikasi sedimen sangat halus yang berasosiasi dengan intensitas dan entropy rendah - zona gelap pada gray scale image dan sedimen kasar yang berasosiasi dengan intensitas dan entropy tinggi - zona terang pada gray scale image. Karakter yang sama juga teramati pada area Batam, yaitu teridentifikasi sedimen halus (berasosiasi dengan intensitas rendah) - sedimen kasar (intensitas tinggi). Namun, berbeda dengan di Bontang, di Batam nilai entropi menunjukkan nilai yang sebaliknya, yaitu nilai tinggi berkorelasi dengan sedimen halus, dan sebaliknya. Hal ini diperkirakan akibat keberadaan batuan dan struktur sedimen.Kata Kunci: Side Scan Sonar, Intensitas, Hambur balik dan Entropi.

show abstract

“…Many previous works have proposed different methods to reconstruct bathymetry from sidescan [ 3 , 4 , 8 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 ]. They can be categorized into different groups according to different criteria, see Table 1 .…”

Section: Introductionmentioning

confidence: 99%

Sidescan Only Neural Bathymetry from Large-Scale Survey

Xie

Bore

Folkesson

2022

Sensors

View full text Add to dashboard Cite

Sidescan sonar is a small and low-cost sensor that can be mounted on most unmanned underwater vehicles (UUVs) and unmanned surface vehicles (USVs). It has the advantages of high resolution and wide coverage, which could be valuable in providing an efficient and cost-effective solution for obtaining the bathymetry when bathymetric data are unavailable. This work proposes a method of reconstructing bathymetry using only sidescan data from large-scale surveys by formulating the problem as a global optimization, where a Sinusoidal Representation Network (SIREN) is used to represent the bathymetry and the albedo and the beam profile are jointly estimated based on a Lambertian scattering model. The assessment of the proposed method is conducted by comparing the reconstructed bathymetry with the bathymetric data collected with a high-resolution multi-beam echo sounder (MBES). An error of 20 cm on the bathymetry is achieved from a large-scale survey. The proposed method proved to be an effective way to reconstruct bathymetry from sidescan sonar data when high-accuracy positioning is available. This could be of great use for applications such as surface vehicles with Global Navigation Satellite System (GNSS) to obtain high-quality bathymetry in shallow water or small autonomous underwater vehicles (AUVs) if simultaneous localization and mapping (SLAM) can be applied to correct the navigation estimate.

show abstract

Reconstructing Seabed Topography from Side-Scan Sonar Images with Self-Constraint

Cited by 11 publications

References 32 publications

Robust Semi-Supervised Non-Negative Matrix Factorization With Structured Normalization

Robust Semi-Supervised Non-Negative Matrix Factorization With Structured Normalization

Seabed Characterization through Image Processing of Side Scan Sonar Case Study: Bontang and Batam

Sidescan Only Neural Bathymetry from Large-Scale Survey

Contact Info

Product

Resources

About