.png){kind=icon}
.png){kind=icon}
Return
Seabed Sediment Classification Based on Multibeam Data Using a Multifeature Fusion Algorithm
.png){kind=icon}
PRE
AI
DOI：10.1109/JOE.2025.3636403.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The seabed sediment serves as a critical indicator of seabed conditions, and its accurate classification is essential for rapid assessment of the seabed environment. In this article, we propose a novel multifeature fusion algorithm based on bathymetry and backscatter intensity data acquired by a multibeam bathymetry system. To address the challenge of the limited availability of labeled samples, a sample enhancement algorithm is proposed to effectively expand the data set while preserving sample consistency. To improve the feature expression ability of single-source data, we extract various texture features from the backscatter intensity data using the gray-level co-occurrence matrix, while different topographic features are derived from the bathymetry data. Given the high dimensionality and redundancy of these features, the random forest algorithm is employed to quantify feature importance, and a feature selection method is proposed based on importance ranking. Experimental results demonstrate that the selected feature combination achieves higher classification accuracy and stronger feature expression ability. By integrating the sample enhancement and multifeature fusion algorithms, the random forest model attains a classification accuracy of 96.73% on real data. This work establishes a viable framework for seabed sediment classification by leveraging the sample enhancement and multifeature fusion.
Keywords:
Sediments
Feature extraction
Classification algorithms
Bathymetry
Backscatter
Surfaces
Acoustics
Gray-scale
Accuracy
Kernel
Feature selection
machine learning
seabed sediment classification
texture features
topography features
Journal
IF:
5.3
Papers:
2.6K
Citations:
7.4K
View more organizations