High diversity seabed habitats, such as shellfish aggregations, play a significant role in marine ecosystem sustainability but are susceptible to bottom disturbance induced by anthropogenic activities. Regular monitoring of these habitats with effective mapping methods is therefore essential. Multibeam echosounder (MBES) has been widely used in recent decades for seabed characterization due to its non-destructive manner and extensive spatial coverage compared to traditional methods like bottom sampling. Nevertheless, bottom sampling remains essential to link ground truth with acoustic seabed classification. Using seabed samples and MBES measurements, machine learning techniques are commonly employed to model their relationships and generate classification maps of an extended seabed. However, limited ground truth data, resulting from constraints in regulations, budget, or time, may impede the development of robust machine learning models. To address this challenge, we applied a semi-supervised machine learning method to classify seabed sediments of a blue mussel (Mytilus edulis) cultivation area in the Oosterschelde, the Netherlands. We utilized nine boxcore samples to generate pseudo-labels on MBES data. These pseudo-labels enlarged the training data size, facilitated the training of three comprehensive machine learning algorithms (Gradient Boosting, Random Forest, and Support Vector Machine), and helped to classify the study site into mussel and non-mussel areas. We found the geomorphological and backscatter-related features to be complementary for mussel culture detection. Our classification results were demonstrated effective through expert knowledge of this cultivation area and brought insights for future research on natural mussel habitats.

@en

To facilitate the conservation of seafloor habitats and planning of offshore activities, there is a growing need for mapping marine benthos in an effective and efficient way. Acoustic data acquired by multi-beam echosounders (MBES) have been extensively used for large-scale and high-resolution seafloor characterization. A deeper understanding of the relationship between backscatter data and sediment compositions can help to identify the benthos occurrence from the MBES data. With two multi-spectral MBES datasets collected near the western Wadden Sea islands in the North Sea, we investigated the potential of mapping marine benthos through backscatter classification. Two unsupervised classification methods, i.e., Bayesian classification, which mainly exploits the backscatter strength from incident angles larger than 20°, and hierarchical clustering of the backscatter strength at different angular ranges, were employed and the results were compared. The classification results from both methods showed a good correspondence with sediment properties such as the median grain size. Moreover, based on a principal component analysis of bottom sample properties, the hierarchical clustering results indicated a better distinction between contributions from the gravel content and benthos occurrence, e.g., sand mason worm density, than Bayesian classification, through involving the backscatter angular variations. Classification for multiple frequencies, on the other hand, showed little difference regarding the relationship with bottom sample properties. Although the backscatter difference between frequencies was also found to positively correlate with certain sample properties, using multi-spectral features for acoustic classification in this study did not reveal additional information compared to single-frequency classification results.@en

Detailed knowledge of both the sedimentological and ecological characteristics of the seafloor is essential when undertaking bottom-disturbing activities, but can be a challenge to obtain. Through backscatter data at different frequencies, collected with a multi-spectral multi-beam sonar, information on the structure of both the sediment surface and subsurface, and potentially also on the presence and distribution of benthic organisms, can be derived. We conducted two surveys at sea in summer 2021, in which we used an R2Sonic 2026 multi-spectral multi-beam sonar in the southern North Sea. Boxcore samples were taken to gather information on macrobenthos densities and sediment characteristics. The two studied areas were found to differ in seafloor morphology and correspondingly in the composition of the sediment composition and benthos distribution. Backscatter strength was used to classify the seafloor via the Bayesian method and via hierarchical clustering of angular variation. Relationships between the classification results for three frequencies and sediment and ecological variables were studied through redundancy analysis (RDA), for which hierarchical clustering of the angular variation in backscatter strength showed a higher model fit than Bayesian classification. We found that the density of the sand mason worm Lanice conchilega and percentages of dead shells, gravel and sand contributed most to the backscatter-based classification, with lower contributions of the percentages of mud and living bivalves. Our results suggest that acoustic backscatter can be used to delineate distinct seafloor regions, corresponding with concurrent gradients in ecological and sedimentological variables.

@en

Acoustic classification using single-beam and multi-beam echosounders has been widely applied in characterizing seabed sediments. Although previous studies have shown a better discrimination of fine and coarse sediments using multi-spectral echosounder data, analysis regarding comprehensive seabed sediment properties is still needed. In this study, we used single-beam data of 24 kHz, as well as multi-beam data of 90 and 300 kHz to investigate the benefits of multi-spectral backscatter data in describing sediment properties including median grain size; weight percentages of gravel, sand, and mud; volume percentages of stones, shell fragments, and living bivalves; as well as density of acoustically hard animals (molluscs and the tube-building worm Lanice conchilega). We classified data of each frequency in an unsupervised manner, using K-means clustering for the single-beam echo time series and Bayesian classification for the multi-beam backscatter. Compared with the top-layer sediment properties, we found classification of 90 and 300 kHz consistent with variations of median grain size and L. conchilega density, whereas classification of 24 kHz can also be related to the percentages of shell fragments and stones. In addition, one acoustic class of 24 kHz might indicate a higher gravel content in the subsurface of the study area. Although quantitative relationships between backscatter and sediment properties are still difficult to achieve given a limited number of samples, using multi-spectral backscatter data is a potential approach to characterize seabed sediments from various perspectives.

@en

Classification of aerial point clouds with high accuracy is significant for many geographical applications, but not trivial as the data are massive and unstructured. In recent years, deep learning for 3D point cloud classification has been actively developed and applied, but notably for indoor scenes. In this study, we implement the point-wise deep learning method Dynamic Graph Convolutional Neural Network (DGCNN) and extend its classification application from indoor scenes to airborne point clouds. This study proposes an approach to provide cheap training samples for point-wise deep learning using an existing 2D base map. Furthermore, essential features and spatial contexts to effectively classify airborne point clouds colored by an orthophoto are also investigated, in particularly to deal with class imbalance and relief displacement in urban areas. Two airborne point cloud datasets of different areas are used: Area-1 (city of Surabaya—Indonesia) and Area-2 (cities of Utrecht and Delft—the Netherlands). Area-1 is used to investigate different input feature combinations and loss functions. The point-wise classification for four classes achieves a remarkable result with 91.8% overall accuracy when using the full combination of spectral color and LiDAR features. For Area-2, different block size settings (30, 50, and 70 m) are investigated. It is found that using an appropriate block size of, in this case, 50 m helps to improve the classification until 93% overall accuracy but does not necessarily ensure better classification results for each class. Based on the experiments on both areas, we conclude that using DGCNN with proper settings is able to provide results close to production.@en

Functional classification of the road is important to the construction of sustainable transport systems and proper design of facilities. Mobile laser scanning (MLS) point clouds provide accurate and dense 3D measurements of road scenes, while their massive data volume and lack of structure also bring difficulties in processing. 3D point cloud understanding through deep neural networks achieves breakthroughs since PointNet and arouses wide attention in recent years. In this paper, we study the automatic road type classification of MLS point clouds by employing a point-wise neural network, RandLA-Net, which is designed for consuming large-scale point clouds. An effective local feature aggregation (LFA) module in RandLA-Net preserves the local geometry in point clouds by formulating an enhanced geometric feature vector and learning different point weights in a local neighborhood. Based on this method, we also investigate possible feature combinations to calculate neighboring weights. We train on a colorized point cloud from the city of Hannover, Germany, and classify road points into 7 classes that reveal detailed functions, i.e., sidewalk, cycling path, rail track, parking area, motorway, green area, and island without traffic. Also, three feature combinations inside the LFA module are examined, including the geometric feature vector only, the geometric feature vector combined with additional features (e.g., color), and the geometric feature vector combined with local differences of additional features. We achieve the best overall accuracy (86.23%) and mean IoU (69.41%) by adopting the second and third combinations respectively, with additional features including Red, Green, Blue, and intensity. The evaluation results demonstrate the effectiveness of our method, but we also observe that different road types benefit the most from different feature settings.

@en