A comparison of Landsat 8, RapidEye and Pleiades products for improving empirical predictions of satellite-derived bathymetry

A Convolutional Neural Network with Spatial Location Integration for Nearshore Water Depth Inversion

Nearshore water depth plays a crucial role in scientific research, navigation management, coastal zone protection, and coastal disaster mitigation. This study aims to address the challenge of insufficient feature extraction from remote sensing data in nearshore water depth inversion. To achieve this, a convolutional neural network with spatial location integration (CNN-SLI) is proposed. The CNN-SLI is designed to extract deep features from remote sensing data by considering the spatial dimension. In this approach, the spatial location information of pixels is utilized as two additional channels, which are concatenated with the input feature image. The resulting concatenated image data are then used as the input for the convolutional neural network. Using GF-6 remote sensing images and measured water depth data from electronic nautical charts, a nearshore water depth inversion experiment was conducted in the waters near Nanshan Port. The results of the proposed method were compared with those of the Lyzenga, MLP, and CNN models. The CNN-SLI model demonstrated outstanding performance in water depth inversion, with impressive metrics: an RMSE of 1.34 m, MAE of 0.94 m, and R2 of 0.97. It outperformed all other models in terms of overall inversion accuracy and regression fit. Regardless of the water depth intervals, CNN-SLI consistently achieved the lowest RMSE and MAE values, indicating excellent performance in both shallow and deep waters. Comparative analysis with Kriging confirmed that the CNN-SLI model best matched the interpolated water depth, further establishing its superiority over the Lyzenga, MLP, and CNN models. Notably, in this study area, the CNN-SLI model exhibited significant performance advantages when trained with at least 250 samples, resulting in optimal inversion results. Accuracy evaluation on an independent dataset shows that the CNN-SLI model has better generalization ability than the Lyzenga, MLP, and CNN models under different conditions. These results demonstrate the superiority of CNN-SLI for nearshore water depth inversion and highlight the importance of integrating spatial location information into convolutional neural networks for improved performance.

Investigating the Shallow-Water Bathymetric Capability of Zhuhai-1 Spaceborne Hyperspectral Images Based on ICESat-2 Data and Empirical Approaches: A Case Study in the South China Sea

Accurate bathymetric and topographical information is crucial for coastal and marine applications. In the past decades, owing to its low cost and high efficiency, satellite-derived bathymetry has been widely used to estimate the depth of shallow water in coastal areas. However, insufficient spectral bands and availability of in situ water depths limit the application of satellite-derived bathymetry. Currently, the investigation about the bathymetric potential of hyperspectral imaging is relatively insufficient based on datasets of the Ice, Cloud, and Land Elevation Satellite-2 (ICESat-2). In this study, Zhuhai-1 hyperspectral images and ICESat-2 datasets were utilized to perform nearshore bathymetry and explore the bathymetric capability by selecting different bands based on classical empirical models (the band ratio model and the linear band model). Furthermore, experimental results achieved at the South China Sea indicate that the combination of blue (2 and 3 band) and green (9 band) bands and the combination of red (10 and 12 band) and near-infrared (29 band) bands are most suitable to achieve nearshore bathymetry. Correspondingly, the highest accuracy of bathymetry reached root mean square error values of 0.98 m and 1.19 m for different band combinations evaluated through bathymetric results of reference water depth. The bathymetric accuracy of Zhuhai-1 image is similar with that of Sentinel-2 when employing the blue and green bands. The combination of red and near-infrared bands has a higher bathymetric accuracy for Zhuhai-1 image than that for Sentinel-2 image.

Satellite-derived bathymetry based on machine learning models and an updated quasi-analytical algorithm approach

Retrieving the water depth by satellite is a rapid and effective method for obtaining underwater terrain. In the optical shallow waters, the bottom signal has a great impact on the radiation from the water which related to water depth. In the optical shallow waters, the spatial distribution characteristic of water quality parameters derived by the updated quasi analysis algorithm (UQAA) is highly correlated with the bottom brightness. Because the bottom reflection signal is strongly correlated with the spatial distribution of water depth, the derived water quality parameters may helpful and applicable for optical remote sensing based satellite derived bathymetry. Therefore, the influence on bathymetry retrieval of the UQAA IOPs is worth discussing. In this article, different machine learning algorithms using a UQAA were tested and remote sensing reflectance at water depth in situ points and their detection accuracy were evaluated by using Worldwiew-2 multispectral remote sensing images and laser measurement data. A backpropagation (BP) neural network, extreme value learning machine (ELM), random forest (RF), Adaboost, and support vector regression (SVR) machine models were utilized to compute the water depth retrieval of Ganquan Island in the South China Sea. According to the obtained results, bathymetry using the UQAA and remote sensing reflectance is better than that computed using only remote sensing reflectance, in which the overall improvements in the root mean square error (RMSE) were 1 cm to 5 cm and the overall improvement in the mean relative error (MRE) was 1% to 5%. The results showed that the results of the UQAA could be used as a main water depth estimation eigenvalue to increase water depth estimation accuracy.

Nearshore Bathymetry Retrieval from Wave-Based Inversion for Video Imagery

A wavelet-based method for bathymetry retrieval using a sequence of static images of the surface wave field, as obtained from video imagery, is proposed. Synthetic images of the water surface are generated from a numerical Boussinesq type model simulating the propagation of irregular waves. The spectral analysis is used to retrieve both wave periods and wavelengths by evaluating the spectral peaks in the time and spatial domains, respectively. The water depths are estimated using the linear dispersion relation and the results are validated with the model’s bathymetry. To verify the proposed methodology, 2D and 3D simulations considering effects of wave shoaling and refraction were performed for different sea conditions over different seafloors. The method’s ability to reproduce the original bathymetry is shown to be robust in intermediate and shallow waters, being also validated with a real case with images obtained with a shore-based video station. The main improvements of the new method compared to the consideration of a single image, as often used in Satellite Derived Bathymetry, is that the use of successive images enables the consideration of different wave periods, improving depth estimations and not requiring the use of subdomains or filters. This image processing methodology shows very positive results to provide bathymetry maps for shallow marine environments and can be useful to monitor the nearshore with high time- and space-resolution at low cost.

Evaluation of Satellite-Derived Bathymetry from High and Medium-Resolution Sensors Using Empirical Methods

This study evaluates the accuracy of bathymetric maps generated from multispectral satellite datasets acquired from different multispectral sensors, namely the Worldview 2, PlanetScope, and the Sentinel 2, in the bay of Elounda in Crete. Image pre-processing steps were implemented before the use of the three empirical methods for estimating bathymetry. A dedicated correction and median filter have been applied to minimize noise from the sun glint and the sea waves. Due to the spectral complexity of the selected study area, statistical correlation with different numbers of bands was applied. The analysis indicated that blue and green bands obtained the best results with higher accuracy. Then, three empirical models, namely the Single Band Linear Algorithm, the Multiband Linear Algorithm, and the Ratio Transform Algorithm, were applied to the three multispectral images. Bathymetric and error distribution maps were created and used for the error assessment of results. The accuracy of the bathymetric maps estimated from different empirical models is compared with on-site Single beam Echo Sounder measurements. The most accurate bathymetric maps were obtained using the WorldView 2 and the empirical model of the Ratio Transform algorithm, with the RMSE reaching 1.01 m.

Satellite-derived bathymetry using Landsat-8 and Sentinel-2A images: assessment of atmospheric correction algorithms and depth derivation models in shallow waters

Satellite-derived bathymetry (SDB) has an extensive prospect in nearshore bathymetry for its high efficiency and low costs. Atmospheric correction and bathymetric modeling are critical processes in SDB, and examining the performance of related algorithms and models will contribute to the formulation of reliable bathymetry strategies. This study explored the effectiveness of three general atmospheric correction algorithms, namely Second Simulation of a Satellite Signal in the Solar Spectrum (6S), Atmospheric correction for OLI 'lite' (ACOLITE), and QUick Atmospheric Correction (QUAC), in depth retrieval from Landsat-8 and Sentinel-2A images using different SDB models over Ganquan Island and Oahu Island. The bathymetric Light Detection and Ranging (LiDAR) data was used for SDB model training and accuracy verification. The results indicated that the three atmospheric correction algorithms could provide effective corrections for SDB. For the SDB models except log-transformed band ratio model (LBR) and support vector machine (SVM), the impact of different atmospheric corrections on bathymetry was basically the same. Furthermore, we assessed the performance of six different SDB models: Lyzenga's model (LM), generalized additive model (GAM), LBR, SVM, multilayer perceptron (MLP), and random forest (RF). The bathymetric accuracy, consistency of bathymetric maps and generalization ability were considered for the assessment. Given sufficient training data, the accuracy of the machine learning models (SVM, MLP, RF) was generally superior to that of the empirical inversion models (LM, GAM, LBR), with the root mean square error (RMSE) varied between 0.735 m to 1.177 m. MLP achieved the best accuracy and consistency. When the depth was deeper than 15 m, the bathymetry error of all the SDB models increased sharply, and LM, LBR and SVM reached the upper limit of depth retrieval capability at 20-25 m. In addition, LM and LBR were demonstrated to have better adaptability in heterogeneous environment without training data.

Improved Filtering of ICESat-2 Lidar Data for Nearshore Bathymetry Estimation Using Sentinel-2 Imagery

The accurate estimation of nearshore bathymetry is necessary for multiple aspects of coastal research and practices. The traditional shipborne single-beam/multi-beam echo sounders and Airborne Lidar bathymetry (ALB) have a high cost, are inefficient, and have sparse coverage. The Satellite-derived bathymetry (SDB) method has been proven to be a promising tool in obtaining bathymetric data in shallow water. However, current empirical SDB methods for multispectral imagery data usually rely on in situ depths as control points, severely limiting their spatial application. This study proposed a satellite-derived bathymetry method without requiring a priori in situ data by merging active and passive remote sensing (SDB-AP). It realizes rapid bathymetric mapping with only satellite remotely sensed data, which greatly extends the spatial coverage and temporal scale. First, seafloor photons were detected from the ICESat-2 raw photons based on an improved adaptive Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm, which could calculate the optimal detection parameters for seafloor photons by adaptive iteration. Then, the bathymetry of the detected seafloor photons was corrected because of the refraction that occurs at the air–water interface. Afterward, the outlier photons were removed by an outlier-removal algorithm to improve the retrieval accuracy. Subsequently, the high spatial resolution (0.7 m) ICESat-2 derived bathymetry data were gridded to match the Sentinel-2 data with a lower spatial resolution (10 m). All of the ICESate-2 gridded data were randomly separated into two parts: 80% were employed to train the empirical bathymetric model, and the remaining 20% were used to quantify the inversion accuracy. Finally, after merging the ICESat-2 data and Sentinel-2 multispectral images, the bathymetric maps over St. Thomas of the United States Virgin Islands, Acklins Island in the Bahamas, and Huaguang Reef in the South China Sea were produced. The ICESat-2-derived results were compared against in situ data over the St. Thomas area. The results showed that the estimated bathymetry reached excellent inversion accuracy and the corresponding RMSE was 0.68 m. In addition, the RMSEs between the SDB-AP estimated depths and the ICESat-2 bathymetry results of St. Thomas, Acklins Island, and Huaguang Reef were 0.96 m, 0.91 m, and 0.94 m, respectively. Overall, the above results indicate that the SDB-AP method is effective and feasible for different shallow water regions. It has great potential for large-scale and long-term nearshore bathymetry in the future.

Identification of Potential Surface Water Resources for Inland Aquaculture from Sentinel-2 Images of the Rwenzori Region of Uganda

Aquaculture has the potential to sustainably meet the growing demand for animal protein. The availability of water is essential for aquaculture development, but there is no knowledge about the potential inland water resources of the Rwenzori region of Uganda. Though remote sensing is popularly utilized during studies involving various aspects of surface water, it has never been employed in mapping inland water bodies of Uganda. In this study, we assessed the efficiency of seven remote-sensing derived water index methods to map the available surface water resources in the Rwenzori region using moderate resolution Sentinel 2A/B imagery. From the four targeted sites, the Automated Water Extraction Index for urban areas (AWEInsh) and shadow removal (AWEIsh) were the best at identifying inland water bodies in the region. Both AWEIsh and AWEInsh consistently had the highest overall accuracy (OA) and kappa (OA > 90%, kappa > 0.8 in sites 1 and 2; OA > 84.9%, kappa > 0.61 in sites 3 and 4), as well as the lowest omission errors in all sites. AWEI was able to suppress classification noise from shadows and other non-water dark surfaces. However, none of the seven water indices used during this study was able to efficiently extract narrow water bodies such as streams. This was due to a combination of factors like the presence of terrain shadows, a dense vegetation cover, and the image resolution. Nonetheless, AWEI can efficiently identify other surface water resources such as crater lakes and rivers/streams that are potentially suitable for aquaculture from moderate resolution Sentinel 2A/B imagery.

Assessment of Empirical Algorithms for Shallow Water Bathymetry Using Multi-Spectral Imagery of Pearl River Delta Coast, China

Pearl River Delta (PRD), as one of the most densely populated regions in the world, is facing both natural changes (e.g., sea level rise) and human-induced changes (e.g., dredging for navigation and land reclamation). Bathymetric information is thus important for the protection and management of the estuarine environment, but little effort has been made to comprehensively evaluate the performance of different methods and datasets. In this study, two linear regression models—the linear band model and the log-transformed band ratio model, and two non-linear regression models—the support vector regression model and the random forest regression model—were applied to Landsat 8 (L8) and Sentinel-2 (S2) imagery for bathymetry mapping in 2019 and 2020. Results suggested that a priori area clustering based on spectral features using the K-means algorithm improved estimation accuracy. The random forest regression model performed best, and the three-band combinations outperformed two-band combinations in all models. When the non-linear models were applied with three-band combination (red, green, blue) to L8 and S2 imagery, the Root Mean Square Error (Mean Absolute Error) decreased by 23.10% (35.53%), and the coefficient of determination (Kling-Gupta efficiency) increased by 0.08 (0.09) on average, compared to those using the linear regression models. Despite the differences in spatial resolution and band wavelength, L8 and S2 performed similarly in bathymetry estimation. This study quantified the relative performance of different models and may shed light on the potential combination of multiple data sources for more timely and accurate bathymetry mapping.

Influence of Different Satellite Imagery on the Analysis of Riparian Leaf Density in a Mountain Stream

In recent decades, technological advancements in sensors have generated increasing interest in remote sensing data for the study of vegetation features. Image pixel resolution can affect data analysis and results. This study evaluated the potential of three satellite images of differing resolution (Landsat 8, 30 m; Sentinel-2, 10 m; and Pleiades 1A, 2 m) in assessing the Leaf Area Index (LAI) of riparian vegetation in two Mediterranean streams, and in both a winter wheat field and a deciduous forest used to compare the accuracy of the results. In this study, three different retrieval methods—the Caraux-Garson, the Lambert-Beer, and the Campbell and Norman equations—are used to estimate LAI from the Normalized Difference Vegetation Index (NDVI). To validate sensor data, LAI values were measured in the field using the LAI 2200 Plant Canopy Analyzer. The statistical indices showed a better performance for Pleiades 1A and Landsat 8 images, the former particularly in sites characterized by high canopy closure, such as deciduous forests, or in areas with stable riparian vegetation, the latter where stable reaches of riparian vegetation cover are almost absent or very homogenous, as in winter wheat fields. Sentinel-2 images provided more accurate results in terms of the range of LAI values. Considering the different types of satellite imagery, the Lambert-Beer equation generally performed best in estimating LAI from the NDVI, especially in areas that are geomorphologically stable or have a denser vegetation cover, such as deciduous forests.

A Cross-Resolution, Spatiotemporal Geostatistical Fusion Model for Combining Satellite Image Time-Series of Different Spatial and Temporal Resolutions

Dense time-series with coarse spatial resolution (DTCS) and sparse time-series with fine spatial resolution (STFS) data often provide complementary information. To make full use of this complementarity, this paper presents a novel spatiotemporal fusion model, the spatial time-series geostatistical deconvolution/fusion model (STGDFM), to generate synthesized dense time-series with fine spatial resolution (DTFS) data. Attributes from the DTCS and STFS data are decomposed into trend and residual components, and the spatiotemporal distributions of these components are predicted through novel schemes. The novelty of STGDFM lies in its ability to (1) consider temporal trend information using land-cover-specific temporal profiles from an entire DTCS dataset, (2) reflect local details of the STFS data using resolution matrix representation, and (3) use residual correction to account for temporary variations or abrupt changes that cannot be modeled from the trend components. The potential of STGDFM is evaluated by conducting extensive experiments that focus on different environments; spatially degraded datasets and real Moderate Resolution Imaging Spectroradiometer (MODIS) and Landsat images are employed. The prediction performance of STGDFM is compared with those of a spatial and temporal adaptive reflectance fusion model (STARFM) and an enhanced STARFM (ESTARFM). Experimental results indicate that STGDFM delivers the best prediction performance with respect to prediction errors and preservation of spatial structures as it captures temporal change information on the prediction date. The superiority of STGDFM is significant when the difference between pair dates and prediction dates increases. These results indicate that STGDFM can be effectively applied to predict DTFS data that are essential for various environmental monitoring tasks.

Atmospheric correction for satellite-derived bathymetry in the Caribbean waters: from a single image to multi-temporal approaches using Sentinel-2A/B

Different atmospheric correction (AC) procedures for Sentinel-2 satellites are evaluated for their effectiveness in retrieving consistent satellite-derived bathymetry (SDB) over two islands in the Caribbean (Buck and Culebra). The log-ratio method for SDB, which allows use of minimal calibration information from lidar surveys (25 points in this study), is applied to several Sentinel-2A/B scenes at 10 m spatial resolution. The overall performance during a one-year study period depends on the image quality and AC. Three AC processors were evaluated: ACOLITE Exponential model (EXP), ACOLITE Dark Spectrum Fitting model (DSF), and C2RCC model. ACOLITE EXP and ACOLITE DSF produce greater consistency and repeatability with accurate results in a scene-by-scene analysis (mean errors ∼1.1 m) for depths up to 23 m (limit of lidar surveys). In contrast, C2RCC produces lower accuracy and noisier results with generally higher (>50%) errors (mean errors ∼2.2 m), but it is able to retrieve depth for scenes in Buck Island that have moderately severe sunglint. Furthermore, we demonstrate that a multi-temporal compositing model for SDB mapping, using ACOLITE for the input scenes, could achieve overall median errors <1 m for depths ranging 0-23 m. The simple and effective compositing model can considerably enhance coastal SDB estimates with high reliability and no missing data, outperforming the traditional single image approaches and thus eliminating the need to evaluate individual scenes. The consistency in the output from the AC correction indicates the potential for automated application of the multi-scene compositing technique, which can apply the open and free Sentinel-2 data set for the benefit of operational and scientific investigations.

Bathymetry Time Series Using High Spatial Resolution Satellite Images

The use of the new generation of remote sensors, such as echo sounders and Global Navigation Satellite System (GNSS) receivers with differential correction installed in a drone, allows the acquisition of high-precision data in areas of shallow water, as in the case of the channel of the Encañizadas in the Mar Menor lagoon. This high precision information is the first step to develop the methodology to monitor the bathymetry of the Mar Menor channels. The use of high spatial resolution satellite images is the solution for monitoring many hydrological changes and it is the basis of the three-dimensional (3D) numerical models used to study transport over time, environmental variability, and water ecosystem complexity.