Bathymetry of the Coral Reefs of Weizhou Island Based on Multispectral Satellite Images

Rongyong Huang,Lin Mu,Wenhuan Wang,Jikun Wang,Kefu Yu,Yinghui Wang

doi:10.3390/rs9070750

Abstract

Shallow water depth measurements using multispectral images are crucial for marine surveying and mapping. At present, relevant studies either depend on the use of other auxiliary data (such as field water depths or water column data) or contain too many unknown variables, thus making these studies suitable only for images that contain enough visible wavebands. To solve this problem, a Quasi-Analytical Algorithm (QAA) approach is proposed in this paper for estimating the water depths around Weizhou Island by developing a QAA to estimate the diffuse attenuation coefficients and simplifying the parameterization of the bathymetric model. The approach contains an initialization sub-approach and a novel global adjustment sub-approach. It is not only independent of other auxiliary data but also greatly reduces the number of unknowns. Experimental results finally demonstrated that the Root Mean Square Errors (RMSEs) were 1.01 m and 0.77 m for the ZY-3 image and the WorldView-3 (WV-3) image, respectively, so the approach is competitive to other QAA bathymetric methods. Besides, the global adjustment sub-approach was also seen to be superior to common smoothing filters: if the Signal-to-Noise Ratio (SNR) is as low as 42, i.e., ZY-3, it can smooth the water depths and improve the accuracies, otherwise can avoid the over-smoothing of water depths.

Highlights

Shallow seawater depth measurements are considered to be very important, as they are directly relevant to the applications of environmental management, exploration, and research [1,2,3,4].For example, the natural development of coral reefs can strongly influence the physical structure of their environment
Thereafter, we further attempt to propose a global adjustment method to refine the base of the estimated water depths depending on the following two conditions: (a) most of the water depths observed between a pixel and its adjacent pixels should be similar to each other; and (b) if the resolution of the image is high, most of the water depths of the pixels that are close to the transient coastline should be approximately zero
When the statistics are further estimated, we find that the maximum value of the absolute errors (MAX) is 4.69 m, the mean of the absolute errors (MAE) is 1.02 m, and the root mean square error (RMSE) is 1.24 m, where the errors are calculated by using the differences between the water depths estimated from from the the ZY-3

Summary

Introduction

Shallow seawater depth measurements are considered to be very important, as they are directly relevant to the applications of environmental management, exploration, and research [1,2,3,4].For example, the natural development of coral reefs can strongly influence the physical structure of their environment. Shallow seawater depth measurements are considered to be very important, as they are directly relevant to the applications of environmental management, exploration, and research [1,2,3,4]. Information associated with water depth is fundamental for characterizing and discriminating between different kinds of coral reef habitats, such as patch reefs, spur-and-groove systems around reef fronts, and sea grass beds [1]. The knowledge of the water depth can facilitate estimations of the bottom albedo, which can improve the habitat mapping [5]. Shallow water depth measurements have consistently represented an important component of marine surveying and mapping. As the echosounders are usually fixed in ships [7,10], such echosounder

Methods

Results

Discussion

Conclusion