Abstract-Determination of bathymetric information is key element for near off shore activities and hydrological studies such as coastal engineering applications, sedimentary processes and hydrographic surveying. Remotely sensed imagery has provided a wide coverage, low cost and time-effective solution for bathymetric measurements. In this paper a methodology is introduced using Ensemble Learning (EL) fitting algorithm of Least Squares Boosting (LSB) for bathymetric maps calculation in shallow lakes from high resolution satellite images and water depth measurement samples using Eco-sounder. This methodology considered the cleverest sequential ensemble that assigns higher weights as Boosting for those training sets that are difficult to fit. The LSB ensemble using reflectance of Green and Red bands and their logarithms from Spot-4 satellite image was compared with two conventional methods; the Principal Component Analysis (PCA) and Generalized Linear Model (GLM). The retrieved bathymetric information from the three methods was evaluated using Echo Sounder data. The LSB fitting ensemble resulted in RMSE of 0.15m where the PCA and GLM yielded RMSE of 0.19m and 0.18m respectively over shallow water depths less than 2m. The application of the proposed approach demonstrated better performance and accuracy compared with the conventional methods.Index Terms-Bathymetry, PCA, GLM, least square boosting.
I. INTRODUCTIONAccurate bathymetric information is so important for costal science applications, shipping navigations and environmental studies of marine areas [1]. Mapping underwater features as rocks, sandy areas, sediments accumulation and coral reefs needs up to date water depths information [2], [3]. Water depths data are essential also for accomplishing sustainable management [4], bathymetric information constitutes a key element hydrological modeling, flooding estimation and degrading or sediments removing [5], [6]. Manuscript received November 25, 2014; revised April 27, 2015. The work was supported by Mission Department, Egyptian Ministry of Higher Education (MoHE), Egypt-Japan University of Science and Technology (E-JUST) and partially supported by JSPS "Core-to-Core Program, B. Asia-Africa Science Platforms".H. Mohamed and Abdelazim Negm are with the Environmental Engineering Department, School of Energy and Environmental Engineering, Egypt-Japan University of Science and Technology, Alexandria, Egypt (e-mail: hassan.mohamed@ejust.edu.eg, negm@ejust.edu.eg).M. Zahran is with the Department of Geomatics Engineering, Faculty of Engineering at Shoubra, Benha University, Egypt (e-mail: mohamed.zahran01@feng.bu.edu.eg).O. Saavedra is with the Department of Civil Engineering, Tokyo Institute of Technology, Oookayama, Meguro, Tokyo, Japan. He is also with E-JUST, Egypt (e-mail: saavedra.o.aa@m.titech.ac.jp).Sonar remains the primary method for obtaining discreet water depth measurements with high accuracy [7]. Single beam sonar on survey vessel can acquire single point depths along sparsely surveying scan lines up to 50...