Three-dimensional of Coastal Front Reconstruction Using Radarsat-1 Sar Satellite Data

Natural phenomena that are imaged using remote sensing satellite data can be reconstructed in 3-D. This process can be accomplished either by active or passive methods. The active methods interfere with the reconstructed phenomena, either mechanically or radiometrically. The radiometric methods reconstruct the 3-D from the reflected or backscattered information about the specific objects or phenomena. However, passive methods use a sensor to measure the radiance reflected or emitted by the object's surface to infer its 3-D structure. 3-D reconstruction of natural phenomena plays tremendous role to understand a complex system such as the dynamic processes of coastal waters. Three-dimensional (3D) computer visualization has tremendous demands for complex phenomena studies. Coastal waters are considered as complex system because of they are dominated by complex system. In this regard, this study aims to present a method that is based on fuzzy B-spline to reconstruct 3D of coastal water phenomena such as front from two-dimensional RADARSAT-1 SAR data. In doing so, fuzzy B-spline algorithm is integrated with Volterra model and velocity bunching model. Volterra algorithm is used to determine the sea surface current along the front zone while velocity bunching model implemented to acquire the information about significant wave height. fuzzy B-spline reconstructed 3-D front with smooth graphic feature. Indeed, fuzzy B-spline tracked the smooth and rough surface. Finally, fuzzy B-spline algorithm can keep track of uncertainty with representing spatially clustered gradient of flow points across the front. In conclusion, the fuzzy B-spline algorithm can be used for 3-D front reconstruction with integration of velocity bunching and Volterra algorithm. INTRODUCTION Front plays tremendous role to understand the mechanisms of coastal water circulation, marine productivities and coastal pollutant martials spreading (Simpson and Pingree 1978; Bowden 1983;Robinson 1995, Marghany 2011). Although, conventional methods for front studies depend on in-situ measurements of sea water temperature and salinity they might be costly and time consuming. Isothermal, isohaline contours and water mass diagrams are established procedures for front detection nevertheless front cannot be visualized in large scale surface ocean (Simpson 1981; Bowden 1983; Marghany 2012). In this paper, we address how 3-D front can be reconstructed from single SAR data (namely the RADARSAT-1 SAR) using integration of Volterra kernel (Ingland and Garello, 1990), velocity bunching and Fuzzy B-spline models (Marghany et al., 2010 and Marghany and Mazlan 2010). In this regard, synthetic aperture radar (SAR) is able to identify front as a result of abruptly changes of surface wave pattern across front led to exceedingly change cross backscatter of SAR data. Therefore, Johannessen et al., (1996) stated that SAR images can sometimes be used to interpret frontal dynamics, including growth and decay of meanders. Recently, Jiang et al., (2009) exploited various remote sensing data. Satellite images obtained from the Advanced Very High Resolution Radiometer (AVHRR), the Moderate Resolution Imaging Spectroradiometer (MODIS), the Sea-viewing Wide Field-of-view Sensor (SeaWiFS) and RADARSAT-1 SAR S1 mode data to study coastal water plume and front which is also captured in S1 mode data (Klemas 2011). There are about three hypothesis that examined are: (i) the use of Volterra model to detect front flow pattern in RADARSAT-1 SAR CHH band; (ii) the use of velocity bunching model to acquire significant wave height from RADARSAT-1 SAR data; and (iii) to utilize fuzzy B-spline to remodel 3-D of front surface. In fact, scientists have used conventional mathematical algorithms to comprehend the complexity of various system interaction. In this regard, imaging coastal feature in synthetic aperture radar (SAR) required such standard mathematical algorithms have been reported recently by (Zaki, 2007; Messaoudi et al., 2007; Stephen 2009; Adeyemo and Fred 2009; Mehmet 2009; Ugwu, 2009; Akintorinwa and Adesoji 2009; Boumaza et al., 2009; Anjamrooz 2011; Anjamrooz et al., 2011; Khadijeh et al., 2011; Guillermo et al., 2011; Murat 2011; Mustafa 2011). METHODS AND EQUATION 3D Model for front reconstruction There are three algorithms involved for 3-D front reconstruction; Volterra, velocity bunching and Fuzzy B-spline algorithms. Significant wave heights are simulated from RADARSAT-1 SAR image by using velocity bunching model. Fuzzy B-spline used significant wave height information to reconstruct 3-D front. Moreover, front flow pattern is modeled by Volterra model. The RADARSAT-1 SAR fine mode data were acquired on March 26, 2004, over the coastline of Kuala Terengganu, Malaysia (103° 5' E to 103° 9'E and 5° 20' N to 5° 27' N). The RADARSAT-1 SAR fine mode data are acquiring information using C band HH polarized of frequency 5.3 GHz. The swath width of RADARSAT-1 SAR fine mode sensor is 50 km, with the range resolution of 8-9 km. There are two numbers of looks for The RADARSAT-1 SAR and the incident angle of 35°-49° (RADARSAT 2012). Volterra Model In refereeing to Ingland and Garello, (1990), Volterra series can be used to model nonlinear imaging mechanisms of surface current gradients by RADARSAT-1 SAR image. As result of that Volterra linear kernel is contained most of RADARSAT-1 SAR energy which used to simulate current flow along range direction. Following Ingland and Garello, (1990) Volterra kernel filter has the following expression: