Comparison of Classification Techniques on Fused Optical and SAR Images for Shoreline Extraction: A Case Study at Northeast Coast of Peninsular Malaysia

Abstract

Shoreline is a very important element to identify exact boundary at the coastal areas of a country. However, in order to identify land-water boundary for a large region using traditional ground survey technique is very time consuming. Alternatively, shoreline can be extracted by using satellite images that minimizes the mapping errors. The trend of extracting shoreline has been shifted from image processing to machine learning and data mining techniques. By using machine learning technique, the satellite images could be classified into land and water classes in order to extract shoreline. However, the result is meaningless if it has cloud and shadow on the water-land boundary. In this study, we compare the accuracy and Kappa Coefficient of six machine learning techniques namely Maximum Likelihood, Minimum Distance, Mahalanobis Distance, Parallelepiped, Neural Network and Support Vector Machines on three type of images; single optical multispectral, single SAR and fused image. A case study for this research is done alongside Tumpat beach, located at the Northeast Coast of Peninsular Malaysia. All the machine learning techniques have been tested on the three types of images. The experimental results show that classification using SVM on single multispectral image has the highest accuracy among all. However, the classified of fused image using SVM is considered much more accurate because it can cater the cloud and shadow problem. Additionally, the classification on 5 and 10 m fused images also tested and the result shows that with the increase of spatial resolution of fused image, the classification accuracy also increases.

A Fusion Approach for Water Area Classification Using Visible, Near Infrared and Synthetic Aperture Radar for South Asian Conditions

“…In an effort to develop a more robust water extraction technique that is tailored to overcome challenges in humid tropical climates of South Asia, we aim to fuse complementary strengths of remote sensing data types. Various studies have targeted fusion of multiple sensor products for various goals, such as shoreline extraction [4], change detection [5], retrieving daily normalized difference vegetation index (NDVI) and leaf area index (LAI) [6], and temporal aggregation for land cover mapping [7]. Studies by Kaplan and Avdan [1], Huang et al [8], and Irwin et al [12] have monitored wetlands and surface water by different fusion techniques.…”

A Learning Framework With Multispectral Band-Differentiated Encoding for Remote Sensing Water Body Detection

“…[13] employed K-means clustering based on the pixel coordinates of RGB three-band values as three-dimensional coordinates. Additionally, [14] applied Bayesian principles, and [15], [16], [17] designed support vector machine (SVM) solutions, all focusing on binary classification of water bodies against other land features. Solbo et al [18] employed a primitive, fully connected artificial neural network approach for surface water mapping in SAR images.…”

Multi-Temporal Sentinel-2 Images for Classification Accuracy

“…For the highresolution image, an optical image from Sentinel-2 is used along with the SAR image from Sentinel 1A. Table 1: Detail of information spectral bands of sentinel-2 images Sentinel-2A (Optical Data) Sentinel-2 was introduced by the ESA to perform global spatial resolution monitoring as part of the EUs Copernicus program (Manaf et al, 2016;Hagolle et al, 2015;Segl et al, 2015). Sentinel-2 images cover 13 wavelengths in the visible, Near Infrared (NIR) and Shortwave Infrared (SWIR).…”