OIL SPILL IDENTIFICATION IN SAR IMAGE USING CURVELET TRANSFORM AND SVM

Abstract

At present, the problem of marine pollution caused by oil spill accidents is increasingly serious. Rapid and accurate automatic recognition of SAR images provides an important prerequisite for the handling and decision of oil spill accidents. This paper proposes a feature extraction method for SAR images based on Curvelet transform. First, we performed discrete Curvelet transform and selected the low-frequency component as a new image matrix, which contains the main information. Then, the Principal Component Analysis (PCA) technique was applied to select the best features to reduce the dimension. Finally, the Support Vector Machine (SVM) classifier was used to distinguish between “oil slicks" and “look-alikes oil slicks” and verify the validity of the extracted features. Experiments were performed on the different datasets, and the results proved that the accuracy of recognition is improved with Curvelet transform. In addition, compared with other neural network algorithms, Curvelet transform is an effective way to extract a reduced set of discriminative features for SAR images.

Existing System

Principal component analysis (Principal Component Analysis, PCA), Linear Discriminant Analysis (LDA)

Proposed System

This paper proposes a SAR image recognition method using Curvelet transform. Firstly, the SAR image is decomposed by Curvelet transform, and retained the main information of the SAR image. Then the PCA dimensionality reduction algorithm is used to extract the image features, and the SVM classifier is used for verification and identification.

Architecture

BLOCK DIAGRAM