Oil Spills Identification in SAR Image Using Optimized RBF Network Model

The Synthetic Aperture Radar (SAR) can work on all weather conditions for high-resolution monitoring of oil spills. Efficient eigenvectors can be extracted to optimize the RBF network model which is used for distinguishing oil spill from SAR images. The eigenvectors can be computed using the Measured Dark zone boundary determined SAR images. Such eigenvectors could be valid input parameters for building the incentive functions that are used for training SAR image samples. Using the error between the output value and the actual value as a constraint, we can adjust the weighting factor, center and width of the radial basis, which consequently accelerate the convergence speed of the model. Finally, the oil spill is determined according to the linear output layer activation function. Experimental results show that the optimization model of RBF has higher accuracy in recognition of "oil slicks" and "look-alikes oil slicks" image.