Proceedings Article | 19 October 2012
KEYWORDS: Image filtering, Digital filtering, Speckle, Synthetic aperture radar, Wavelet transforms, Denoising, Wavelets, Nonlinear filtering, Statistical analysis, Mathematical modeling
At present, there are two types of method to detect ships in SAR images. One is a direct detection type, detecting ships
directly. The other is an indirect detection type. That is, it firstly detects ship wakes, and then seeks ships around wakes.
The two types all effect by speckle noise. In order to improve the accuracy of ship detection and get accurate ship and
ship wakes parameters, such as ship length, ship width, ship area, the angle of ship wakes and ship outline from SAR
images, it is extremely necessary to remove speckle noise in SAR images before data used in various SAR images ship
detection. The use of speckle noise reduction filter depends on the specification for a particular application. Some
common filters are widely used in speckle noise reduction, such as the mean filter, the median filter, the lee filter, the
enhanced lee filter, the Kuan filter, the frost filter, the enhanced frost filter and gamma filter, but these filters represent
some disadvantages in SAR image ship detection because of the various types of ship. Therefore, a mathematical
function known as the wavelet transform and multi-resolution analysis were used to localize an SAR ocean image into
different frequency components or useful subbands, and effectively reduce the speckle in the subbands according to the
local statistics within the bands. Finally, the analysis of the statistical results are presented, which demonstrates the
advantages and disadvantages of using wavelet shrinkage techniques over standard speckle filters.
