Constant false alarm rate detection method in polarimetric synthetic aperture radar imagery based on whitening filter under inverse beta distribution

Abstract

The constant false alarm rate (CFAR) detection method based on multi-look polarisation whitening filter (MPWF) has not been obtained in the polarimetric synthetic aperture radar imagery when the texture variable obeys the Inverse Beta distribution. To solve the problem, a new CFAR detection method with analytical expression for false alarm rate is proposed. First, the probability density function (PDF) of the output via MPWF is derived based on the hypothesis of the texture variables obeying the Inverse Beta distribution in the product model. Secondly, the analytical expression of false alarm rate about detection threshold is obtained by integrating the PDF of MPWF output, and the procedure of the CFAR detection method is designed. Finally, the performance comparisons are made among the new method, Wishart CFAR detection algorithm and the two-parameter CFAR method via simulated and real data of AIRSAR. The results show that figure of merit of the new method is the highest and the new method gives the best performance in Inverse Beta distribution case.

Keywords

• radar detection
• radar imaging
• synthetic aperture radar
• radar polarimetry
• remote sensing by radar
• image texture
• statistical distributions
• geophysical image processing
• filtering theory
• constant false alarm rate detection method
• polarimetric synthetic aperture radar imagery
• polarisation multilook polarisation whitening filter
• texture variable
• CFAR detection method
• analytical expression
• detection threshold
• MPWF output
• probability density function
• AIRSAR data
• inverse beta distribution