Automatic Threshold Selection Method for SAR Edge Detection

Abstract

In order to reduce the manual intervention and calculate amount in edge detection processing, an automatic threshold selection method for edge detection of synthetic aperture radar (SAR) images is proposed. The proposed automatic threshold selection (ATS) method determines the optimal thresholds by establishing a relationship between different thresholds and corresponding detection results. The thresholds chosen by the proposed ATS method are very close to the thresholds determined by manual intervention. Comparing with conventional method, ratio-based detectors with ATS obviously reduce the calculate amount and ensure the accuracy of edge detection at the same time, benefiting from less manual intervention and almost the same thresholds. The experimental results on SAR images show that ratio-based detectors combined with the proposed ATS method not only achieve the accurate detection results, but also effectively reduce the time consumption during thresholds selection.