基于ROC融合准则的SAR边缘检测算法

Edge detection algorithm of SAR images based on ROC fusion rule

根据ROC(receiver operating characteristics)技术能评估分类器在所有可能工作阈值下总体性能的特点,建立包含边缘像素点相关分析与ROC分类决策的ROC融合准则。依据该准则组合多种SAR边缘检测算子,并得到合成孔径雷达(SAR)影像的"理想"边缘检测结果。实验结果表明,本文方法能融合多种边缘检测算子的优点,有较强的开放性与目标适应性,并且不需要手工设置阈值,自动化程度高,有很强的工程实用性。

ROC（receiver operating characteristics） analysis method,which can evaluate the classifier performance under the condition of all working thresholds,is applied to the edge detection algorithm of SAR images.By combining the edge pixels correspondence analysis and ROC classification decision,ROC fusion rule is built.And ＂ideal＂ edge detection result is obtained using this rule which includes many SAR edge detectors.The experimental results show that the algorithm integrates the merit of different edge detectors,possesses opening interface,object suitability and high automation,and it can assist in practical implementations.