中文摘要：

形状检测在目标识别中有着重要作用。广义Hough变换（GHT）具有很好的全局特性，是常用的目标形状检测算法。但由于SAR影像具有强噪声等特点，常用的广义Hough变换算法可能会产生定位不准确的问题，甚至会出现错检的情况。文中将GHT中投票不集中的区域看做“模糊投票点”，构造了隶属函数，并据此设计了模糊广义Hough变换算法（FGHT），通过优化参考点的位置达到检测结果精确定位的目的。实验对比证明，该算法在强噪声情况下具有更好的鲁棒性。

英文摘要：

Shape detection plays an important role in object recognition. Generalized Hough Transform（GHT） arithmetic is now applied abroad on object shape detection on account of its good global character. But because SAR image has strong noise, the position may not be accurate by the usual Generalized Hough Transform, even the result of the detection is wrong. In this paper, we regarded the unfocused area as a ＂fuzzy voting point＂, and constructed a subjection function. According to it, the FGHT arithmetic had been designed. This arithmetic could make the result of the detection accurate through optimizing the position of the reference points. From experiments we could see that our arithmetic was more robust on strong noise condition.