利用油罐目标相对集中而虚假目标相对较离散的分布特点,提出了一种基于区域生长原理的聚类方法.对于实际的遥感光学图像,所提方法既能将油罐目标聚类定位油库区域,又能自动去除虚假油罐目标.实验表明,该方法在油库检测定位上准确率高,是一种实用有效、运算量小的聚类方法.

Oil storage tanks are usually distributed closely in groups, while the false targets generally distract from each other. A region-growing clustering method is proposed in this paper, which can not only localize oil tanks but also remove the false oil tank targets automatically. The experimental results with real remote sensing images indicate that this method can localize oil house accurately, effectively, and at meantime is a practical and fast clustering method.