基于变异系数法的SAR船舶检测优化研究

发布时间：2018-04-08 09:03

  本文选题：船舶检测　切入点：高分辨SAR　出处：《遥感技术与应用》2017年02期

【摘要】：船舶检测在民用和军用领域都具有广阔的前景。利用高分辨率SAR图像可以提高海上船舶检测精度,但同时也面临船舶旁瓣、海表漂浮物及人工设施等物体的干扰。针对这些不足,提出了一种基于最佳熵双阈值算法对SAR影像检测的基础上,利用多种几何属性特征组合优化初步检测结果的算法。首先,基于最佳熵双阈值算法对研究区域的Radarsat-2影像进行初步检测,得到潜在船只目标;其次,计算潜在目标的核密度、长宽比、潜在像元数目等3个特征;最后,利用客观权重分配方法——变异系数法,针对3个特征进行权重分配,降低虚警率,达到优化船只检测结果的目的,同时利用AIS数据、K-CFAR算法检测结果及黄河口的ALOS-2数据验证了该算法的有效性和实用性。
[Abstract]:Ship detection has broad prospects in both civil and military fields.In order to solve these problems, a new algorithm based on optimal entropy double threshold algorithm for SAR image detection is proposed, which optimizes the initial detection results by combining various geometric attributes.The objective weight allocation method, variation coefficient method, is used to distribute the weights according to the three characteristics to reduce the false alarm rate and to optimize the detection results of ships.At the same time, the validity and practicability of the algorithm are verified by using the AIS data and the ALOS-2 data of the Yellow River estuary.
【作者单位】： 中国科学院遥感与数字地球研究所;中国科学院大学;成都理工大学;
【基金】：高分辨率对地观测系统重大专项(01-Y30A03-9001-12/13)
【分类号】：TN957.52
，

本文编号：1720926