遥感影像主特征线检测

发布时间：2018-01-25 13:12

本文关键词： 主特征线格式塔法则视觉显著性累计权重矩阵直线统计矩阵　出处：《遥感学报》2017年02期 　论文类型：期刊论文

【摘要】：受到成像规律性差,背景纹理复杂及强烈噪声的影响,直线检测方法通常难以适应遥感影像处理的需求。有鉴于此,论文提出一种具有视觉显著性的遥感影像主特征线检测方法。论文首先论证了利用已提取直线为基元,基于格式塔法则构建主特征线的可行性;其次对直线在主特征线上的复杂投影情况进行了详细的剖析,并给出了主特征线的定义;接着建立了主特征线累计权重矩阵及直线统计矩阵,依据格式塔原则分析直线权重分布规律,以此构建了直线的权重模型,同时探讨不同直线在同一主特征线上权重分配规律;最后依据上述分析结果提出了具体的算法步骤。通过多幅含有强烈噪声的光学与SAR遥感卫星影像实验结果表明,相对于其他聚类算法,论文算法能够在杂乱无序的直线集中提取较为清晰的主特征线,并且实验效果基本符合人工视觉感知,便于机器对遥感影像的清晰理解。
[Abstract]:Under the influence of poor imaging regularity, complex background texture and strong noise, line detection methods are usually difficult to meet the needs of remote sensing image processing. This paper presents a visual salient detection method for the main feature line of remote sensing image. Firstly, the paper demonstrates the feasibility of constructing the main feature line based on the Gestalt rule by using the extracted line as the basic element. Secondly, the complex projection of a line on the main feature line is analyzed in detail, and the definition of the main feature line is given. Then the main characteristic line cumulative weight matrix and linear statistical matrix are established. According to the Gestalt principle the linear weight distribution law is analyzed and the linear weight model is constructed. At the same time, the law of weight distribution of different lines on the same main characteristic line is discussed. Finally, based on the above analysis results, the specific algorithm steps are proposed. The experimental results of several optical and SAR remote sensing satellite images with strong noise show that compared with other clustering algorithms. In this paper, the algorithm can extract the main feature line in the chaotic straight line, and the experimental results basically accord with the artificial visual perception, so it is convenient for the machine to understand the remote sensing image clearly.
【作者单位】：辽宁工程技术大学测绘与地理科学学院;中国测绘科学研究院;
【基金】：国家自然科学基金(编号:41271374,61540056,41401535) 对地观测技术国家测绘地理信息局重点实验室开放基金项目(编号:K201402) 资源与环境信息系统国家重点实验室开放基金~~
【分类号】：TP751
【正文快照】： 引用格式:戴激光,张力,朱恩泽,廖健驰,方鑫鑫,李晋威.2017.遥感影像主特征线检测.遥感学报,21(2):228 238Dai J G,Zhang L,Zhu E Z,Liao J C,Fang X X and Li J W.2017.Principal line detection in remote sensing image.直线作为一种重要的几何信息,在立体匹配、图像压缩、

【相似文献】