改进的形状上下文在舰船图像匹配中的应用  

The Improved Shape Context In the Application of Warship Image Matching Algorithm

作　　者：李黎[1] 戴军[1] 杨旭[1] 

机构地区：[1]西南技术物理研究所,四川成都610041

出　　处：《数字技术与应用》2016年第3期76-78,80,共4页Digital Technology & Application

摘　　要：结合舰船图像的特点,对经典的形状上下文算法在抗几何形变和复杂度上做了改进。首先提取出舰船的二值轮廓图,利用css角点检测,做为舰船轮廓特征点,通过改进的形状上下文算法对特征点生成描述子,以此作为舰船轮廓的特征,然后利用相似性度量定义的代价函数计算特征之间的相似度。针对图像匹配技术的研究,结合改进后的算法,建立舰船图像特征数据库,对舰船图像进行匹配验证。实验表明,该算法能够有效的对舰船图像进行匹配识别。According to the characteristics of the image, improve the classic shape context algorithm on resistance to geometric deformation and the complexity. First of all, extract the ship binary contour map as a ship outline feature points by CSS comer detection. Then, create the descriptor with feature points by improving the shape of the context algorithm as the ship contour features. Then, calculate the similarity between features using the cost function defined by similarity measure. Establish the database of ship image feature according to the research of image matching technology and combined with the improved algorithm, then, do the ship image matching. Experiments show that this algorithm can effectively image matching recognition for ships.

关 键 词：形状上下文 角点检测 目标匹配 相似性度量 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]