基于图像处理的船舶水尺刻度识别  

Ship draft mark recognition based on image processing

作　　者：冉鑫[1] 彭将辉[1] 

机构地区：[1]上海海事大学商船学院,上海201306

出　　处：《上海海事大学学报》2012年第2期6-9,共4页Journal of Shanghai Maritime University

基　　金：上海市重点学科建设项目(S30602);上海海事大学校基金(20110006)

摘　　要：为克服传统船舶水尺刻度识别方法的不足,提出一种基于图像处理的自动识别方法.首先对船舶水尺图像进行二值化等预处理,采用结构特征提取算法检测图像中船舶水尺刻度的三叉点特征.三叉点特征从细化处理后的图像上提取,是图像中多个彼此相邻的同类特征点的集合.利用三叉点的位置特征构造图像的特征模板;利用模板匹配方法实现船舶水尺刻度的自动识别.该方法可有效解决采集的船舶水尺数字图像与特征模板尺寸不一致的问题.In order to overcome the limits of the traditional ship draft survey methods, an automatic rec- ognition method based on image processing is presented. After some image pre-processing like binaryza- tion, a structure feature extraction algorithm is used to detect the trigeminal point features of ship draft in the image. The trigeminal point features of ship draft are extracted from the thinned images, and they are the aggregates of the similar feature points which are adjacent to each other. The position features of the trigeminal point of ship draft is utilized to construct the feature template of image, and then the ship draft marks can be identified automatically by using the template matching method. The proposed method can solve the problem caused by the size mismatch between the feature template and the gathered ship draft digital image effectively.

关 键 词：船舶水尺 三叉点特征 模板匹配 图像处理 

分 类 号：U674.703.1[交通运输工程—船舶及航道工程] TP391.41[交通运输工程—船舶与海洋工程]