| 面向对象的船载激光扫描数据船舶目标检测方法 |
| |
| 引用本文： | 邵磊,董广军,姚强强,张阿龙.面向对象的船载激光扫描数据船舶目标检测方法[J].测绘科学技术学报,2017(5):481-485. |
| |
| 作者姓名： | 邵磊 董广军 姚强强 张阿龙 |
| |
| 作者单位： | 信息工程大学,河南郑州,450001 |
| |
| 基金项目： | 国家自然科学基金项目(61272146) |
| |
| 摘 要： | 提出了一种基于对象的船载激光扫描数据内河船舶目标检测方法。该方法基于船舶目标的高程信息和几何特征,利用投影和图像形态学运算方法快速确定包含船舶目标的兴趣区域;然后针对船舶目标数据的空间分布特征,对兴趣区域点云数据进行欧式距离分割和优化处理,生成实体对象,并构建高差、外包围盒和局部密度等特征构成的向量,采用支持向量机进行二元分类,提取出船舶目标。该方法成功地检测出船载激光扫描数据中的船舶目标,船舶目标检测的准确率较高。
|
| 关 键 词： | 船载激光扫描 兴趣区域 点云分割 特征提取 船舶目标 |
Object-Oriented Extraction of Ship Targets from Boat-Based Laser Scanning Data |
| |
| Abstract: | An object-based method for detecting riverboat targets from boat-based laser scanning data is proposed in this paper.The elevation information and geometric features of the riverboat targets combining with projection and image morphological operation methods are employed to determine the region of interest which includes these riverboat targets.Then by using the spatial distribution feature of riverboat targets,segmentation and optimization based on the Euclidean distance of the region of interest are exploited to generate entity objects,meanwhile characteristics of each entity object such as height,outer boundary box,and local density are calculated after segmentation.Finally,the support vector machine(SVM) is used for binary classification to extract the ship targets.Experimental results show that riverboat targets can be detected successfully and the overall accuracy of riverboat targets extraction is satisfactory. |
| |
| Keywords: | boat-based laser scanning region of interest segmentation feature extraction boat targets |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
