优化区域生长的高分辨率影像道路精确提取

发布时间：2018-06-25 04:37

本文选题：种子点自动获取 + 梯度值　；参考：《科学技术与工程》2017年33期

【摘要】：鉴于区域生长存在需提供种子点及易过生长的缺陷,改进了区域生长算法对高分辨率遥感影像进行道路精确提取。首先利用格网对原始影像进行划分,将在道路灰度范围内的外边缘格网以同质性指标依次排序作为候选种子点;将满足道路连续性和一致性要求的路径格网起始点作为种子点;并通过道路指数加以验证得到最终种子点,以此避免盲目选择种子点的不足。其次在引入梯度值进行道路生长后,以二值化的彩色分量与生长结果进行信息融合方式来优化算法,达到缓解过生长的目的。实验结果表明算法是合理有效的,能够较为精确地提取出道路区。
[Abstract]:In view of the defects of region growth which need to provide seed points and be easy to grow, an improved region growth algorithm is proposed for accurate road extraction from high-resolution remote sensing images. Firstly, the original image is divided by grid, and the outer edge grid in the range of road grayscale is ranked as the candidate seed point by homogeneity index, and the starting point of the path grid that meets the requirements of road continuity and consistency is taken as the seed point. Finally, the final seed points are obtained by road index, so as to avoid the shortage of blind selection of seed points. Secondly, after introducing gradient value for road growth, the information fusion of binary color component and growth result is used to optimize the algorithm to alleviate overgrowth. The experimental results show that the algorithm is reasonable and effective, and can extract the road area accurately.
【作者单位】：中国地质大学(武汉)信息工程学院;河南财经政法大学资源与环境学院;
【基金】：国家863计划资助项目(2007AA092102) 高分辨率对地观测重大专项(07-Y30A05-9001-12/13)资助
【分类号】：TP751

【相似文献】