A comparison of multiple-point statistics and two-point statistics for spectral-spatial land cover classification | |
Shi, Zhongkui ; Li, Peijun | |
2016 | |
英文摘要 | Incorporating spatial information quantified by geostatistical tools to improve the classification of remotely sensed data is a promising method. Most existing studies only focused on the development of geostatistics-based spectral-spatial classification methods. This paper investigates primarily the advantages and disadvantages of two different types of geostatistical methods (two-point statistics and multiple-point statistics) for spectral-spatial land cover classification, by quantitatively comparing the classification results through an accuracy assessment. Two study areas were selected to evaluate the performance of two methods. Experimental results demonstrate that the classification result by including spatial information using multiple-point statistics obtained a higher accuracy than that by using two-point statistics. ? 2016 IEEE.; EI; 255-259 |
语种 | 英语 |
出处 | 4th International Workshop on Earth Observation and Remote Sensing Applications, EORSA 2016 |
DOI标识 | 10.1109/EORSA.2016.7552808 |
内容类型 | 其他 |
源URL | [http://ir.pku.edu.cn/handle/20.500.11897/449409] ![]() |
专题 | 地球与空间科学学院 |
推荐引用方式 GB/T 7714 | Shi, Zhongkui,Li, Peijun. A comparison of multiple-point statistics and two-point statistics for spectral-spatial land cover classification. 2016-01-01. |
个性服务 |
查看访问统计 |
相关权益政策 |
暂无数据 |
收藏/分享 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。
修改评论