9 October 2009 Reed wetland extraction in the Yellow River Delta Nature Reserve based on knowledge inference technology
Author Affiliations +
Proceedings Volume 7471, Second International Conference on Earth Observation for Global Changes; 747111 (2009) https://doi.org/10.1117/12.836336
Event: Second International Conference on Earth Observation for Global Changes, 2009, Chengdu, China
Abstract
With the reduction of sediments into the sea, the area of reed wetland, which is the key habitat of red-crowned crane, has been shrinking in the Yellow River Delta Nature Reserve, China. With Landsat Thematic Mapper (TM) images and field observations, we mapped the reed wetland using the knowledge inference technology. Six wetland types were extracted using a supervised classification method. To resolve the confusions between reeds and other wetland types, a set of rules were established. Firstly, reed wetland was separated from mudflat wetland, rearing and shrimp pond and water body by using the normalized digital vegetation index (NDVI). Secondly, reed wetland was distinguished from paddy field by using image texture information. Thirdly, the reed wetland was separated from the Chinese tamarisk by using the principal transformation. All these rules were built by using ERDAS Imagine's knowledge engineer. Reed wetland classification was conducted by using the neighbor analysis technology. The accuracy assessment shows that the knowledge-based classification obtained an overall accuracy of 89.02% and kappa coefficient of 0.89, which was better than the traditional supervised classification.
© (2009) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Xiaomin Fu, Xiaomin Fu, Hong Wang, Hong Wang, Ling Li, Ling Li, } "Reed wetland extraction in the Yellow River Delta Nature Reserve based on knowledge inference technology", Proc. SPIE 7471, Second International Conference on Earth Observation for Global Changes, 747111 (9 October 2009); doi: 10.1117/12.836336; https://doi.org/10.1117/12.836336
PROCEEDINGS
7 PAGES


SHARE
Back to Top