摘要
本文基于2011年～2013年绿潮爆发期间的六景GOCI影像,利用目视解译方法构建了测试数据集,对比分析了各算法及其不同波段选择方式的绿潮探测能力。结果表明:NDVI算法的绿潮探测能力明显优于OSABI、KOSC、RVI、EVI 4种算法,且性能稳定可靠,可作为GOCI绿潮业务监测的首选算法;IGAG算法具有一定的潜力,但其绿潮探测能力具有显著的不确定性;从不同波段的选择来看,红光波段选择波段5或6差别不大;近红外波段选择波段7或波段8差异显著且变化规律不一致。所以,未来发展新的GOCI绿潮探测算法时,应综合利用GOCI近红外两个波段的信息。
Based on the spectral difference in red and near-infrared bands of green macro-algae and seawater,many green macro-algae detection algorithms have been put forward.The Geostationary Ocean Color Imager(GOCI),which was launched in 2010,made it possible to monitor the hourly movement of green macro-algae;however,the comparison of different macro-algae detection algorithms has not been performed.Based on six selected GOCI images from 2011 to 2013,the green macro-algae and seawater samples were selected,and the discrimination capabilities of macro-algae monitoring algorithms and their band combinations were analyzed quantitatively.We found that:(1)The extraction capability of NDVI was significantly better than OSABI,KOSC,RVI and EVI,which indicated that NDVI could be regarded as the preferred algorithm in GOCI green macro-algae monitoring application.The IGAG algorithm showed good potential,but it had significant uncertainty.(2 ) When monitoring the green macro-algae with the six above algorithms,choosing between GOCI band 5 (660nm)or 6 (680nm)as red band was flexible,while results between GOCI band 7(745nm)or 8(865)as NIR band showed noticeable variability.The new GOCI algorithm in the future should utilize the information from both NIR bands simultaneously.
出处
《遥感信息》
CSCD
2014年第5期44-50,共7页
Remote Sensing Information
基金
中国科技部欧洲空间局"龙计划"合作三期项目(10470)
关键词
绿潮
GOCI
探测能力
green macro-algae
GOCI
differentiation capability
