基于不同滤波的水稻物候期提取  

Study on Phenology Extraction of Paddy Rice Based on Different Filtering Methods

作　　者：凌洋[1] 耿利宁[2] 景元书[2] 胡荣辰[1] 孙啸[1] 

机构地区：[1]江苏省气象局,南京210000 [2]南京信息工程大学,南京210044

出　　处：《科学技术与工程》2014年第35期16-22,共7页Science Technology and Engineering

基　　金：公益性行业(气象)科研专项(GYHY20090622);"十二五"农村领域国家科技计划课题(2011BAD32B01);江苏省"青蓝工程"项目资助

摘　　要：基于2010年中分辩率成像光谱仪(moderate-resolution imaging specroradiometer,MODIS)数据,以江苏省为研究区域,采用不同信号滤波方法,对水稻物候期的提取方法进行了分析和研究。在获得水稻的增强型植被指数(enhanced uegetatisn index,EVI)数据基础上,进行了HANTS(harmonic analysis of time series)滤波和小波变换滤波的对比分析,使用小波滤波重构后的数据结合Matlab软件进行了水稻物候期的格点化提取,并验证了结果的准确性。结论如下:二者都能较好的去除噪声,还原原始信息,但对云噪声污染较严重地区,小波滤波相比HANTS滤波效果更好,新的移栽期提取方法和小波滤波的物候期提取方法能够较为准确的反应真实的水稻物候情况,经站点数据检验,能很好地反应真实水稻物候情况。Aiming to extract the rice phenology of Jiangsu provinces using a variety of signal tiltenng methods baseu on multi-temporal MODIS imagery, in order to extract the phenology of paddy rice ,the EVI data of paddy rice was fil- tered by HANTS （harmonic analysis of time series） filtering and Wavelet filtering, and then Wavelet filtering was chosen combined with Matlab software to extract the grid data of phenology after compared the result of HANTS filtering with Wavelet filtering. At last, the accuracy of the result was in vestigated. In conclusion, The Wavelet filtering is more ef- fective than HANTS filtering in cloud-contaminated areas. The new method which extract the transplanting date and the wavelet filtering method which extract the other rice phenology can accurately reflect the real rice phenology.

关 键 词：水稻物候期提取 HANTS滤波 小波滤波 格点数据 

分 类 号：P407.8[天文地球—大气科学及气象学]