The Application of Continuous Spatial Scaling Model of NDVI in Validation and Its Sensibility Analysis to Various Image Characteristics
The Application of Continuous Spatial Scaling Model of NDVI in Validation and Its Sensibility Analysis to Various Image Characteristics
复制标题
DOI：
10.17265/2162-5298/2016.10.006
复制
发表时间：
2016-10
期刊：
影响因子：
--
通讯作者：
H. Luan;Q. Tian;Xinxin Zhang;Qin Nie;Hui Li;Huoping Pan
中科院分区：
文献类型：
--
作者：
H. Luan;Q. Tian;Xinxin Zhang;Qin Nie;Hui Li;Huoping Pan
基于分形理论，建立了归一化植被指数(NDVI)的连续空间尺度模型(CSSM)，以解决定量遥感中空间尺度放大带来的问题。该模型能够定量描述NDVI在连续尺度上的转换关系。(1)以GeoEye-1影像及其NDVI CSSM影像为基础，对ETM+NDVI影像进行了验证，得出以下结论：ETM+影像由于条带较差，去条纹效果有限，误差较大，整个实验影像的误差在25%左右，ETM+NDVI影像不适合直接实际应用；(2)以沙田比地(北海市，中国)为实验区，研究了四幅影像(两幅覆盖面较大和较小的ETM+影像，一幅是GeoEye-1影像，一幅是HJ-1B CCD1影像)。计算并比较了四种图像的最合适尺度水平，更好地理解了各种图像特征(覆盖面积、空间分辨率和成像质量)对确定NDVI CSSM尺度水平的影响。
Based on the fractal theory, this study establishes a Continuous Spatial Scaling Model (CSSM) of the Normalized Difference Vegetation Index (NDVI) to address issues arising with spatial up-scaling in quantitative remote sensing. This model is able to quantitatively describe transformation relationships of the NDVI on continuous scales. Then the following experiments are accomplished: (1) the validation of ETM+ NDVI imagery is implemented based on the GEOEYE-1 image and its NDVI CSSM, and the following conclusion is obtained: because of bad stripes in the ETM+ image and the limited effect of destriping, the ETM+ NDVI image had a rather large error, and the error for the entire experimental imagery is about 25%, so the ETM+ NDVI product is not suitable for direct practical application; (2) Shatian Byland (Beihai City, China) is taken as the experimental area, and four images (two ETM+ images with wider and smaller coverage, respectively, a GEOEYE-1 image, and an HJ-1B CCD1 image) are studied. The most suitable scale levels are computed and compared for the four images, and a better understanding is obtained of the impact of various image characteristics (area of coverage, spatial resolution, and imaging quality) on determining the scale level for the NDVI CSSM.