Application of BRDF model in land cover mapping
-
摘要: 植被的反射异质性特征可以反映植被的结构和光谱特性, 有助于更有效地识别植被.在塔里木河下游开展的研究中, 使用了MISR的多角度观测数据, 利用核驱动和RPV模型反演获取地表BRDF信息, 使用SVM方法对MISR天底角反射率数据和BRDF信息组合进行土地覆被分类研究, 对比分析了BRDF对分类效果的影响.发现：① BRDF信息可以为半干旱区土地覆被制图提供附加的有用信息, 提高制图的精度. ②核驱动模型和RPV模型都能较好地模拟研究区地表反射的状况. ③空间结构差异较大的草地和林地类型使用BRDF信息后的用户精度明显提高.Abstract: The reflection heterogeneity of vegetation can reflect the structural and spectral characteristics of vegetation, and therefore it can help to identify vegetation more effectively.Misr multi-angle observation data has been explored in the sutdy of the lower Tarim River.Meanwhile, the kernel driven and RPV model inversion is used to obtain information of surface BRDF, and SVM method is used to study the land use and the cover of classification by the combination of MISR nadir reflectance data and BRDF information. By comparative analysis of the effect of the BRDF of the classification results, the findings are as following: ① BRDF information can provide additional information for the land cover mapping in the semi-arid area and improve the accuracy of the mapping. ② Both the kernel driven model and the RPV model can simulate the surface reflection in the study area. ③ After using BRDF, the identification accuracy of grassland and forest land increased obviously.
-
Key words:
- BRDF /
- kernel driven model /
- RPV /
- misr /
- the lower Tarim river
-
图 6 各土地覆被类型核驱动模型参数平均值
Fig. 6 Average of kernel driven models parameters of each land cover type
表 1 土地覆被类型定义
Tab. 1 Land-cover classification system
土地类型 样地数量/个 覆盖度/% 描述 灌木 2034 >5 灌木, 半灌木植物群落 林地 1154 >5 胡杨林 水体 95 0 水库, 天然水体 未利用地 670 < 5 沙地, 盐碱地 耕地 383 >40 农田 草地 206 >5 盐生草本植物群落 注:本处覆盖度是指在一个MISR像元大小范围内(即275×275 m2), 所有植被包括胡杨林、灌木、盐生草本植物和农作物垂直投影面积占像元面积的百分比.不同土地覆被类型像元中的植被可能包含多种植被类型. 
下载: 导出CSV
表 2 RPV和核驱动模拟的精度评价
Tab. 2 Accuracy assessment of RPV and kernel driven model
RPV模型 核驱动模型 RMSE均值 0.007 718 0.007 238 RMSE标准差 0.002 39 0.002 9
下载: 导出CSV
表 3 MISR多角度观测数据集
Tab. 3 MISR multi-angle observation DataSet
数据集 描述 天底角 AN相机的蓝、绿、红光和近红外共4个波段的地表反射率数据 天底角+ $k$ , $r$ 0, $b$ } AN相机的4个波段的地表反射率数据以及RPV 模型反演的 $k$ , $r_0$ , $b$ 值 天底角+ $f_{\rm iso}$ , $f_{\rm vol}$ , $f_{\rm geo}$ } AN相机的4个波段的地表反射率数据以及核驱动 模型反演的 $f_{\rm iso}$ , $f_{\rm vol}$ , $f_{\rm geo}$ 值 天底角+ $f_{\rm iso}$ , $f_{\rm vol}$ , $f_{\rm geo}$ + $k$ , $r$ 0, $b$ AN相机的4个波段的地表反射率数据以及核驱动模型反演的 $f_{\rm iso}$ , $f_{\rm vol}$ , $f_{\rm geo}$ 值 和RPV模型反演的 $k$ , $r_0$ , $b$ 值
下载: 导出CSV
表 4 SVM法使用天底角数据集分类的混淆矩阵
Tab. 4 Confusion matrix of SVM on nadir DataSet
类型 灌木 林地 水体 未利用地 耕地 草地 总数 用户精度 灌木 1 175 114 0 36 3 10 1 338 0.88 林地 464 315 0 16 0 0 795 0.40 水体 0 2 70 0 0 0 72 0.97 未利用地 156 0 0 301 0 1 458 0.66 耕地 4 0 0 2 228 3 237 0.96 草地 58 19 0 7 0 44 128 0.34 总数 1857 450 70 362 231 58 3 028 生产者精度 0.63 0.70 1.00 0.83 0.99 0.76 0.704 425
下载: 导出CSV
表 5 SVM法使用天底角_RPV模型参数数据集分类的混淆矩阵
Tab. 5 Confusion matrix of SVM on nadir plus RPV model parameters DataSet
类型 灌木 林地 水体 未利用地 耕地 草地 总数 用户精度 灌木 1 081 185 0 51 3 18 1 338 0.81 林地 266 506 0 20 0 3 795 0.64 水体 1 0 69 0 1 1 72 0.96 未利用地 110 15 0 329 0 4 458 0.72 耕地 6 0 0 4 226 1 237 0.95 草地 34 18 0 2 1 73 128 0.57 总数 1498 724 69 406 231 100 3 028 生产者精度 0.72 0.70 1.00 0.81 0.98 0.73 0.754 293
下载: 导出CSV
表 6 SVM法使用天底角_核驱动模型参数分类的混淆矩阵
Tab. 6 Confusion matrix of SVM on nadir plus kernel driven model parameters DataSet
类型 灌木 林地 水体 未利用地 耕地 草地 总数 用户精度 灌木 1 122 146 0 53 5 12 1 338 0.84 林地 299 472 0 22 0 2 795 0.59 水体 1 1 70 0 0 0 72 0.97 未利用地 113 7 0 338 0 0 458 0.74 耕地 4 0 0 2 229 2 237 0.97 草地 36 19 0 2 0 71 128 0.55 总数 1575 645 70 417 234 87 3 028 生产者精度 0.71 0.73 1.00 0.81 0.98 0.82 0.760 238
下载: 导出CSV
表 7 SVM法使用天底角_核驱动模型参数_RPV模型参数分类的混淆矩阵
Tab. 7 Confusion matrix of SVM on nadir plus kernel driven and RPV model parameters DataSet
类型 灌木 林地 水体 未利用地 耕地 草地 总数 用户精度 灌木 1 089 177 0 48 3 21 1 338 0.81 林地 257 517 0 18 0 3 795 0.65 水体 2 0 69 0 1 0 72 0.96 未利用地 114 14 0 324 0 6 458 0.71 耕地 7 0 0 2 227 1 237 0.96 草地 28 20 0 1 2 77 128 0.60 总数 1497 728 69 393 233 108 3 028 生产者精度 0.73 0.71 1.00 0.82 0.97 0.71 0.760 568
下载: 导出CSV
-
[1] ROSS J. The Radiation Regime and Architecture of Plant Stands[M]. Boston, MA: W Junk, 1981. [2] MYNENI R B, ASRAR G. Radiative transfer in three-dimensional atmosphere vegetation media[J]. Journal of Quantitative Spectroscopy and Radiative Transfer, 1993, 49: 585-598. doi: 10.1016/0022-4073(93)90003-Z [3] CHEN J M, CIHLAR J. Quantifying the effect of canopy architecture on optical measurements of leaf area index using two gap size analysis methods[J]. IEEE Transactions on Geoscience and Remote Sensing, 1995, 33: 777-787. doi: 10.1109/36.387593 [4] MARTONCHIK J V, BRUEGGE C J, STRAHLER A H. A review of reflectance nomenclature used in remote sensing[J]. Remote Sensing Reviews, 2000, 19: 9-20. doi: 10.1080/02757250009532407 [5] LI X, STRAHLER A H. Geometric-optical modeling of a conifer forest canopy[J]. IEEE Transactions on Geoscience & Remote Sensing, 1985, ge-23(5): 705-721. http://ieeexplore.ieee.org/xpl/articleDetails.jsp?reload=true&arnumber=4072365&pageNumber%3D33832%26rowsPerPage%3D100 [6] ASNER G P. Contributions of multi-view angle remote sensing to land-surface and biogeochemical research[J]. Remote Sensing Review, 2000, 18: 137-162. doi: 10.1080/02757250009532388 [7] LIANG S, STRAHLER A H. Land surface bi-directional reflectance distribution function (BRDF): Recent advances and future prospects[J]. Remote Sensing Reviews, 2000, 18: 83-511. [8] MARTONCHIK J V, PINTY B, VERSTRAETE M M. Note on: An improved model of surface BRDF-atmospheric coupled radiation[J]. IEEE Transactions on Geosciences and Remote Sensing, 2002, 40: 1637-1639. doi: 10.1109/TGRS.2002.802505 [9] WANNER W, LI X, STRAHLER A H. On the derivation of kernels for kernel driven models of bidirectional reflectance[J]. Journal of Geophysical Research, 1995, 100: 21077-21090. doi: 10.1029/95JD02371 [10] ROUJEAN J L, LEROY M, DESCHAMPS P Y. A bidirectional reflectance model of the Earth's surface for the correction of remote sensing data[J]. Journal of Geophysical Research, 1992, 97: 20455-20468. doi: 10.1029/92JD01411 [11] LOVELAND T R, REED B C, BROWN J F, et al. Development of a global land cover characteristics database and IGBP DIScover from 1 km AVHRR data[J]. International Journal of Remote Sensing, 2000, 21: 1303-1330. doi: 10.1080/014311600210191 [12] LANGLEY S K, CHESHIRE H M, HUMES K S. A comparison of single date and multitemporal satellite image classifications in a semi-arid grassland[J]. Journal of Arid Environments, 2001, 49: 401-411. doi: 10.1006/jare.2000.0771 [13] CHOPPING M J, RANGO A, RITCHIE J C. Improved semi-arid community type differentiation with the NOAA AVHRR via exploitation of the directional signal[J]. IEEE Transactions on Geoscience and Remote Sensing, 2002, 40: 1132-1149. doi: 10.1109/TGRS.2002.1010900 [14] SU L, CHOPPING M J, RANGO A, et al. Differentiation of semi-arid vegetation types based on multi-angular observations from MISR and MODIS[J]. International Journal of Remote Sensing, 2007, 28: 1419-1424. doi: 10.1080/01431160601085995 [15] XAVIER A S, GALVÃO L S. View angle effects on the discrimination of selected Amazonian land cover types from a principal-component analysis of MISR spectra[J]. International Journal of Remote Sensing, 2005, 26: 3797-3811. doi: 10.1080/09500340500106931 [16] LIESENBERG V, GALVÃO L S, PONZONI F J. Variations in reflectance with seasonality and viewing geometry: Implications for classification of Brazilian savanna physiognomies with MISR/Terra data[J]. Remote Sensing of Environment, 2007, 107(1/2): 276-286. http://www-misr.jpl.nasa.gov/mission/pubs/fulltext/liesenberg2007.pdf [17] 阿布都沙拉木⋅加拉力丁, 克力木⋅买买提, 海米提⋅依米提.塔里木河流域胡杨保护区土地覆盖变化研究[J].干旱区资源与环境, 2010, 24(1): 154-157. http://www.cnki.com.cn/Article/CJFDTOTAL-GHZH201001031.htm [18] BUHE A, TSUCHIYA K, KANEKO M, et al. Land cover of oases and forest in XinJiang, China retrieved from ASTER data[J]. Advances in Space Research, 2007, 39(1): 39-45. doi: 10.1016/j.asr.2006.02.056 [19] 黄粤, 包安明, 王士飞, 等.间歇性输水影响下的2001-2011年塔里木河下游生态环境变化[J].地理学报, 2013, 68(9): 1251-1262. http://www.cnki.com.cn/Article/CJFDTOTAL-DLXB201309009.htm [20] DINER D J, MARTONCHIK J V, BOREL C, et al. Multi-angle Imaging Spectro-Radiometer Level Surface Retrieval Algorithm Theoretical Basis Document[EB/OL]. [2016-02-01]. http://eospso.gsfc.nasa.gov/eos_homepage/for_scientists/atbd/docs/MISR/atbd-misr-10.pdf. [21] DINER D, BECKERT J, REILLY T, et al. Multi-angle Imaging Spectro Radiometer (MISR):Instrument description and experiment overview[J]. IEEE Transactions on Geoscience and Remote Sensing, 1998, 36(4): 1072-1087. doi: 10.1109/36.700992 [22] RAHMAN H, DEDIEU G. SMAC: A simplified method for the atmospheric correction of satellite measurements in the solar spectrum[J]. International Journal of Remote Sensing, 1994, 15(1): 123-143. doi: 10.1080/01431169408954055 [23] RAHMAN H, PINTY B, VERSTRAETE M M. Coupled surface-atmosphere reflectance (CSAR) model 2.Semiempirical surface model usable with NOAA Advanced Very High 93 Resolution Radiometer data[J]. Journal of Geophysical Research, 1993, 98: 20791-20801. doi: 10.1029/93JD02072 [24] MARTONCHIK, JOHN V. Determination of aerosol optical depth and land surface directional reflectances using multiangle imagery[J]. Journal of Geophysical Research Atmospheres, 1997, 102(D14): 17015-17022. doi: 10.1029/96JD02444 [25] ROBERTS G. A review of the application of BRDF models to infer land cover parameters at regional and global scales[J]. Progress in Physical Geography, 2001, 25(4): 483-511. doi: 10.1191/030913301701543154 [26] VAPNIK V. The Nature of Statistical Learning Theory[M]. New York: Springer-Verlag, 1995. [27] MOUNTRAKIS G, IM J, OGOLE C. Support vector machines in remote sensing: A review[J]. Isprs Journal of Photogrammetry & Remote Sensing, 2011, 66(3): 247-259. https://www.researchgate.net/profile/Giorgos_Mountrakis/publication/222302561_Support_vector_machines_in_remote_sensing_A_review/links/004635228ff86ae1ea000000.pdf [28] CHANG C C, LIN C J. LIBSVM: A library for support vector machines[J]. Acm Transactions on Intelligent Systems & Technology, 2011, 2(3): 389-396. http://dame.dsf.unina.it/documents/libsvm.pdf [29] LUCHT W, SCHAAF C B, STRAHLER A H. An algorithm for the retrieval of albedo from space using semiempirical BRDF models[J]. IEEE Transactions on Geoscience & Remote Sensing, 2000, 38(2): 977-998. http://web.gps.caltech.edu/~vijay/Papers/BRDF/lucht-etal-00.pdf [30] CHOPPING M J.Testing LiSK BRDF models over a semi-arid grassland region with visible and near-infrared ATSR-2 and AVHRR data[J].International Journal of Remote Sensing, 2001, 22(17): 3533-3552. doi: 10.1080/01431160010007024 [31] LAVERGNE T, KAMINSKI T, PINTY B, et al. Application to MISR land products of an RPV model inversion package using adjoint and Hessian codes[J]. Remote Sensing of Environment, 2007, 107(1/2): 362-375. https://www.researchgate.net/profile/Bernard_Pinty/publication/222570457_Application_to_MISR_land_products_of_an_RPV_model_inversion_package_using_adjoint_and_Hessian_codes._Remote_Sens._Environ._107(1-2)362-375/links/0912f510656532ff99000000.pdf -
点击查看大图
图(8) / 表(7)
计量
- 文章访问数: 430
- HTML全文浏览量: 122
- PDF下载量: 651
- 被引次数: 0