Forest type identification by random forest classification combined with SPOT and multitemporal SAR data

Abstract

We developed a forest type classification technology for the Daxing′an Mountains of northeast China using multisource remote sensing data. A SPOT-5 image and two temporal images of RADARSAT-2 full-polarization SAR were used to identify forest types in the Pangu Forest Farm of the Daxing′an Mountains. Forest types were identified using random forest (RF) classification with the following data combination types: SPOT-5 alone, SPOT-5 and SAR images in August or November, and SPOT-5 and two temporal SAR images. We identified many forest types using a combination of multitemporal SAR and SPOT-5 images, including Betula platyphylla, Larix gmelinii, Pinus sylvestris and Picea koraiensis forests. The accuracy of classification exceeded 88% and improved by 12% when compared to the classification results obtained using SPOT data alone. RF classification using a combination of multisource remote sensing data improved classification accuracy compared to that achieved using single-source remote sensing data.

Keywords

Random forest classification / Multitemporal / Multisource remote sensing data / Polarization decomposition

References

[1] Aghabalaei A, Maghsoudi Y, Ebadi H. Forest classification using extracted PolSAR features from compact polarimetry data. Adv Space Res, 2016, 57(9): 1939-1950.

[2] Borghys D, Yvinec Y, Perneel C, Pizurica A, Philips W. Supervised feature-based classification of multi-channel SAR images. Pattern Recognit Lett, 2006, 27(4): 252-258.

[3] Breiman L. Random forests. Mach Learn, 2001, 45(1): 5-32.

[4] Bu CGTZY (2007) Ministry of Land and Resources, P.R.C

[5] Cloude SR, Pottier E. An entropy based classification scheme for land applications of polarimetric SAR. IEEE Trans Geosci Remote Sens, 1997, 35(1): 68-78.

[6] Evans DL, Farr TG, VanZyl JJ, Zebker HA. SAR polarimetry: analysis tools and applications. IEEE Trans Geosci Remote Sens, 1988, 26(6): 774-789.

[7] Freeman A, Durden SL. A three-component scattering model for polarimetric SAR data. IEEE Trans Geosci Remote Sens, 1998, 36(3): 963-973.

[8] Furtado LF, Silva TSF, Novo EMLM. Dual-season and full-polarimetric C band SAR assessment for vegetation mapping in the Amazon varzea wetlands. Remote Sens Environ, 2016, 174: 212-222.

[9] Holm WA, Barnes RM (1988) On radar polarization mixed target state decomposition techniques. In: Radar, IEEE national conference—RADAR, pp 249–254

[10] Huett C, Koppe W, Miao Y, Bareth G. Best accuracy land use/land cover (LULC) classification to derive crop types using multitemporal, multisensor, and multi-polarization SAR satellite images. Remote Sens, 2016 8 8 684

[11] Huynen JR. Phenomenological theory of radar targets. Electromagnetic scattering, 1978, New York: Academic Press 653 712

[12] Kasapoglu NG, Anfinsen SN, Eltoft T (2012) Fusion of optical and multifrequency polsar data for forest classification. Paper presented at the 2012 IEEE international geoscience and remote sensing symposium. 22–27 July

[13] Krogager E. Properties of the sphere, diplane, helix (target scattering matrix) decomposition. Mol Ecol, 2006, 15(11): 3205-3217.

[14] Laurin GV, Frate FD, Pasolli L, Notarnicola C, Guerriero L, Valentini R. Discrimination of vegetation types in alpine sites with ALOS PALSAR-, RADARSAT-2-, and lidar-derived information. Int J Remote Sens, 2013, 34(19): 6898-6913.

[15] Lee JS, Grunes MR, Ainsworth TL, Du LJ. Unsupervised classification using polarimetric decomposition and the complex Wishart classifier. IEEE Int Geosci Remote Sens, 1998, 4(5): 2178-2180.

[16] Li X, Chen Y, Tong L, Luo SY. A study on vegetation cover extraction using a Wishart H- classifier based on fully polarimetric Radarsat-2 data. Int J Remote Sens, 2016, 37(12): 2844-2859.

[17] Ling F, Li ZY, Erxue C, Wang QK (2009) Comparison of ALOS PALSAR RVI and Landsat TM NDVI for forest area mapping. In: Asian-Pacific conference on synthetic aperture radar, pp 132–135

[18] Ma N, Hu YF, Zhuang DF, Wang XS. Determination on the optimum band combination of HJ-1AHyperspectral data in the case region of Dongguan based on optimum index factor and J-M distance. Remote Sens Technol Appl, 2010, 25(3): 358-365.

[19] Maghsoudi Y, Collins MJ, Leckie DG. Radarsat-2 polarimetric SAR data for boreal forest classification using SVM and a wrapper feature selector. IEEE J Sel Top Appl Earth Obs Remote Sens, 2013, 6(3): 1531-1538.

[20] Masjedi A, Zoej MJV, Maghsoudi Y. Classification of polarimetric SAR images based on modeling contextual information and using texture features. IEEE Trans Geosci Remote Sens, 2016, 54(2): 932-943.

[21] Niculescu S, Lardeux C, Grigoras I, Hanganu J, David L. Synergy between LiDAR, RADARSAT-2, and Spot-5 images for the detection and mapping of wetland vegetation in the Danube Delta. IEEE J Sel Top Appl Earth Obs Remote Sens, 2016, 9(8): 3651-3666.

[22] Rahman MM, Sumantyo JTS. Mapping tropical forest cover and deforestation using synthetic aperture radar (SAR) images. Appl Geomat, 2010, 2(3): 113-121.

[23] Richards JA, Jia XP. Remote sensing digital image analysis: an introduction, 1986 3 Berlin: Springer

[24] Sun H (2006) The research of Spot5 application in forest inventory. Thesis for the Master Degree. Central South University of Forestry and Technology, pp 52–55

[25] Touzi R, Landry R, Charbonneau FJ. Forest type discrimination using calibrated C-band polarimetric SAR data. Can J Remote Sens, 2004, 30(3): 543-551.

[26] Uhlmann S, Kiranyaz S. Integrating color features in polarimetric SAR image classification. IEEE Trans Geosci Remote Sens, 2014, 52(4): 2197-2216.

[27] Wang LH, Zhao ZY. Automatically Classifying and Identifying the TM remote sensing images of forest mixed with conifer and broadleaves using improved BP ANN. Sci Silvae Sin, 2005, 41(6): 95-100.

[28] Yamaguchi Y, Yajima Y, Yamada H. A four-component decomposition of POLSAR images based on the coherency matrix. IEEE Geosci Remote Sens Lett, 2006, 3(3): 292-296.