کاربرد منحنی ROC در ارزیابی روش‌های طبقه‌بندی پیکسل- ‌پایه روی تصاویر هوایی UltraCam-D برای تفکیک تاج درختان در توده‌های خالص بلوط ایرانی در جنگل‌های زاگرس

نوع مقاله : علمی- پژوهشی

نویسنده

هیات علمی دانشگاه شیراز

چکیده

مدیریت پایدار جنگل در زاگرس نیازمند برآورد تراکم تاج‌پوشش درختان جنگلی است و صحت نقشه‌های مربوط به این ویژگی که از سنجش از دور به‌دست آمده است، باید با روش‌های مناسب مورد ارزیابی قرار گیرد. این پژوهش با هدف ارزیابی نتایج سه طبقه‌بندی‌کننده پیکسل- ‌پایه روی تصاویر هوایی UltraCam-D برای تفکیک تاج درختان بلوط ایرانی در جنگل‌های زاگرس به‌وسیله منحنی عملیاتی دریافت‌کننده (ROC) انجام شد. یک قطعه‌نمونه با مساحت 30 هکتار در بخشی از توده‌های خالص بلوط ایرانی این جنگل‌ها انتخاب شد و موقعیت مکانی و محدوده تاج همه درختان آن به‌طور کامل مساحی شدند. تصویر هوایی UltraCam-D منطقه موردنظر بااستفاده از طبقه‌بندی‌کننده‌های حداکثر احتمال (ML)، شبکه عصبی مصنوعی (ANN) و ماشین بردار پشتیبان (SVM) طبقه‌بندی شد. سپس نتایج حاصل از طبقه‌بندی بااستفاده از منحنی ROC و شاخص‌های صحت کلی و ضریب کاپا و نقشه واقعیت زمینی مورد ارزیابی قرار گرفت. نتایج نشان داد بیشترین سطح زیر منحنی ROC مربوط به «تاج درختان» در طبقه‌بندی‌کننده ML بوده است (894/0) و در مقابل، کمترین سطح زیر منحنی مربوط به طبقه‌بندی‌کننده SVM بود (819/0). حساسیت و ویژگی «تاج درختان» در طبقه‌بندی‌کننده ML (به‌ترتیب 999/0 و 999/0) بیشتر از دو طبقه‌بندی‌کننده دیگر بود. اگرچه دقت طبقه‌بندی‌کننده SVM در تفکیک «تاج درختان» حداکثر مقدار ممکن بود (000/1)، اما صحت این طبقه در طبقه‌بندی‌کننده‌ ML (999/0) بیشتر بود. به‌طور کلی این پژوهش نشان داد که منحنی ROC قادر به ارزیابی صحت و دقت روش‌های طبقه‌بندی پیکسل- پایه موردبررسی روی تصاویر هوایی UltraCam-D به‌منظور تفکیک «تاج درختان» بوده است.

کلیدواژه‌ها

عنوان مقاله [English]

Application of ROC curve to assess pixel-based classification methods on UltraCam-D aerial imagery to discriminate tree crowns in pure stands of Brant`s oak in Zagros forests

نویسنده [English]

  • Seyyed Yousef Erfanifard
Assistant Professor, Department of Natural Resources and Environment, College of Agriculture, Shiraz University, ‎Shiraz, I.R.‎
چکیده [English]

Sustainable forest management in Zagros Mountains entails accurate information on tree crown density, which could be possibly derived from remote sensing data. Moreover, those remote sensing products need to be objectively evaluated. In this study, the results of three pixel-based classifiers of UltraCam-D aerial imagery were evaluated for classifying Brant`s oak (Quercus brantii Lindl.) crowns in Zagros forests in western Iran. This was carried out by means of receiver operating characteristic (ROC) curve. Therefore, a 30 ha plot was selected in pure Brant`s oak stand, in which the location and crown area of all trees were mapped. The UltraCam-D aerial imagery was classified by maximum likelihood (ML), artificial neural networks (ANNs) and support vector machines (SVMs) classifiers. The classification results were then evaluated by ROC curve and were presented by overall accuracy and Kappa coefficient. Results showed that the ML classified returned the largest area under ROC curve of "tree crowns" (0.894), whereas the lowest rate was found for SVM classifier (0.819). Sensitivity and specificity of "tree crowns" in ML classifier (0.999 and 0.999, respectively) were higher than those in two other classifiers. Although the precision of SVM classifier was the highest in discriminating "tree crowns" (1.000), the achieved accuracy of tree “crown class” was higher for ML classifier (0.999). This study concluded that using ROC curve enables an evaluation accuracy and precision of common pixel-based classifiers of such aerial imagery to discriminate tree crowns.

کلیدواژه‌ها [English]

  • Tree crown
  • Zagros
  • pixel-based classifiers
  • ROC curve
  • UltraCam-D

مراجع

- Abdolahi, H. and Shataee, Sh. 2012. Comparative evaluation of IRS-P6-LISS-III and LISS IV images for canopy cover mapping of Zagros forests (Case study: Javanroud Forests). Iranian Journal of Wood & Forest Science and Technology, 19 (1): 43-60 (In Persian).
- Ahmadi Sani, N., Darvishsefat, A.A., Zobeiri, M. and Farzaneh, A. 2008. Potentiality of ASTER images for forest density mapping in Zagros (Case study: Marivan forests). Iranian Journal of Natural Resources, 61 (3): 603-614 (In Persian).
- Ahmadpour, A., Solaimani, K., Shokri, M. and Ghorbani, J. 2011. Comparison of three common methods in supervised classification of satellite data for vegetation studies. Iranian Journal of Applied RS & GIS Techniques in Natural Resource Science, 2 (2): 69-81.
- Amini, M.R., Shataee, Sh., Ghazanfari, H.O. and Moaieri, M.H. 2008. Changes in Zagros's forests extension using aerial photos and satellite imagery (Case study, Armerdeh forests of Baneh). Iranian Journal of Agricultural Sciences and Natural Resources, 15 (2): 1-12 (In Persian).
- Alatorre, L.C., Andrés, R., Cirujano, S., Beguería, S. and Carrillo, S. 2011. Identification of mangrove areas by remote sensing: The ROC curve technique applied to the Northwestern Mexico coastal zone using Landsat imagery. Remote Sensing, 3: 1568-1583.
- Alonso, M.C. and Malpica, J.A. 2010. Satellite imagery classification with LiDAR data.International Archives of the Photogrammetry, Remote Sensing and Spatial Information Science. Japan, 9-12 Aug. 2011, 38 (8): 730-735.
- Brookhouse, M., Brack, C. and McElhinny, C. 2010. The distance to structural complement (DiSCo) approach for expressing forest structure described by Aerial Photograph Interpretation data sets. Forest Ecology and Management, 260 (7): 1230-1240
- Camps-Valls, G., Gomez-Chova, L., Calpe-Maravilla, J., Martin-Guerrero, J.D., Soria-Olivas, E., Alonso-Chorda, L. and Moreno, J. 2004. Robust support vector method for hyperspectral data classification and knowledge discovery. IEEE Transactions on Geoscience and Remote Sensing, 42: 1530-1542.
- Camps-Valls, G., Tuia, D., Gomez-Chova, L., Jimenez S. and Malo, J. 2012. Remote Sensing Image Processing. Morgan & Claypool Publishers, 176 p.
- Chen, C.H. 2007. Signal and image processing for remote Sensing. Taylor & Francis Publishers, 672 p.
- Colgan, M.S., Baldeck, C.A., Feret J.B. and Asner, G.P. 2012. Mapping savanna tree species at ecosystem scales using support vector machine classification and BRDF correction on airborne hyperspectral and LiDAR data. Remote Sensing, 4: 3462-3480.
- Congalton, G.C. and Green, K. 1999. Assessing the accuracy of remotely sensed data: principles and practices. Lewis Publishers, 133 p.
- Darvishsefat, A.A., Rafieyan, O., Babaii Kafaki, S. and Mataji, M. 2010. Evaluation of UltraCam-D images capability for tree species identification using object-based method in the even-aged mixed forestation. Iranian Journal of Forest, 2 (2): 165-176 (In Persian).
- De Jong, S.M. and Van Der Meer, F.D. 2005. Remote sensing image analysis: including the spatial domain. Springer, 359 p.
- Dobreva, I.D. and Klein, A.G. 2011. Fractional snow cover mapping through artificial neural network analysis of MODIS surface reflectance. Remote Sensing of Environment, 115 (12): 3355-3366.
- Erfanifard, Y., Zobeiri, M., Feghhi, J. and Namiranian, M. 2007. Estimation of crown cover on aerial photographs using shadow index (Case study: Zagros forests, Iran). Iranian Journal of Forest and Poplar Research, 15 (3): 278-288 (In Persian).
- Franklin, S.E. 2001. Remote sensing for sustainable forest management. Lewis Publishers, 407 p.
- Ghanbari, F. and Shataee, Sh. 2011. Investigation on forest extend changes using aerial photos and ASTER imagery (Case study: Border forests in south and southwest of Gorgan city). Iranian Journal of Wood & Forest Science and Technology, 17 (4): 1-18 (In Persian).
- Jazirehi, M.H. and Ebrahimi Rostaghi, M. 2003. Silviculture in Zagros. University of Tehran Press, 560 p (In Persian).
- Jones, S. and Reinke, K. 2009. Innovations in remote sensing and photogrammetry. Springer, 453 p.
- Klobucar, D., Pernar, R., Loncaric S. and Subasic, M. 2008. Artificial neural networks in the assessment of stand parameters from an IKONOS satellite image. Croatian Journal of Forest Engineering, 29: 201-211.
- Li, Z., Chen, J. and Baltsavias, E. 2008. Advances in photogrammetry, remote sensing and spatial information sciences. Taylor & Francis Publishers, 527 p.
- Mathieu, R., Naidoo, L., Cho, M.A., Leblon, B., Main, R., Wessels, K., Asner, G.P., Buckley, J., VanAardt, J., Erasmus, B.F.N. and Smit, I.P.J. 2013. Toward structural assessment of semi-arid African savannahs and woodlands: The potential of multitemporal polarimetric RADARSAT-2 fine beam images. Remote Sensing of Environment, 138: 215-231.
- Moeinazad Tehrani, S.M., Darvishsefat, A.A. and Namiraniyan, M. 2008. Evaluation of FCD model for estimation of forest density using Landsat 7 imagery (Case study: Chalus forst). Iranian Journal of Forest and Poplar Research, 16 (1): 124-138 (In Persian).
- Mondal, A., Kundu, S., Chandniha, S., Shukla, R. and Mishra, P.K. 2012. Comparison of support vector machine and maximum likelihood classification technique using satellite imagery. International Journal of Remote Sensing and GIS, 1: 116-123.
- Negri, R.G., Dutra, L.V. and Sant'Anna, S.J.S. 2014. An innovative support vector machine based method for contextual image classification. ISPRS Journal of Photogrammetry and Remote Sensing, 87: 241-248.  
- Nouredini, S.A.R., Bonyad, A.A. and Pourshakori, F. 2012. Investigation on the some textural indices from aerial photographs for forest canopy density mapping (Case study: Khoramabad, Lorestan). Iranian Journal of Forest, 4 (4): 353-363 (In Persian).
- Pacifici, F., Chini M. and Emery, W.J. 2009. A neural network approach using multi-scale textural metrics from very high-resolution panchromatic imagery for urban land-use classification. Remote Sensing of Environment, 113 (6): 1276-1292.
- Paine, D.P. and Kiser, J.D. 2012. Aerial photography and image interpretation (3rd ed.). John Wiley & Sons, Inc., 629 p.
- Parma, R. and Shataee, Sh. 2010. Capability study on mapping the diversity and canopy cover density in Zagros forests using ETM+ images (Case study Ghalajeh forests, Kirmanshah province). Iranian Journal of Forest, 2 (3): 231-242 (In Persian).
- Pouteau, R., Meyer, Y., Taputuarai, R. and Stoll, B. 2012. Support vector machines to map rare and endangered native plants in Pacific islands forests. Ecological Informatics, 9: 37-46.
- Rafieyan, O., Darvishsefat, A.A., Babaii, S. and Matajii, A. 2011. Evaluation of pixel-based and object-based classification methods for tree idenstification using aerial images (Case study: a forestation in Camestan-Nur). Iranian Journal of Forest, 3 (1): 35-47 (In Persian).
- Schowengerdt, R.A., 2007. Remote sensing: models and methods for image processing (3rd ed.). Academic Press, 515 p.
- Teodoro, A.C., Goncalves, H., Pais-Barbosa, J., Veloso-Gomes F. and Taveira-Pinto, F. 2010. Identification of beach features/patterns through artificial neural networks techniques using IKONOS data. ISPRS Symposium. Austria, 5-7 July, 38 (7B): 574-579.
- Vapnik, V.N. 2000. The Nature of Statistical Learning Theory (2nd ed.). Springer, 314 p.
- Vaughn, N.R., Moskal L.M. and Turnblom, E.C., 2012. Tree species detection accuracies using discrete point LiDAR and airborne waveform LiDAR. Remote Sensing, 4: 377-403.
- Zhang, H., Zhang Y. and Lin, H. 2012. A comparison study of impervious surfaces using optical and SAR remote sensing images. International Journal of Applied Earth Observation and Geoinformation, 18: 148-156.
 

فایل ها

هم رسانی

ارجاع به این مقاله

آمار