Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran

Volume: 66 Number: 1 January 1, 2016
EN TR

Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran

Abstract

Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran

Abstract: Due to rapid population growth over recent decades, changes of urban areas have significantly impacted the environment. Urban is a heterogeneous and highly fragmented environment which has made them a challenging area for remote sensing imagery. The reliability of the information delivered by remote sensing applications in urban area highly depends on the quality of spatial and spectral data. Accordingly, the objective of this study is to analyze the impact of incorporation of Hyperion imagery and textural characteristics of high resolution panchromatic ALI imagery in classifying of urban region of south west of Tehran. To this end, we extracted textural information from panchromatic ALI imagery using gray-level co-occurrence matrix (GLCM) method. Classification was carried out by SVM method in five scenarios: Classification of spectral band of CNT method, classification of spectral bands plus texture with window size 3, size 5, size 7 and size 9. The classification results show that the urban areas of south west of Tehran are insufficiently characterized by the Hyperion satellite imagery. A quantitative assessment of the results demonstrated that the use of texture information improved urban land covers classification. As a result, combining of texture information with Hyperion imagery decreases class confusion specifically in heterogonous classes. The GLCM features show great potential for land use cover classification in heterogeneous areas with rich textural information.

Keywords: Hyperspectral imagery, image texture, GLCM, remote sensing, SVM classification.

İran Tahran şehri güney bölgesinde kent alanlarının sınıflandırılmasında SVM yöntemi ile hiperspektral görüntü ve tekstür bilgilerinin birlikte kullanılması

Özet: Son yıllarda hızlı nüfus artışı ve kentsel alanlardaki değişimler çevreyi önemli bir şekilde etkilemiştir. Kentsel alanlar heterojenik ve parçalanmış bir yapıya sahiptir, bu durum uzaktan algılama görüntüleri açısından zorlu bir durum yaratmaktadır. Kentsel alanlarda uzaktan algılama uygulamalarından elde edilen bilgilerin güvenilirliği  mekansal ve spektral verilerin kalitesine bağlı olarak değişmektedir. Dolayısıyla, bu çalışmanın amacı Tahran'ın güney batısındaki kentsel bölgede Hyperion görüntüleri ve yüksek çözünürlüklü pankromatik ALI görüntülerinin dokusal özelliklerinin esas etkisini analiz etmektir. Bu amaçla, gri-seviyeli eş-oluşum matrisi (gray-level co-occurrence matrix) (GLCM) yöntemini kullanarak pankromatik ALI görüntülerinden yapısal bilgi ayıklanmıştır. Sınıflandırma beş senaryo halinde SVM yöntemi ile gerçekleştirilmiştir: CNT yöntemiyle spektral bantların sınıflandırılması, spektral bant sınıflandırılması pencere boyutu 3, boyut 5, boyut 7 ve boyut 9. Sınıflandırma sonuçları Tahran güney batı kentsel alanların  Hyperion uydu görüntüleri ile yeterince karakterize edilemediğini göstermektedir. Sonuçlar yapısal bilgilerin kullanımı ile kentsel arazi sınıflandırmalarının daha başarılı bir şekilde yapılabildiğini göstermektedir. Sonuç olarak, Hyperion görüntüleri ile yapısal bilgilerinin birleştirilmesi heterojenik sınıflandırmada karışıklığı azaltmaktadır. GLCM özellikleri içerdikleri zengin yapısal bilgi ile heterojen alanlarda arazi kullanım sınıflandırmaları için büyük bir potansiyel gösterirler.

Anahtar Kelimeler: Hiperspektral görüntü, görüntü tekstürü, GLCM, uzaktan algılama, SVM sınıflandırması

Received (Geliş tarihi): 09 February 2015 - Revised (Düzeltme tarihi): 19 February 2015 -   Accepted (Kabul tarihi): 19 February 2015

To cite this article: Yazdi, A.M.,  Eisavi, V., Shahsavari, A., 2016. Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran. Journal of the Faculty of Forestry Istanbul University 66(1): 90-103. DOI: 10.17099/jffiu.01280

Keywords

References

  1. Bokoye, A.I., Dionne, P., 2004. Urban material characterization from the Hyperion hyperspectral imager: Application to downtown Montreal (Quebec, Canada). Image and Signal Processing for Remote Sensing 5238: 569- 574. Conference No.9, Barcelona, Spain, Doi: http://dx.doi.org/10.1117/12.511150
  2. Camps-Valls, G., Bruzzone, L., 2005. Kernel-based methods for Hyperspectral image classification, IEEE Transactions on Geoscience Remote Sensing 43: 1351–1362.
  3. Carleer, A. P., Wolff, E, 2006. Urban land covermulti-level region-based classification of VHR data by selecting relevant features. International Journal of Remote Sensing 27: 1035−1051.
  4. Carlson, T.N., Sanchez-Azofeifa, G.A. 1999. Satellite remote sensing of land use changes in and around San José, Costa Rica. Remote Sensing of Environment 70: 247–256.
  5. Cavalli, M.R., Fusilli, L., Pascucci, S., Pignatti, S., Federico Santini, F., 2008. Hyperspectral Sensor Data Capability for Retrieving Complex Urban Land Cover in Comparison with Multispectral Data: Venice City Case Study (Italy). Sensors 8: 3299-3320. Doi: http://dx.doi.org/10.3390/s8053299
  6. Clausi, D. A., Yue, B., 2004. Comparing co-occurrence probabilities and Markov random fields for texture analysis of SAR sea ice imagery. IEEE Transaction on Geoscience and Remote Sensing 42: 215-228.
  7. Colgan, M. S., Baldeck, C. A., Féret, J. B., 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(11): 3462-3480. Doi: http://dx.doi.org/10.3390/rs4113462
  8. Féret, J., Asner, G. P., 2013. Tree species discrimination in tropical forests using airborne imaging spectroscopy. IEEE Transactions on Geoscience Remote Sensing 51: 73-84.

Details

Primary Language

English

Subjects

-

Journal Section

-

Publication Date

January 1, 2016

Submission Date

February 9, 2015

Acceptance Date

-

Published in Issue

Year 2016 Volume: 66 Number: 1

APA
Maleknezhad Yazdi, A., Eisavi, V., & Shahsavari, A. (2016). Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran. Journal of the Faculty of Forestry Istanbul University, 66(1), 90-103. https://doi.org/10.17099/jffiu.01280
AMA
1.Maleknezhad Yazdi A, Eisavi V, Shahsavari A. Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran. J FAC FOR ISTANBUL U. 2016;66(1):90-103. doi:10.17099/jffiu.01280
Chicago
Maleknezhad Yazdi, Ahmad, Vahid Eisavi, and Ali Shahsavari. 2016. “Incorporation of Hyperspectral Imagery and Texture Information in a SVM Method for Classifying Urban Area of Southern Regions of Tehran, Iran”. Journal of the Faculty of Forestry Istanbul University 66 (1): 90-103. https://doi.org/10.17099/jffiu.01280.
EndNote
Maleknezhad Yazdi A, Eisavi V, Shahsavari A (January 1, 2016) Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran. Journal of the Faculty of Forestry Istanbul University 66 1 90–103.
IEEE
[1]A. Maleknezhad Yazdi, V. Eisavi, and A. Shahsavari, “Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran”, J FAC FOR ISTANBUL U, vol. 66, no. 1, pp. 90–103, Jan. 2016, doi: 10.17099/jffiu.01280.
ISNAD
Maleknezhad Yazdi, Ahmad - Eisavi, Vahid - Shahsavari, Ali. “Incorporation of Hyperspectral Imagery and Texture Information in a SVM Method for Classifying Urban Area of Southern Regions of Tehran, Iran”. Journal of the Faculty of Forestry Istanbul University 66/1 (January 1, 2016): 90-103. https://doi.org/10.17099/jffiu.01280.
JAMA
1.Maleknezhad Yazdi A, Eisavi V, Shahsavari A. Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran. J FAC FOR ISTANBUL U. 2016;66:90–103.
MLA
Maleknezhad Yazdi, Ahmad, et al. “Incorporation of Hyperspectral Imagery and Texture Information in a SVM Method for Classifying Urban Area of Southern Regions of Tehran, Iran”. Journal of the Faculty of Forestry Istanbul University, vol. 66, no. 1, Jan. 2016, pp. 90-103, doi:10.17099/jffiu.01280.
Vancouver
1.Ahmad Maleknezhad Yazdi, Vahid Eisavi, Ali Shahsavari. Incorporation of hyperspectral imagery and texture information in a SVM method for classifying urban area of southern regions of Tehran, Iran. J FAC FOR ISTANBUL U. 2016 Jan. 1;66(1):90-103. doi:10.17099/jffiu.01280