Home Articles Land cover mapping using remotely sensed observation

Land cover mapping using remotely sensed observation

H. S. Lim
School of Physics
Universiti Sains Malaysia
Minden 11800 Penang, Malaysia.
Tel: +604-6577888, Fax: +604-6579150
[email protected]

M. Z. MatJafri
School of Physics
Universiti Sains Malaysia
Minden 11800 Penang, Malaysia.
Tel: +604-6577888, Fax: +604-6579150
[email protected]

K. Abdullah
School of Physics
Universiti Sains Malaysia
Minden 11800 Penang, Malaysia.
Tel: +604-6577888, Fax: +604-6579150
[email protected]

N. M. Saleh
School of Physics
Universiti Sains Malaysia
Minden 11800 Penang, Malaysia.
Tel: +604-6577888, Fax: +604-6579150

C. J. Wong
School of Physics
Universiti Sains Malaysia
Minden 11800 Penang, Malaysia.
Tel: +604-6577888, Fax: +604-6579150

Sultan AlSultan
Remote Sensing Center of Envir. Counsultant
ISPRS, commission 7 WG VII/7, Middle East Coordinator
Malaz, Al Nurii St., P.O.Box.92038
Riaydh City 11653, Saudi Arabia.
Tel: +966504890977 Fax: +96614767828
[email protected]

ABSTRACT
This study presents the land cover mapping using supervised classification technique namely maximum likelihood, minimum distance-to-mean and parallelepiped. The remotely sensed data used in this study was the Landsat TM image acquired on 20 January 1999 over the Gulf of Saudi Arabia. The objective of this study was to test the feasibility of high spatial resolution Landsat TM image for land cover mapping. Training sites were selected within each scene and land cover classes were assigned to each classifier. Accuracy assessment was performed in this study to determine the quality of the land cover map. The maximum likelihood classifier produced superior result with the highest accuracy in this study. This study clearly classified the land cover features using the multispectral classification technique for urban planning and development purposes.

INTRODUCTION
The increasing availability of remote-sensing images, acquired periodically by satellite sensors on the same geographical area, makes it extremely interesting to develop the monitoring systems capable of automatically producing and regularly updating land-cover maps of the considered site (Bruzzone, et al., 2002). The objective of this study was to estimate the coverage area of the seasonal agricultural vegetation over AlQassim, Saudi Arabia using two different seasons of satellite Landsat TM images. In this study, we are using three standard supervised classification techniques for the analysis. Finally, the post classification analysis of accuracy assessment was performed based on the Kappa statistic and overall accuracy.

STUDY AREA
The study area in the Arabian Peninsula, located between latitude 12°N and 32°N and between longitude 20°E and 35°E (Figure 1). This particular geographical position gives the area a great bioclimatic diversity. The desert of the Arabian Peninsular is located as a part of the hot desert, which extends from the Sahara in Africa in the west to the Thar Desert in Indo-Pakistan sub-continent in the east.


Figure 1. The study area in the central of Saudi Arabia

DATA ANALYSIS AND RESULTS
Two high spatial resolution satellite Landsat TM scenes were chosen in this study for seasonal agricultural vegetation analysis on 28 February 1994 and 27 November 1994. The aim of the classification analysis is to categorize all of the pixels in the imageries into two classes; vegetation and non-vegetation. Basically, the process can be divided into three simple steps, the pre-processing, data classification and output. In the pre-processing, the classes were established by using polygons for training sites. They are delineated by spectrally homogeneous sub areas, which have, class name given. In the classification stage, three supervised classification methods were selected to classify the images. Maximum Likelihood, Minimum Distance-to-Mean, and Parallelepiped were applied in the present study. Two methods of accuracy assessment used in this study were the Kappa statistic and overall accuracy. The Kappa statistic is a statistical method of assessing the accuracy that takes into account the chance of random agreement. This statistic has been used by many researchers in their studies [Selamat, et al., (2002), Dymond and Johnson, (2002)]. The produced results in this study are shown in Table 1 and the accuracy assessment results are shown in Table 2 and 3. Finally, the coverage of seasonal agricultural vegetation was determined. The classified seasonal agricultural vegetation maps are shown in Figure 2 for 28-2-1994 and Figure 3 for 27-11-1994.

Table 1. Statistic analysis for the increasing vegetation and urban areas

Classes 28-2-1994(km2) 27-11-1994 (km2)
Vegetation 13.0023 11.7712
Non- Vegetation 100.9557 102.1868
Total 113.9580 113.9580

Table 2. The Kappa coefficient for the two images

Classification method Kappa coefficient
28-2-1994 27-11-1994
Maximum Likelihood 0.8651 0.9102
Minimum Distance-to-Mean 0.7152 0.7925
Parallelepiped 0.5961 0.6584

Table 3. The overall classification accuracy for the two images

Classification method Overall classification accuracy (%)
28-2-1994 27-11-1994
Maximum Likelihood 86.2154 90.2561
Minimum Distance-to-Mean 75.2152 78.2151
Parallelepiped 54.0210 61.3653


Figure 2. The seasonal agricultural vegetation map: 28-2-1994 [Colour Code: Green = vegetation and Blue = Non-vegetation]


Figure 3. The seasonal agricultural vegetation map 27-11-1994 [Colour Code: Green = vegetation and Blue = non-vegetation]

CONCLUSION

This analysis has demonstrated the necessity of a spatial approach in studying seasonal agricultural vegetation over AlQassim, Saudi Arabia. The Maximum Likelihood classifier produced high degree of accuracy. This study classified two seasonal agricultural vegetation coverage maps with a reasonable accuracy.

ACKNOWLEDGEMENTS:

We would like to thank the technical staff who participated in this project. Thanks are extended to USM for support and encouragement.

REFERENCES:

  • Bruzzone, L., Cossu, R. and Vernazza, G. (2002). Combining parametric and non-parametric algorithms for a partially unsupervised classification of multitemporal remote-sensing images. Information Fusion 3, 289 –297.
  • Dymond, C. C. and Johnson, E. A., 2002, Mapping vegetation spatial patterns from modeled water, temperature and solar radiation gradients. Journal of Photogrammetry and Remote Sensing, 57, 69–85.
  • Rees, W.G., Williams, M. and Vitebsky, P., 2003, Mapping land cover change in a reindeer herding area of the Russian Arctic using Landsat TM and ETM+ imagery and indigenous knowledge. Remote Sensing of Environment, 85, 441–452.
  • Selamat, I., Nordin, L., Hamdan, N., Mohti, A. and Halid, M., 2002, Evaluation of TiungSAt data for land cover/use mapping application. Proceeding of the Seminar Kumpulan Pengguna TiungSAT-1, Jabatan Remote Sensing dan Sains Geoinformasi, Fakulti Kejuruteraan Dan Sains Geoinformasi Universiti Teknologi Malaysia and Astronautic Technology (M) Sdn Bhd.