Impacts of spatial resolution on land cover classification

Chanida Suwanprasit and Naiyana Srichai

Prince of Songkla University Phuket Campus

APAN 33rd Meeting 13-17 February 2012

Outline Introduction Objective Methodology Results Conclusions

2/20

Spatial Resolutionis a measurement of the spatial detail in an image, which is a function of the design of the sensor and its operating altitude above the Earth’s surface (Smith, 2012).

3/20

Classification Factors Number of mixed Pixel Number of ROIs Scale or spatial resolution Spectral resolution Temporal resolution

Objective To examine effects of pixel size on

land use classification in Kathu district,

Phuket, Thailand

5/20

Study area: Kathu, Phuket 7/20

Kathu

Kamala

Patong

Imagery Source Resolution (m) Band Spectral Type

LANDSAT 5 TM

30 1 (Blue) 0.45 – 0.52 m

30 2 (Green) 0.52 – 0.60 m

30 3(Red) 0.63 – 0.69 m

30 4 (NIR) 0.78 – 0.90 m

30 5 (NIR) 1.55 – 1.75 m

60 6 (TIR) 10.40 – 12.5 m

30 7(MIR) 2.80 – 2.35 m

THEOS

15 1 (Blue) 0.45 -0.52 m

15 2 (Green) 0.53 – 0.60 m

15 3 (Red) 0.62 – 0.69 m

15 4 (NIR) 0.77 – 0.90 m

Data set specification6/20

Band 1 (Blue) Band 2 (Green) Band 3 (Red)

Band 4 (NIR) Band 5 (NIR) Band 7 (MIR)

Landsat 5 Spectral Bands10/20

Band 1 (Red) Band 2 (Green)

Band 3 (Blue) Band 4 (NIR)

THEOS Spectral Bands11/20

True Color

THEOS Landsat 5

9/20

RGB (4,3,2)

THEOS Landsat 5

13/20

Process Overview

THEOS Landsat 5

Classes• Forest• Built-up• Road• Water• Agriculture• Grassland• Bare land

UnsupervisedK-Mean

SupervisedSVMs

Training area

Test area

Control points

THEOS LandSat 5

Land use Classification Map

Data Set

12/20

THEOS Landsat 5

Unsupervised Classification:K-Mean (7 Classes) 14/20

Support Vector Machines : SVMs

THEOS Landsat

Forest

Grassland

Bare land

Water

Built - up

Road

16/20

Class Confusion Matrix

Class

THEOS Landsat-5

Prod. Acc. (%)

User Acc. (%)

Prod. Acc. (%)

User Acc. (%)

Forest 97.47 96.81 100.00 100.00

Built-up 62.37 71.18 97.02 97.57

Road 74.89 64.62 90.15 90.59

Water 99.87 99.29 83.25 78.71

Bare land 76.78 91.31 60.88 66.78

Grassland 89.49 95.23 96.02 91.85

Agriculture 92.21 84.22 76.69 75.37

Overall Accuracy 90.65% (Kappa Co.= 0.88) 89.00% (Kappa Co.=0.87)

17/20

Conclusion THEOS gave a higher classification accuracy than Landsat 5

for identifying land use in this study. More Spectral bands from Landsat 5 with 30m is not appropriated for

selecting clearly ROIs than THEOS with 15m resolution. The better resolution image greatly reduce the mixed-pixel problem, and there is

the potential to extract much more detailed information on land-use/land cover structures.

18/20

References Duveiller, G. and P. Defourny (2010). "A conceptual

framework to define the spatial resolution requirements for agricultural monitoring using remote sensing." Remote Sensing of Environment 114(11): 2637-2650.

Randall B. Smith (2012). "Introduction to Remote Sensing Environment (RSE)". Website: http://www.microimages.com.

19/20

Acknowledgement Faculty of Technology and Environment

Prince of Songkla University, Phuket Campus

Geo-Informatics and Space Technology Development Agency (Public Organization)

UniNet

20/20

Thank you for your kind attention