SEA GRASS MAPPING FROM SATELLITE DATA

Transcription

1 JSPS National Coordinators Meeting, Coastal Marine Science May 2008 Melaka SEA GRASS MAPPING FROM SATELLITE DATA Mohd Ibrahim Seeni Mohd, Nurul Hazrina Idris, Samsudin Ahmad

2 1. Introduction PRESENTATION OUTLINE 2. Objectives of Study 3. Study of Sea Grass Features from Satellite Data 4. Results 5. Concluding Remarks

3 INTRODUCTION Mapping of sea grass is important to fishing industry and ocean science studies. Remote sensing satellites provide large area coverage and a range of temporal scale which allow the parameters to be studied continuously. Previous study used the AVNIR-2 (Advanced Visible and Near Infrared Radiometer type 2) data from ALOS Satellite for sea grass mapping.

4 OBJECTIVES To extract the sea grass features from LANDSAT TM satellite data. To map the sea bottom features in the coastal waters of Sibu Island, Malaysia.

5 LANDSAT TM SATELLITE CHARACTERISTICS The data used was acquired on November 25, Altitude Orbit Inclination Repeat coverage Approximately 705 km Polar, sun-synchronous days

6 Swath width Spatial resolution Wavelength 185 km (at nadir) 30 m / 120 m band 1: µm (visible blue) band 2: µm (visible green) band 3: µm (visible red) band 4: µm (near infrared) band 5: µm (near infrared) band 6: µm (thermal) band 7: µm (infrared)

7 STUDY AREA

8 LANDSAT TM DATA PROCESSING The technique for extracting bottom- type information depends upon the fact that bottom-reflected radiance is approximately a linear function of the bottom reflectance and an exponential function of the water depth. Thus, the measured radiance are transformed according to the following equation (Lyzenga( Lyzenga,, 1981),

9 X i = Ln (L i L si ) X j = Ln (L j L sj ) where, L i = measured radiances in band i L si = deep-water radiances in band i L j = measured radiances in band j L sj = deep-water radiances in band j

10 If X i is plotted versus X j and water depth varied, the data points will fall along a straight line whose slope is K i / K j where K i and K j is the attenuation coefficient of water in band i and band j, respectively. If the bottom reflectance is changed, the data points will fall along a parallel line which is displaced from the first. By measuring the amount of displacement, a change in bottom reflectance can be detected even if the water depth is unknown.

11 The amount of displacement is given by, Y i = [ K j ln (L i L si ) K i ln (L j L sj )] ( K 2 i + K 2 j ) 1/2

12 The technique used for extracting bottom- type features combines the information in band 1 and band 3 of the satellite data. This procedure was implemented on the LANDSAT TM data by calculating the variable Y i at each point in the scene and using this variables as a depth-invariant index of the bottom type. The depth invariant index was density sliced into three sea bottom types, namely sea grass, coarse sand and fine sand.

13 Graf X i vs X j X y = x X1

14 RAW LANDSAT TM IMAGE Band combination (RGB): 3, 2, 1 respectively.

15 Band combination (RGB): 3, 2, 1 respectively.

16 DEPTH INVARIANT INDEX IMAGE

17 Depth invariant index Features Seagrass Fine sand Course sand Depth invariant index

18 SEA GRASS DISTRIBUTION LEGEND Sea Grass Fine Sand Course Sand

19 CONCLUDING REMARKS In this study, three bottom-type type features have been found surrounding Sibu Island i.e. seagrass,, fine sand and course sand. This result needs to be verified by ground truth observation and multitemporal LANDSAT TM data need to be used to analyze the capability of LANDSAT data for sea grass studies.

20 ACKNOWLEDGEMENTS We would like to thank Prof. T. Yanagi of Kyushu University, Japan and the Japan Society for Promotion of Science (JSPS) for making this study possible.