1117 
MORPHOLOGICAL ANALYSIS OF FLOOD INUNDATED REGIONS 
IN PADDY AREAS USING ALOS/PALSAR DATA AND 
ITS DISTRIBUTION ON THE GOOGLE EARTH 
DESIGN OF THE FUTURE DISASTER MANAGEMENT SYSTEM (FDMS) 
Yasuharu Yamada 3 
“National Institute for Rural Engineering, National Agriculture and Food Research Organization (NARO), 2-1-6, 
Kannondai, Tsukuba, Ibaraki 305-8609 Japan - yamaday@affrc.go.jp 
Commission IV/8, VIII/2 
KEY WORDS: Floods, SAR, Morphology, Paddy Field, Geoweb 
ABSTRACT: 
Large floods occurred at the northern part of the Central Plain of Thailand in May of 2006. The ordinary class flood appeared in 
October of 2006 in the middle stream region of the Chao Phraya river of Thailand. The ALOS satellite obtained the PALSAR radar 
data at the flood by the emergency observation requests. The PALSAR data can easily to detect the water surface because its wave 
length is L-band. But it is rather difficult in the paddy field areas since usually paddy fields have water surface in its farmland block. 
Mathematical morphology method is one of the solutions of this problem. The author carried out the geometric correction and 
morphological filtering to the PALSAR data. The result images almost coincide with the reported flood regions. The morphological 
analysis shows good ability to the discrimination to the flood water surface from ordinary paddy field’s inundated water. The ability 
of the PALSAR data recognition of the ground objects seems to be improved comparing to JERS-l/SAR. Addition to that, this 
method requires only once satellite data acquisition, light computer resources and short time to execute the task. Therefore this image 
processing procedure is suitable for the disaster early warning system. Then the next task is how to hand the result of satellite 
analysis to the decision maker or inhabitants. There are many kinds of Web based GIS platforms which become an interface with 
natural hazard stakeholders and inhabitants. The author tried to superimpose the ALOS/PALSAR data and its result of morphological 
analysis upon the Google earth image. And also the overlay procedure was done for the geo-morphological survey map for the 
prediction of flood time inundation along the Chao Phraya River in Thailand. This kind system will be a demonstration of the 
decision support tool or hazard information platform for large flood using satellite data and the integrated ISO’s GIS standard (GML). 
Furthermore its concept will be extended to the distributed and collaborated web-GIS. To put it concretely, the following thing is 
considered. There are much of agricultural social facilities such as irrigation canals, head works, drainage pumps, water gates, etc. 
And those irrigation facility control system exists in rural areas. If the GIS data and its network are made by the GML standard, an 
usual GIS, such as agricultural product traceability system, sensor network system checking irrigation facilities, can be easily 
transformed into a disaster management GIS. And the disaster management GIS will be able to be integrated into one system with the 
usual GIS used by the concerning state and local governments or irrigation districts. Addition to that, the airborne Ku-band 
side-looking radar is developing in Japan. Its resolution is very good and it can get the ground condition of the disaster regions even 
at night or cloudy weather. The combination of those equipment can be the “Future Disaster Management System (FDMS)” concept 
just like the “Future Combat Systems (FCS)” of US Army. 
1. INTRODUCTION 
PALSAR data can be easy to detect the water surface because 
its wave length is L-band. But it is rather difficult in the paddy 
areas since usually paddy fields have water surface. 
Mathematical morphology method is one of the solutions to 
this conflict (Yamada, Y., 2003; 2007ac). The author carried out 
the geometric correction to the PALSAR data which had 
obtained at the 2006-May flood and the 2006-October flood in 
the Central Plain of Thailand. The morphological analysis 
shows good ability to the discrimination of the flood water 
surface from paddy water. This mathematical morphology 
method needs very light calculation and computer hardware 
requirements. And the most important merit of this method is to 
be demanded only once satellite scene at the flooding time. The 
conventional method of detecting flooded areas is to compare 
the before and after the flood. Therefore the result of this 
morphological method is very suitable for the early warning 
system of flood disaster. The author tried some experiment for 
making the future disaster management system. 
2. MORPHOLOGICAL PROCEDURE 
2.1 Data used 
The data used in this analysis are ALOS/PALSAR Fine mode, 
level 1.5 on May 25 th and May27th, 2006 for the northern part 
of the Central Plain of Thailand; and on October 12 th , 2006 for 
the entire Chao Phraya river basin. 
2.2 Mathematical morphology procedure 
The mathematical morphology procedure is the followings. (1) 
The region under DN=2200 is considered as the water surface 
including paddy field and flood inundated areas. (2) 2 times 
morphological dilations at 3 by 3 size (3) morphological open 
operation at 12 by 12 size. The last procedure is for sieving out 
the minute polygons. 
2.3 Results 
2.3.1 Northern Central Plain of Thailand