Radar Backscattering of Vegetation from ERS-1 SAR
Data and a Radiative Canopy Model
Lau Yok King, Mohd. Ibrahim Seeni Mohd. And Samudin Ahmad
Centre for Remote Sensing,
Faculty of Geoinformation Science and Engineering
Universit Teknologi Malaysia,
Locked Bag 791, 80990 Johor Bahru, Malaysia
Phone : +(6 07) 550 2880 Fax : +(6 07) 556 6163
E-mail : mism@fksg.utm.my
Abstract
Satellite remote sensing using the visible and the infrared wavelengths has been used successfully in various applications related to earth resource studies and monitoring of the environment. However, the use of radar remote sensing has many advantages in comparison with optical remote sensing techniques. By far the most important factor is the virtual insensitivity of radar to atmospheric conditions. This allows the regular collection of site observations independent of cloud cover or time of overpass. On the other hand, the interpretation of radar imagery over land area is not as straightforward as that of the more commonly used visible and infrared remote sensors. This study was carried out in the framework of EC/ASCEAN ERS-1 Radar Remote Sensing Project.
The objective of the study is to derive backscatter values from three main classes of vegetation, namely, matured rubber and oil palm plantations as well as paddy in the study area of Kuala Terengganu from ERS-1 multi-temporal data to analyze if it is possible to differentiate between the vegetation from the data. Backscatter simulations have been carried out by using the University of Texas at Arlington (UTA) Radioative Canopy Model in order to make comparisons with the ERS-1 data and to improve physical understanding of the backscatter properties of the vegetation classes under study. From the results, it can be concluded that the differentiation of matured oil palm and matured rubber is difficult even with multi-temporal data. But, in the case of paddy it is possible to differentiate it from other vegetation with multi-temporal data due to the difference in backscatter values during the growing cycle.
Introduction
Satellite remote sensing using the visible and the infrared wavelengths have been used very successfully in various applications related to earth resources studies and monitoring of the environment. However, it is not very suitable in the tropics because of the cloud cover problem. Radar remote sensing can overcome the problem although it may not possess some of the advantages of sensing in the optical wavelengths.
On the other hand, the interpretation of radar imagery over land is not as straightforward as that of the more commonly used visible and infrared remote sensors. Usually, special image processing techniques must be applied on the radar imagery to make it more readily interpretable. Furthermore, the interpretation of the backscattering process that underlies the radar image formation must be well understood with respect to the physical characteristics of he targets under observation and the specifies of the radar instrument.
This paper reports on some of the results obtained in the Terengganu study area (Figure 1) of the MAL-2 EC-ASEAN ERS-1 Radar Remote Sensing Project.
Figure 1: Location map of study area.
Backscatter from ERS-1 SAR DATA
The digital values for the 3 dominant vegetation i.e. matured rubber and oil palm as well as paddy were extracted at various locations. A mean value was obtained and converted to dB values. These values were obtained from the SAR data acquired on 8 different date. Figure 2 shows the graph of dB values for these vegetation types obtained from the ERS-1 SAR data acquired during the eight different dates.
Figure 2: Backscatter values for paddy, rubber and oil palm
extracted from the multi-data ERS-1 data.
