Abstract SAR is a new technologies for earth surface observation, which are important measures for land use survey to acquiring field data. Information of crops, farm land as well as its environmental conditions can be obtained all the time under all weather conditions through microwave remote sensing and image procession. Information, such as land use type, crop type recognition, growth rate monitoring, impact assessment of drought and logging, and yield estimation, is the basis for sound scientific agricutural management.
1.Introduction
In today's world, rapid population growth, accompanied by food and resource shortages, as well as environmental degradation is threatening social and economic development. China's population is over one-forth of the total world population but its arable land per capital is much lower even than most developing countries. When 2000 comes, the population will reach 1.3 billion, meanwhile, with the development of national economy, urban and industrial land expanding, the arable land has reduced year by year. It is clear that changes in land conditions and food supply have a significant influence on sustainable development. Recently, as the progress of China's opening and reforming, the structure and area of cropping has greatly affect the national ecorcomic development of our country. So in China the land use moitoring is very important for obtaining the information about the land dynamic timely and provide scientific decision-making.
At present, land use survey and crop growth monitoring by remote sensing technology (TM, SPOT, NOAA etc.) are carried out through acquiring optical information But in the south part of China, due to the cloudy and rainy weather conditions, data acquirement by optical remote sensing technology for crop growth monitoring and land use investigation become the biggest problem. Generally, farming and land monitoring are with a high requirement for real time data. For multi-spectral remote sensing systems, such as Landsat TM which is in common use today, long coverage cycle and bad weather conditions often make them impossible to get field data. So, the land use survey and crop monitoring in the south part of China becomes more difficult.
According to the exist problems of remote sensing sensor, SAR data can promote the power for land use survey, crop type recognition and and supply a new information source for agricuture investigation.
2. Composite procession of SAR and TM
In order to appraise the potential of SAR data for landuse survey and crop recognition, two kinds of remote sensing data are used. One is Landsat data of 1997, the other is Radarsat SAR data( called as Radarsat SAR standard mode(S2)) received in 1997.
In composite procession, geometric registration of two kinds of data and noise smoothing are carried out. synthetic Aperture Radar imagery inherently has speckle due to the coherent illumination. The existence of speckle in SAR data is possibly due to two reasons. First there exists noise in the RADARSAT data. Second, the intensity of radar response from land cover may be controlled by many structural parameters; the random variation of these parameters can cause completely different radar response. Before the data is analyzed, it is necessary to remove this speckle. Although many smoothing effect is often at the sacrifice of edge information because protection edges and smoothing edges and smoothing speckle is contradictory. However, radar data by its very nature is inherently noise. High-pass and textural filtering would be subject to the creation of artifacts unless radar noise is reduced. It is important that radar noise be reduced prior to subsequent processing. so the Enhanced lee Filter is selected to smooth the speckle in this study.
Radarsat data has a characteristic for representing the object physical feature, Landsat TM has a characteristic for representing the object spectral feature. In order to extract the advantage of two remote sensing sensor data, a composite of Radar sat data and Landsat TM data was made in this research. During composite procession, channels 4, 5, 3 of Landsat TM was selected for HIS transform because main information of TM is contained in the three channels.
After transform, colorimetric variable and saturation variable(H and S) as well as the intensity variable of SAR image are re-HIS transformed:
Select channels 3, 4, 5 and SAR image for KL transform.
False color composite is performed on G of the transformed image, original data of Landsat TM Channel 4 and the first variable of the KL transform image correspondingly to R,G and B channels. Thus, the main information of SAR and TM is compounded.
