Abstract
It is well known that rice is an important world crop, particularly in Asia. The population of Asia accounts for approximately 60% of the global population, about 92% of the world's rice production, and 90% of global rice consumption. With such a large population and high levels of rice consumption, an effective rice crop monitoring tool is needed. The unique capabilities of SAR (synthetic aperture radar) to penetrate clouds, haze and darkness allows for efficient and timely data collection which is useful for applications such as rice crop monitoring. Recent research has proven RADAR's capabilities to discriminate between rice and non-rice areas and has demonstrated great potential for more in-depth research of rice crops. RADARSAT, Canada's earth observation SAR satellite, offers several beam position allowing for frequent repeat coverage opportunities. This is important for monitoring highly dynamic and relatively quick growing crops such as rice. For this paper, seven RADARSAT images were acquired over a study area in Guangdong Province in southern china. The full rice crop growing season in this part of the world is 3 months. The acquired data set spans the full growing season with a more frequent acquisition schedule for the peak of the growth paddies in the area, and relate the results to the various growth stages of the plant-from flooded paddies, to full grown plants, to harvested areas.
1.0 Introduction and Background
The monitoring of the earth's surface with satellite imagery has, in the past, been largely dependent on the weather and time of day. Today, microwave sensors such as Canada's RADARSAT, are broadening the world of weather conditions. There are many applications of remote sensing imagery which can benefit from particularly dynamic phenomena such as crop monitoring (Brown et al., 1993; 1995). Rice is a very important plant because it provides a staple food for a large part of the world's population, especially in Asia. In some regions, it is also important economically as an export product. Since most rice producing regions, it is also important economically as an export product. Since most rice producing regions are found in tropical and sub-tropical regions with high amounts of rainfall and therefore cloud cover, the traditional approach of using vegetation indices in the visible and infra-red (VIR) region are not practical. The fragmented land cover and small size of many of the rice paddies also precludes the use of a low-resolution high
repeat coverage sensor like NOAA's AVHRR which can often use a temporal composting technique to overcome the persistent cloud cover problem. Fortunately, the presence of water beneath the plant canopy, typical of paddy rice production systems, enhances the sensitivity of radar backscatter to rice (Le Toan et al., 1989; Aschbacher and Paudyal, 1993; Kurosu et al., 1993; Brisco et al., 1996; Le Toan et al., 1997). This is due to the domination of the backscatter process by canopy scattering and water surface-canopy interaction terms, because no direct surface contribution comes from the underlying water surface. Consequently, RADARSAT in particular and Synthetic Aperture Radar (SAR) satellite systems in general, are very promising tools fro developing a rice mapping and monitoring capability.
In most of China, the crop calendar for rice spans approximately three months. Figure 1 graphically demonstrates the 3 major stages of this growth cycle. The cycle begins with flooded rice paddies, into which young rice are transplanted early in the season. RADAR response at this stage is dark due to the small amount of scattering. The crop then mature and reach their peak growth stage around mid-cycle. The RADAR response from the growing rice plants increases as the plants grows in height and volume, and increase in moisture content. After the plants have fully matured, the moisture levels decrease, senescing begins, and the plants are harvested. There are wide varieties of harvesting techniques used, which can alter the spectral response at this stage in the cycle.
Figure 1: Rice Growth Curve
The purpose of this study is to assess radar's capability for detecting this growth curve based on the radar backscatter from rice targets over a full growth cycle.
