Racurs
Русский info@racurs.ru

Overview

Area of application

Process flowsheets

Detailed specifications

Examples

 NEWS  COMPANY  DEALERS  PRODUCTS  SERVICES  TRAINING  SUPPORT  RESOURCES  CONTACTS  ACCOUNT  SITE MAP 
 PHOTOMOD  PHOTOMOD GeoMosaic  PHOTOMOD UAS  PHOTOMOD Radar  PHOTOMOD Lite  PHOTOMOD Cloud 

Area of application

SAR Applications by the Example of Radarsat Data.

 

Radarsat is the first Canadian space radar system. Radarsat-1 satellite was launched in 1995, Radarsat-2 satellite is scheduled to start in 2005-2006. Side-looking synthetic aperture radars (SAR) based on Radarsat satellites allow to get the Earth surface images independently of weather conditions and light.

Some examples of RADARSAT data use were provided by Canada Center for Remote Sensing

Agriculture
Foresty
Geoscience
Hydrology
Mapping
Marine Applications
SAR stereo products applications

Agriculture

One of the first and still most prominent applications of remote sensing data is the large-scale evaluation of agricultural lands and crops. Data and information can readily be obtained on the status of crops (their health, amounts, and relative distribution), the overall volume of biomass in a region or nation, and soil moisture conditions. From these and other data, fore casts of harvest yields can be made, smoothing out considerably the economic wrinkles involved in the process of planning international purchases and sales of agricultural products, and allowing national agricultural planning to be carried out with greater confidence. Special applications include such things as the evaluation of soil erosion in a region, and evaluation of the effectiveness of agricultural engineering measures to combat erosion, flooding, drought, etc.

 

Deriving Crop Information from RADARSAT-1 Imagery ( Manitoba, Canada).

Satellite optical imagery can provide the information necessary for crop monitoring, but with the current suite of satellites it can be difficult to acquire cloud free imagery throughout the growing season.

Research studies have repeatedly demonstrated that timing of image acquisition is very important to the success of crop mapping with optical imagery. Consequently, unless optical imagery is available during key stages of crop development these data alone will not provide the information necessary for operational field-level crop monitoring. Since microwaves penetrate cloud cover, SAR acquisitions are far more reliable. Microwaves are sensitive to the structure of crops (size and geometry of the leaves, stalks and fruit) and to crop moisture levels. As a result, radar imagery could be an important component of a crop monitoring system.

Based on a data set acquired over Carman, Manitoba (Canada), a study by the Canada Centre for Remote Sensing clearly demonstrated that multi-date RADARSAT-1 imagery could provide accurate information about crop types. Timing of image acquisition was important and the greatest success at separating crop types was achieved with imagery gathered during the period of seed development. The greatest confusion was observed among the various small grain crops (wheat, barley, and oats), since the structure of these crops is very similar. As expected, larger incidence angles like those associated with RADARSAT-1's Fine Beam Modes, were required to establish crop type.

Crop structure and moisture levels change as the crop moves from one developmental stage to the next. These changes in growth stage result in variations in radar backscatter and must be considered during crop mapping. Crop condition indicators - in particular Leaf Area Index and crop height - were correlated with radar backscatter. Based on these results, RADARSAT-1 can provide some information on crop condition and within field variability.

Imagery gathered by Synthetic Aperture Radars can provide crop type and crop condition information. However, for SARs like RADARSAT-1 (C-HH) and ERS-2 (C-VV) that acquire data in only a single frequency and polarization, multi-temporal data sets are required. Future space-borne SARs such as RADARSAT-2 and Envisat ASAR will provide imagery in multiple polarizations. These sensors promise to significantly improve the amount of crop information provided in a single acquisition and will be exciting new tools for crop monitoring. CCRS is currently involved in several research campaigns to explore the potential of these future radar sensors for agriculture.

Reference:
H. McHairn, J. Ellis, J.J. van der Sanden, T. Hirose, R.J. Brown; "Providing crop information using RADARSAT-1 and optical data"; International Journal of Remote Sensing, 2000 (submitted).

Foresty

Applications of remote sensing in forestry include measurement of the forest cover, assessment of the health status of forest areas and even specific tree types, evaluation of the progress and damage wrought by forest fires, and tracking of the aftereffects of clearing operations. These data can be used to develop strategies for the efficient and ecologically sound management of forest resources.

 

RADARSAT Images Forest Management (Alberta, Canada ).

This RADARSAT image, was collected over Whitecourt, Alberta on March 5, 1996. Whitecourt is an active logging region, in the foothills of the Rocky Mountains, west of Edmonton. The dark areas on the image are clearcuts (A) which contrast quite sharply with the untouched Boreal forest. The Boreal forest in this region is comprised of mostly spruce, pine and some aspen.

In the centre of the image, there is a very steep sided valley (C). The radar highlights the topography of the valley by illuminating one slope and shadowing the other. Towards the upper left portion of the image is a feature with a number of lines (B) radiating from it. This is a pumping station, and the linears are gas pipelines running out from it in many directions. Some of the access roads are visible in the image (D), although it is difficult to discriminate between the roads and the pipelines.

Processed by RADARSAT International Inc.
Source Image © Canadian Space Agency

Geoscience

A wealth of information useful in geological investigations car be obtained via Almaz, including images detailing the structure of geological formations (folds, valleys, fractures) and the nature of volcanic activity. The data permit regions to be identified where mineral deposits are likely to be found, and allow mineral surveys to be conducted from space. One special feature of applying SAR technology to geological investigations is the ability to obtain information from areas which are undercloud cover for extended periods. Another unique aspect is the ability to obtain data on volcanic activitv and lava flows in near-real time, under any conditions of weather and illumination.

 

Geological Mapping Bathurst Island, Nunavut, Canada

This RADARSAT image was acquired on March 21, 1996, over Bathurst Island in Nunavut. The most notable feature on this image is the striking representation of the geology. The dark embayment in the middle of the image (A) is Bracebridge Inlet which joins the Arctic ocean (Viscount Melville Sound), west of the imaged area. Polar Bear Pass is the broad valley continuing from the Inlet to the east.

The geology of the island is characterized by a remarkable pattern of folds. The upper few kilometres of the layered rocks have been deformed into a series of folds, clearly visible on the RADARSAT image.

The light tones on this image (C) represent areas of limestone, while the darker areas (B) are comprised of siltstone. The contact between these two materials is sharp and easily identified on the RADARSAT image.

Processed by RADARSAT International Inc.
Source image © Canadian Space Agency

Hydrology

Snow Mapping Ottawa, Ontario, Canada.

In many areas of the world, a significant proportion of the freshwater available for consumption, irrigation and hydroelectric generation is supplied from snowpack runoff. In order to make efficient use of runoff water, resource agencies must be able to make early predications of the amount of water stored in the form of snow. Areal extent, Snow Water Equivalent (SWE) and snowpack wetness are the most important parameters to be determined in this process. Research using SAR data for snow studies has shown promising results.

Figure 1. RADARSAT (Standard Beam 7) 12 January 1996
Figure 2. RADARSAT (Standard Beam 6) 19 January 1996
Figure 3. Difference between two radar images 12 and 19 January

Ottawa RADARSAT images allow a comparison between wet and dry snow conditions and can change detection monitoring.

Mapping

Radar surveillance from space is ideal for compiling topographical maps of remote regions where access is difficult or impossible. It can also be used to improve existing maps, and to identify and measure changes caused by natural events (earthquakes, mudslides, avalanches, droughts, fires, floods, etc ), human interventions (water management practices, road building, reclamation of wetlands, dams, residential development, forest management, etc. ), and climatic changes (such as ice thawing and desertification).
Radar cartography is also useful for making precise determinations of the boundaries of fields and forest areas with other types of land use.

Extracting Roads from Stereo RADARSAT Data.

RADARSAT's many beam modes offer a variety of stereoscopic configurations of a given location that are very different in terms of geometry and radiometry. Researchers at CCRS have undertaken an exhaustive study to evaluate the geometric and radiometric parameters of the SAR system for mapping application.

The researchers generated three stereo pairs with fine mode images (F1-F5) and standard mode images (S1-S7 & S4-S7) and used them to extract planimetric features, such as roads. They separated the roads into four categories according to 1:50,000 Canadian map standards: highways (four or more hard-surface all-weather lanes), main roads (two and more hard-surface all-weather lanes), secondary roads (two and more loose or stabilized lanes), and unclassified "city streets" (see Figure 1). They used the entire data set (more than 900 km of roads) in the statistical accuracy evaluation. From the comparison of the topographic roads and the extracted roads, they computed the omission error and the circular errors with 68% and 90% confidence levels (CE68 and CE90) for each stereo-pair F1-F5, S4-S7 and S1-S7.

Ortho-rectified F5 RADARSAT-SAR fine mode image with main (in dark), secondary (in red) roads and city streets (in white) overlaid.

Marine Applications

 

Figure 1.

 

Figure 2.

 

Figure 3.

 

Figure 4.

 

Figure 5.

Figure 6.

We first consider three RADARSAT SAR images of the same ocean area, but acquired under different wind conditions and with different incidence angles. Fig. 1 is a low wind speed case (~2m/s) with a large incidence angle. Here, the ocean signal level is very low and the image is noise dominated, although some ocean features are visible. In fact, the measured radar cross section is essentially equal to the radar noise floor expressed as a noise-equivalent radar cross section. On the other hand, the vessel signatures are readily apparent. Fig. 2 also shows a low wind speed (~5m/s) scene, but with a smaller incidence angle. In this case the ocean signal is much larger. Ocean features such as internal waves are more readily apparent, as are slicks, probably of biological origin, that are advected by the local surface current. Fig. 3 again has a large incidence angle, but in this case there is an atmospheric front present offshore. The wind speed is lower to the east of the front where the signal level is again close to the noise floor. Both ocean features and ships are visible. To the west of the front, the ocean clutter has increased by nearly 8dB and the features and ships are no longer visible. Note that Figs. 1 through 3 were enhanced to optimize the contrast of the water. For the large incidence angle cases (Figs. 1 and 3 ) the land appears saturated, since there is such a large contrast between the land and water signatures at those incidence angles. The contrast is smaller for the smaller incidence angle case (Fig. 2); the land does not appear saturated.

We now apply these trade-offs to some specific marine applications:

Ship Detection. Ship detection amounts to the detection of point targets in a radar clutter background. The relative detectability of ships is shown in Fig. 4 and examples of ships can be seen in Fig. 5. Ship detection improves as the clutter level decreases. Thus, larger incidence angle beams are recommended. Lower resolution beams reduce ship detectability. S4 through S7, W3, F1 through F5, and EH1 through EH6 provide the best detectability, with ScanSAR Narrow far (SCNfar) being somewhat worse due to its larger resolution. However, SCNfar provides a useful trade-off between resolution and coverage and is recommended for surveillance of ocean-going fishing fleets (vessels longer than 50m). SCW provides useful detection for the larger incidence angle portions of its swath, but detection performance is further reduced due to its lower resolution. Caution should be used with EH beams since range ambiguities ("ghost images") sometimes appear.

Slick Detection. Natural and artificial slicks may be imaged through local suppression of ocean clutter caused by damping of short waves; an example is shown in Fig. 6. There must be adequate resolution to resolve the slick and adequate CNR to detect the local clutter suppression due to the slick. Thus, smaller incidence angles are recommended. S1 through S4 offer adequate CNR. W1, W2, or ScanSAR Narrow near (SCNnear) improve the swath coverage but at the expense of resolution. Slicks have been imaged at large incidence angles with low wind, but have a very poor signal-to-noise ratio. At higher winds, the slick material may be mixed downward into the water column, preventing detection.

Oceanic Features. Mesoscale ocean features may be imaged by their characteristic pattern of advected surfactants (Fig. 2) or due to associated shear or convergence zones. Smaller-scale features such as internal waves (Fig. 2)or surface current patterns associated with underwater topographymay be imaged. Ocean signatures depend upon the wind speed. If the wind speed is too large, then the feature becomes lost in the clutter and cannot be imaged. Thus, ocean feature detection is necessarily restricted to a small CNR regime. If the feature location is known, then S1 through S4 are recommended. If the requirement is for surveillance, then larger swath coverage is desirable; W1, W2, or SCNnear are recommended. SCW performs best for the near portion of the image swath.

Atmospheric Features. Wind speed variations over kilometre scales occur near atmospheric fronts ( Fig. 3 ) or may be associated with atmospheric flow phenomena such as gravity waves, boundary layer rolls, or low pressure systems. These often appear in SAR images as local changes in the ocean clutter level. Modes with adequate CNR and large area coverage are best. W1, SCNnear, and the near portion of SCW are suitable, although even the large incidence angle beams work well if the wind speed is high. Note that estimation of the wind speed from the image requires radiometric calibration.

Ocean Waves. For ocean wave imaging, there is an important additional consideration. The platform range-to-velocity ratio controls the (velocity bunching) mapping from the scene into the SAR image domain. The larger this parameter, the more non-linear the wave imaging becomes. This parameter has a value of approximately 115s for the ERS-1 SAR, which is considered to be too large by many researchers since it can result in significant azimuth cut-off of the imaged wave spectrum. However, this value represents a lower limit for the RADARSAT SAR. In order to keep this parameter as small as possible, we recommend S1, W1, SCNnear, and the near edge of SCW, although the image resolution of SCW is usually too large to provide useful ocean wave information.

The predictions we made in 1995 hold up in general. The better than expected performance of RADARSAT has provided a better ability to detect ocean features, particularly slicks, than was anticipated prior to launch.

SAR stereo products applications.

Currently a SAR stereo data provide a wide range of capabilities suitable for a number of applications. Modern SAR systems have unprecedented features enabling surface mapping over vast areas with moderate cost fast delivering time. Stereo DEM’s generated from SAR data have their own niche in the topography products market. One of the SAR stereo advantage is leaded from the most famous radar capability to acquire images over cloud covered regions that is critical point for number of earth sites. Very often SAR stereo products became most preferable over other remote terrain mapping techniques or digitizing of paper maps technology in sense of cost per square unit, delivering time, and final accuracy.
On the most potential capacious area for applying of SAR stereo products is a telecommunication industry. It has strong demand for digital elevation maps supplied at low price level and with sufficient accuracy. These data help to place the transmission towers for antennas network. An additional advantage is that the SAR stereo derived DEM could be combined with orthorectified image and presented as 3-D scene that do more easy the utilization of such kind information.

Subscribe to company news Ul. Yaroslavskaya, 13A, of. 15,
Moscow, Russia, 129366
Tel   +7 495 720-51-27
Fax   +7 495 720-51-28
Last modified: 18.09.2018© Racurs, 2004-2018