The Amazon attracts world attention because of its biodiversity, its mineral resources and areas as yet considered untouched. However, the human eye, photographic cameras and even optical sensors on board of aircraft or satellites have enormous difficulty to explore it, mainly the variation in altitude of the ground. Covered almost all of the time by clouds, an integral part of its climate, and by smoke from the frequent fires, the region is as yet difficult to map. Nevertheless, the sky began to become clearer when some two years ago a company in the town of São José dos Campos, Geoambiente, detected in this difficulty a good opportunity to advance in the development of systems for improving the capture of images, mainly those of the Amazon soils.
The company created a project that was financed through FAPESP’s Small Business Innovation Research (PIPE) program, which proposed the creation of skills in order to extract from the orbital images of the Radarsat-1 satellite material for putting together updated and more precise topography maps. With a scale of 1:250,000 (each centimeter on the paper is equivalent to 2,500 meters on the ground), many maps that portrait the Amazon basin do not contain altimeter data, which would allow for an understanding of the topography of the surface.
The Radarsat-1, launched in 1995, was developed by the Canadian Space Agency (CSA), as the first commercial system for monitoring natural resources and environmental changes on the planet. The satellite is equipped with a microwave image sensor called Synthetic Aperture Radar – SAR, which allows it to “see” under adverse climatic conditions and also during the night. In the same manner it captures images from different observation angles, through which, when put together by computer, make the data of the altitude of the observed areas possible, or that is to say, maps in 3-dimensions, with details of the elevation of the ground.
The imaging radar works like a photographic camera with a flash, which provides its own illumination for the scene. It manages to do this by emitting radio microwaves and capturing the image in the same region of the electromagnetic spectrum (wavelengths of radiation between 1 centimeter and 1 meter). In this manner the Radarsat captures images with a high degree of high space resolution (8 m x 8m), better than those captured by the multi-spectral optical sensors in the visible spectrum and the infrared spectrum of other satellites such as Thematic Mapper (TM) Landsat, launched by NASA, which produces images of the earth at the spatial resolution of 30m x 30m and has been used to estimate the amount of deforestation in the Amazon basin.
Even with a lower spatial resolution, the TM-Landsat images are also used in the Geoambiente project. “They are used in conjunction with those of Radarsat-1 in order to identify roads, settlements, vegetation and agricultural activities on the ground”, explains the geologist Waldir Renato Paradella, a researcher with the National Institute for Space Research (INPE) and the project’s coordinator.
Radarsat’s advantage is that it operates on a wavelength of greater length than the optical remote sensors. In this manner it is not affected by adverse atmospheric conditions. It also has the versatility of gathering images on thirty five different modes of the same region, thus widening the options in acquiring information on the ground, as well as possessing visor systems with stereoscopic pairs (a 3-dimensional perception) used in the extraction of altitudes and the obtaining of level curves for topographic maps.
The powerful imaging radars of the Integrated Systems of Vigilance and Protection of the Amazon (Sivam/Sipam) have also managed to break through the Amazon basin’s cloud cover and capture the topography in greater detail than Radarsat-1, but there are limits to their use. Since the sensors are based on aircraft, the costs are much higher. Furthermore, a large amount of training still needs to be carried out in the country in order to master the technology of transforming data into useful information.
According to Paradella, qualifying the technicians in Brazil is one of the greatest benefits of the Geoambiente proposal. This will make a reduction in the cost of maps obtained through the images of Radarsat-1 possible, which up until now have been put together using technology developed abroad, as well as meeting the demand in the use of this technology and to the formation of professionals by Sipam. “We’re also already negotiating with the Geographic Services Board, an organ of the Brazilian army, towards the possibility of developing joint projects in cartographic application with data from Radarsat-1 for the Amazon basin”, the president of Geoambiente, the geologist Izabel Cecarelli, informs.
Begun two years ago, the project will receive in total around R$ 230,000. The investment has already born its first fruit: the cartographic map of the National Forest of Tapajós, an area of 1,940 km2 in Belterra, in the state of Pará. In making the map, the technicians firstly processed inside the computer two Radarsat-1 images, in tones of gray, obtained at different angles, which made the modeling in 3-dimensions of the topography possible. Using this process, a relative Digital Elevation Model (DEM) was obtained, which allowed for the generation of altimetry data for maps on the scale of 1:250,000. However, the absolute DEM, from which one can extract altimetry data for cartography on the scale of 1:100,000 (each centimeter on the map represents 1,000 meters on land) such as that of Tapajós, is only obtained after extensive field work. The greater the scale, the greater the detail that can be observed on the map.
In the case of Tapajós, a team of three technicians and an officer from the Brazilian army – also a partner in the project – penetrated into the forest for ten days covering more than one thousand kilometers in search of ten previously chosen points. At each point reached a Global Positioning System (GSP) receptor was installed, that provides measurements of the precise latitude, longitude and altitude of the location.
Using this data the computer models the topography of all of the area, allowing for the creation of the cartographic map with altimetry and planimetry data of details such as road network, hydrographic data, urban occupation and native vegetation. “Since access is very often impossible and the costs are high, we visited fifty of the one hundred selected points. With some of the measured field points we could generate the DEM exactly and the others validated the model”, Paradella explains. In September the field work for the second phase of the project began, which will be completed in July of next year and will result in an unprecedented topographical map of the central part of the Carajás Range, also in Pará.
The results already obtained indicate that the scale of 1:100,000 is the best possible for cartography through the high resolution images of Radarsat-1 for the Amazon basin. However, in less than two years, after the launch of Radarsat-2, with an improvement in spatial resolution (3 m x 3 m), this scale could reach 1:50,000 or better, getting close to that obtained by way of the optical sensors of other satellites. In the opinion of Paradella, the more detailed information and the availability of DEM, is opening up the way for Brazil to carry out much greater surveys of the Amazon basin.
But there is still a long way to go. “The fight now is to spread the technology of radar imaging”, Izabel says. Although the radar sensor has now been available for some years, it is still little known in Brazil, including with some government agencies, which got to the point of soliciting cartographic surveys of Amazon basin regions based on optical sensors. Even so, the company has already begun to sell integrated products, results of the fusion of radar images and geophysical data, relative to the Amazon region. In this market for the past seven years, Geoambiente until now has dedicated itself exclusively to image processing with optical sensors, a promising segment in a country where the shortage of information has reached a critical point.
The very growth of the company itself demonstrates this situation: last year it had an income of aroundR$ 3 million, 50% more than that of 2000, and this year it should again have a repeat performance. “We should grow more than 50% in 2002”, forecasts Izabel Cecarelli. The mastery of the Radarsat-1 technology is certainly going to contribute towards further developing its business.
Generation of Digital Elevation Models by way of Radargrammetry with Radarsat-1 Satellite Images (nº 99/06271-7); Modality Small Business Innovation Research (PIPE) program; Coordinator Waldir Renato Paradella – Inpe/Geoambiente; Investment R$ 129,118.00 and US$ 30,126.18