Canadian Firm Provides Turnkey Solution for Indonesian Coastline Mapping Project By Kevin P. Corbley RADARSAT International (RSI) recently completed the data acquisition phase of the largest and most ambitious coastal zone mapping project ever undertaken in Indonesia. This project has proven as noteworthy for its enormous coverage area as for its illustration of how dramatically large remote sensing and GIS mapping programs have changed in Southeast Asia over the past few years. "It relates directly to the international monetary crisis that has hit especially hard in Asia," said Shawn Burns, RSI's Sales director for Asia Pacific. In better economic times, governments purchased the various products and services necessary for major resource management projects individually-such as satellite images and image processing software-and integrated them on their own. But now, integrated information solutions are more often purchased as a more efficient use of money. "This project is an example of how RSI and other remote sensing companies have had to change the way business is conducted," said Burns. "The remote sensing industry can no longer expect to provide data alone; we have to offer complete turnkey solutions." The project in question is the Digital Marine Resource Mapping (DMRM) program, initiated by the government of Indonesia in 1996. Aside from providing synthetic aperture radar from the Canadian RADARSAT satellite, RSI brought in image-processing software and a project team that trained Indonesian technicians in the latest methods of creating GIS maps from radar imagery. The results have included the successful completion of a coastal management GIS, one that could preserve millions of dollars in Indonesian natural resources over the next century. Also, the data processing techniques that were developed can be duplicated and applied toward coastline mapping around the world. Accuracy is Critical The need for effective coastal mapping assumed great significance in 1982, when the United Nations Convention on the Law of the Sea (UNCLOS) ratified the concept of Exclusive Economic Zones (EEZ). This law gave island nations rights of sovereignty to explore and exploit the resources of water, the seafloor, and subsoil within their EEZ, extending out 200 nautical miles from shore. No country understands the consequences-and complexity-of accurately delineating coastlines more fully than does Indonesia, an archipelago of 17,000 islands stretching more than 3,000 kilometers across the Pacific and Indian Oceans. Its 81,000-km shoreline is the longest and most circuitous of any nation in the world. With rights at stake to abundant fisheries, potential hydrocarbon riches, and other resources, Indonesia began the DMRM to map the land-water boundary along their entire coastline. To ensure compliance with UNCLOS 82, the Indonesian Government's National Agency for Coordination of Surveys and Mapping (BAKOSURTANAL) entered into a contract with the Norwegian survey company Blom ASA in 1985, to incorporate state-of-the-art surveying, satellite imaging, bathymetric mapping, and GIS technologies into the project. Coastline mapping lies at the heart of DMRM and has proven to be one of its most challenging aspects. This phase is crucial because the shoreline map will serve as the point of beginning for numerous other mapping and survey programs. In UNCLOS terms, the land-water interface at the lowest astronomical tide is the baseline from which the extent of the EEZ is measured. BAKOSURTANAL contracted Blom Dantarsa Co. (Jakarta) to perform the critical shoreline-mapping phase of the DMRM. Blom Dantarsa is a cooperative venture of two companies with extensive experience in surveying coastal marine environments -Blom ASA (Norway), and PT. Dantarsa Gatra (Indonesia). Acquiring Imagery through Clouds "Originally we planned to use optical satellite imagery to map the shoreline," said Atle Lagestrand, data processing manager at Blom Dantarsa. "However, Indonesia's location along the equator, and the region's typically unstable weather conditions, make optical remote sensing difficult." As in most equatorial regions around the world, Indonesia is often obscured by clouds, which cannot be penetrated by optical imaging sensors. Flying below the clouds in a survey aircraft was not an option because the project area was too large for the economical use of airborne cameras. To overcome these perverse weather and technical conditions, Lagestrand contacted RSI in Richmond, B.C., Canada for assistance. RSI is the worldwide distributor of RADARSAT imagery, providing data to mapping projects ever since that satellite was launched in 1995. RADARSAT's SAR sensor was chosen because it has the ability to collect imagery under all weather conditions, and it accurately delineates land-water boundaries-two critical components of the project. Numerous specifications made image collection difficult. First, the project schedule allowed for no more than six months to complete the acquisitions. Second, and most significantly, image collection had to be done at or near the level of high tide. This meant that the satellite could not necessarily acquire data every single time it passed over the project area. "Coastlines on hydrographic charts are defined at mean high water," explained Lagestrand. RSI calculated that 136 images were needed to cover Blom Dantarsa's project area over the archipelago. Even if clouds were not a factor, optical satellite sensors could not have collected that many high-tide scenes in the time period permitted. "The RADARSAT feature that made it possible to collect so much imagery in a short time period was the satellite's adjustable incidence angle," said Yves Crevier, coordinator of project management and technical support for RSI. "It gives RADARSAT a more frequent revisit capability than any other satellite operating at a comparable level of accuracy on the ground." Blom Dantarsa and RSI personnel worked together in finding times when high tides coincided with satellite overpasses throughout the archipelago. Fine-tuning the Collection Plan A key consideration for any satellite-mapping project is in weighing the various benefits of spatial resolution, swath coverage, and cost against one another. Such was the case in Indonesia. A huge project area required mapping, but funds were limited. RADARSAT offered several options. The satellite's SAR sensor can operate at any one of several beam modes, which translates into seven different combinations of area coverage and spatial resolutions. Scene size can vary from 50x50 km to 500x500 km, with resolution ranging from eight meters to 100 meters. The cost per square kilometer of area covered drops substantially as swath increases and spatial detail declines. "We balanced the available funding with the map scale requirements of the project. The appropriate map scale was dictated by the scale of other spatial data already gathered and stored in the project GIS," said Burns. The participants selected RADARSAT Wide and Fine beam mode imagery. The Wide beam mode scenes offer a moderate 30-meter spatial resolution in 150-km-wide swaths, which are effective at producing 1:100,000 to 1:200,000-scale maps. Fine beam mode imagery offers eight-meter resolution in roughly 50-km-wide swaths. RADARSAT collected 95 percent of the necessary images between July 1998 and January 1999. All images were acquired with incidence angles between 35 and 50 degrees. Earlier RADARSAT projects had demonstrated these were ideal settings to facilitate identification of land-water boundaries in virtually any coastal landscape-rocky, sandy, or escarpment. Processing the Imagery Blom Dantarsa personnel were experienced users of MapInfo and Intergraph software, but had limited expertise in processing SAR satellite images. RSI recognized the need to provide the DMRM group with a turnkey solution to its coastline mapping challenge, rather than just deliver raw satellite imagery. "We created our own project team to provide Blom with the shoreline information in the GIS file formats needed for direct input into their mapping database," said Burns. "Our goal was to train Blom personnel in the image processing procedure so they could continue that phase of the project and keep the database updated in the future." The RSI team devised a processing scheme that would quickly extract the land-water interface lines from the RADARSAT images and export them as suitable GIS files. "RSI employees traveled to Jakarta and trained five members of the Blom staff to perform heads-up digitizing," said Jamie Gairns, a former PCI applications specialist who works full time at RSI headquarters. PCI is the developer of the EASI/PACE image processing software on which the training took place. Heads-up digitizing is a manual process, whereby the system operator views the RADARSAT image on the screen and uses a computer mouse to trace the cursor along the edge of the on-screen coastline. The EASI/PACE software automatically records the traced line as a GIS vector and saves it as a MapInfo or ArcView file for export to the GIS. During this training, Indonesian technicians used ImageWorks, an easy-to-use heads-up image manipultion module within EASI/PACE. "This project has emphasized how important coastal zone mapping is everywhere in the world," said Gairns. "At PCI, we will be able to transfer the processing techniques developed in Jakarta to create semi-automated techniques that will be released as a coastal zone management software module for use with EASI/PACE." Completing the Database By March 1999, about 95 percent of the imagery for the project had been acquired and had been processed into GIS-compatible vectors at the Project Production Center in Jakarta. The coastline base map is now being combined with numerous other bathymetric, topographic, and hydrographic data layers in a National Marine Resources database. The GIS database will provide the baseline maps that are mandated by UNCLOS and required by Indonesia to accurately establish its Exclusive Economic Zone, Territorial Sea, and Contiguous Zone. Additionally, BAKOSURTANAL will make some GIS data available over an intranet for use by other Indonesian agencies, enabling their integration into ongoing navigation, fishing, oil exploration, and environmental monitoring programs. "Providing DMRM participants with a combination of imagery, on-site consultation, image processing training, and new software made this mapping project a success," said Burns. "We are adopting this turnkey team approach for other remote sensing programs worldwide." About the Author: Kevin Corbley is the principal in Corbley Communications. He is located in Denver, Colo., and may be reached at 303-979-3232, or via e-mail at [email protected].
Back
|