05-045
8/17/06
1:35 PM
Page 1073
A User-customized Web-based Delivery System of Hypertemporal Remote Sensing Datasets for Australasia Michael Schmidt, Edward A. King, and Tim R. McVicar
Abstract Long time series of well-calibrated and consistent daily remote sensing data are important for studies of intra- and inter-annual environmental behavior. These data are used to support environmental management, and in most locations are the only historical spatial dataset. The Web-CATS (CSIRO AVHRR Time Series) system provides access to the Australasian Advanced Very High Resolution Radiometer (AVHRR) data archive using the World Wide Web. The data archive consists of multiple daily satellite overpasses from several receiving stations in Australia since July 1981. The data are held online enabling the use of state-of-the-art algorithms to generate on-demand user-customized products. This novel design for operational and dynamic remote sensing data product generation enables Web-CATS users to browse the entire metadata-database and produce consistent time series information from on-line data in near real time. As these algorithms improve, users have the ability to easily re-process their dataset(s).
Introduction Daily remotely sensed data for terrestrial, marine and atmospheric applications have now been routinely collected for more than two decades by the Advanced Very High Resolution Radiometer (AVHRR) sensor on board the NOAA polar orbiting series of satellites. These data, usually coupled with forward modeling in the spectral, spatial, or temporal domain (or some combination), provide the means to extract information. From this information, knowledge in various application domains is then confirmed, refined, or generated. This can be performed using either requirement-driven, or data mining approaches (Hinke et al., 2000; Zhou, 2003). As the amount of data, information and knowledge in every subject increases, improved and novel mechanisms to access and analyze each are necessary (Zhou et al., 2001; Kelmelis et al., 2003). Working within a data-informationknowledge hierarchy (Hicks et al., 2002; Gunnlaugsdottir, 2003), it is apparent that daily satellite measurements acquired over the last 20 years need management to allow
Michael Schmidt is with CSIRO Marine and Atmospheric Research, GPO 3023, Canberra 2601, ACT, Australia (
[email protected]). Edward A. King is with CSIRO Marine and Atmospheric Research, GPO Box 3023, Canberra 2601, ACT, Australia. Tim R. McVicar is with CSIRO Land and Water and the eWater Cooperative Research Centre, GPO Box 1666, Canberra 2601, ACT, Australia. PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
the derivation of consistent geophysical properties. Data intensive geosciences, including remote sensing, are prime candidates for exploiting synergies arising from the rapid development in information technologies, and in particular communication technologies (Xue et al., 2002; Homer et al., 2004). For these large remotely sensed datasets, the need for metadata storage built within a data warehouses construct is especially necessary (Xue et al., 2002; Overton and Wragg, 2003). Due to an increasing number of satellite missions, usually recording data in more bands with improved sensors, the data volume and potential information to be extracted from these data has also increased (Ogrosky, 2003; Petiteville et al., 2004). One such communication technology is the World Wide Web (WWW), which has become a natural avenue to propagate and distribute data, products and applications in order to reach a wide group of users (non-experts as well as experts) with specialization in various disciplines (Bugg et al., 2002; Carrara et al., 2003). In remote sensing, complex data integration processing steps are usually necessary before the data can be used in applications. For example, with direct satellite broadcast and reception with real time data processing, (geospatial) database ingestion, calibration, navigation, and geolocation of the data are performed. Given a successful implementation of these basic steps, further product generation, usually within application-specific domains, or refinement of the data can be performed (for further reading see Cracknell, 1997; King, 2003; Homer et al., 2004). Previously, users of remote sensing data were mostly provided with limited base processing applied to the data, which they then processed with their own algorithms or software packages, with only “in-house” defined standards. Consequently, no data consistency was achieved by different users, even though all worked with the same base data. Existing web-based satellite data providers can be roughly categorized into four major groups: (a) non-commercial sites, such as the NOAA Satellite Active Archive (http://www. class.noaa.gov), where certain data are accessible free of charge (including AVHRR data with users having no ability to select processing options); (b) commercial sites, such as Eurimage (http://www.eurimage.com), that sell certain satellite data and derived products; (c) distributors of both commercial and non commercial data, such as the EarthExplorer
Photogrammetric Engineering & Remote Sensing Vol. 72, No. 9, September 2006, pp. 1073–1080. 0099-1112/06/7209–1073/$3.00/0 © 2006 American Society for Photogrammetry and Remote Sensing S e p t e m b e r 2 0 0 6 1073
05-045
8/17/06
1:35 PM
Page 1074
run by the U.S. Geological Survey (http://edcsns17.cr.usgs. gov/EarthExplorer/); and (d) institutions that provide users with specific products/information derived from satellite images, such as the Australian Bureau of Meteorology (http://www.bom.gov.au). While online data archive search engines are widely in use, examples of dynamic, web-based processing systems mainly focus on the generation of onlinemaps and downloadable vector information, such as Bugg et al. (2002) and Davidson and Miglarese (2003). Providing on-line access to geospatial data with a raster file format is less common (due to the large data volumes); however there are several examples. Alameh (2004) describe a web-based prototype for raster image re-projection, and furthermore, the NASA funded initiative GeoBrain allows the inclusion of various geospatial data archives including remote sensing data, which can be dynamically processed to a set of user customized product levels (Di, 2004). This flexible and modular system follows the concept of the Open GIS Consortium (OGC) and thus allows the integration of various services, but due to its complexity is still a prototype. None of the systems introduced above have the capacity to perform on-demand user-customized processing of long time series, large data volume datasets. This paper describes the development of a fully automatic operational Web-based software system: Web-CATS (CSIRO AVHRR Time Series). Web-CATS allows users to: (a) gain access to the more than 20 years time-series of daily AVHRR data for the Australasian region; and (b) explicitly specify their processing. The system has been generically designed enabling it to be easily applied to data from other locations and sensors: for example, Moderate Resolution Image Spectrometer (MODIS) or the forthcoming National Polar-orbiting Operational Environmental Satellite System (NPOESS) (Townshend and Justice, 2002). The key to our approach is the dynamic data processing and product generation from base data stored on-line; all algorithms and ancillary data used by the algorithms are fully documented allowing new observations to be processed consistently with the previous data. User requests for data processing and product generation are performed on-demand with a computer cluster using the state-of-the-art algorithms implemented in Common AVHRR Processing Software (CAPS) (see Turner et al., 1998; Turner and Davies, 2000). This process has only become possible with recent advances in information technologies such as increased computer processing power and data storage capacities. Coupled with progress in networking, image processing, process automation, and database management, the opportunities for data/information analysis and exchange have widened rapidly. Making the time series of daily satellite data readily available, utilizing Web access, opens many avenues in several application areas for further operational development and scientific research (e.g., McVicar and Jupp, 2002; Davidson and Miglarese, 2003; McVicar et al., 2003; Lu et al., 2003). Web-CATS overcomes a previously daunting impediment to implementing a data-information-knowledge hierarchy.
Data and Databases
Figure 1. Typical daily coverage for one satellite in one direction, in this case for NOAA-14 ascending (afternoon) passes on 04 January 1996; the equivalent area is typically observed four times each day (see text).
variously acquired at receiving stations located at Aspendale, Hobart, Townsville and Perth (see Figure 1). During this period GAC data are used to supplement HRPT data for periods and locations when HRPT data are not available. Since 1992 all HRPT data from Alice Springs, Darwin, Hobart, Townsville, and Perth have been comprehensively archived, initially as part of the Australian contribution to the “Global 1 km project” (Eidenshink and Faundeen, 1994). Data from the different stations have been combined by stitching the different segments from each station to eliminate redundancy and produce a single best-quality HRPT scene for each overpass (King, 2000); only the stitched HRPT scenes are held online on the cluster (King, 2003). Approximately 50 million km2 of Australasia is recorded daily, including the entire Australian land surface, New Zealand, East Timor, Papua New Guinea, Singapore, most of Indonesia, parts of Malaysia and the Philippines, and the surrounding ocean to at least 2000 km from the Australian Coast (see Figure 1). Since 1992, coverage of this area has been obtained at least four times daily as both the day and night overpasses of satellites in both the morning and afternoon orbital planes are recorded (King, 2003). There are periods when up to 24 HRPT acquisitions have been recorded in one day, as all data from all operating NOAA satellites are acquired and archived by the receiving stations. During the more than two decades that data have been acquired there have been numerous changes at the reception stations as a result of hardware replacement(s) and/or operational changes. These are reflected in the variety of data formats, media formats, and media types present in the raw archive (see King (2003) for further details). At the time of this writing, the stitched archive at the CMAR is 7 TB, (nearly 100,000 overpasses) and is stored on-line in Australian Satellite Data Archive (ASDA) format (Turner et al., 1998). When new data are ingested into the on-line storage, the metadata for each overpass are consistently and automatically stored in a relational metadata-database (see Schmidt et al. (2004) for full details).
Data data, acquired daily for the entire Australian continent since July 1981, have been compiled at the CSIRO Marine and Atmospheric Research (CMAR) in Canberra. From 1981 to 1986 data are the lower resolution Global Area Coverage (GAC) (Cracknell, 1997) obtained from NOAA. From 1986 to 1992, higher resolution High Resolution Picture Transmission (HRPT) (Cracknell, 1997) AVHRR data was AVHRR
1074 S e p t e m b e r 2 0 0 6
Databases Storing and organizing large volumes of remote sensing data is challenging (Lewin and Spehe, 1996). Historically searching a satellite data archive was a time-consuming and expensive task as data were stored on tapes; with only a few sites able to access the data using robot-controlled media management systems. In contrast, due to the rapidly PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
05-045
8/17/06
1:36 PM
Page 1075
decreasing price of computing hardware, the complete AVHRR archive of the Australasian region is held online on a LINUX cluster of personal computers (PCs) (see King et al. (2004) for details). The large number of AVHRR data files requires a data warehousing concept in which a relational database holds the information and history of each scene, and the satellite data are stored on the networked file system of the local disks of each PC within the cluster. The database entries can be accessed using Standard Query Language (SQL) queries (see Bowman et al., 1996); a Web interface has been developed to instigate and update previous data searches (see Figure 4). The tables of the relational database used for the AVHRR data warehouse were designed to be time and space efficient. The database scheme has been normalized to the greatest extent possible; consistent with the need for extensibility and flexibility. Database normalization is a series of steps to obtain a database design that allows for consistent storage and efficient access of data in a relational database. These steps reduce data redundancy and the chances of data becoming inconsistent (see Bowman et al., 1996). The core of the database is one large table, which contains a single record with a unique identifier (primary key) for each scene in the archive. Pointers to other tables (foreign keys) describing characteristics common to many scenes (e.g., spacecraft and receiving station) are also included. Several other tables recording file locations and log files refer to the scene table. The complete MYSQL database configuration is described in Schmidt et al. (2004).
Setup and WWW Interface The combination of Hyper Text Markup Language (HTML), Hypertext Preprocessor (PHP), MYSQL relational database, and Apache Web server permits easy user-access to the satellite overpass metadata-database using a Web browser. Database requests are submitted from a remote computer by an authorized user through a Web interface generated with PHP and HTML. The result of the user request is handled by PHP and sent back in HTML to the client computer for display; a schematic diagram of the process is shown in Figure 2. The Web functionality was achieved by utilizing object oriented programming facilities within PHP. The steps through the Web-pages that lead to the processed data products are described in the following subsections and are graphically displayed in Figure 3.
Figure 2. Software and hardware architecture of WebCATS to allow user defined, customized data and product generation.
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
Figure 3. Structural overview of the interacting Webpages within Web-CATS. The rectangles show the main steps of the data and product selection process; the containers labeled MYSQL represent the databases: AVHRR with 17 tables, and Web-CATS with two tables in which information about the users, job-names and job-versions are stored.
Login and Registration Users are required to register before accessing Web-CATS. Registration enables individual users to be identified, and for user-specific data to be stored for use during subsequent visits. Each time a user connects to the Web interface, they identify themselves with a username and password. As they then proceed through a sequence of Web pages choosing their data and customizing their processing options, they are tracked by means of a session cookie stored in their browser. When initiating a data search and processing request the user may provide a name for the request (job-name) which is stored with the request parameters in the database together with information on how many times this job-name has been used (i.e., job-version). This permits re-use of the parameter specification in subsequent searches (see Figure 3); for example, enabling the same products to be regenerated after more data become available or using the same processing parameters to produce a similar product at a number of different locations or times. If a new job-name is provided, the user is presented with a spatio-temporal search page. Spatio-temporal Search After registration, the user is guided to the product selection page or asked to specify parameters for a new spatio-temporal search. In the spatio-temporal search page, shown in Figure 4, the user is asked to define their geographic region of interest (ROI) and their temporal search window. On the same page the user has the option to search for all available data (default) or to specify which particular satellites (currently: NOAA-6, 7, 8, 9, 10, 11, 12, 14, 15, 16, 17), and/or overpass direction (ascending or descending orbit), and/or if data acquired in a specific time-interval of the day are to be included. The internal query searches first for matching single overpasses in one SQL query for time, satellite, and overpass direction. For each single overpass satisfying the first search, then the S e p t e m b e r 2 0 0 6 1075
05-045
8/17/06
1:36 PM
Page 1076
Figure 4. Web interface of the spatio-temporal satellite data search. (A color version is available on the ASPRS Web Site: www.asprs.org).
geographical search routines check for overlap with the userspecified ROI. The geographical search was programmed within PHP and is thus not dependent on supplementary software packages including a (potentially slow) geographical information system for polygon intersection testing. All bounding polygons (computed from the satellite ephemeris data) for the scenes are stored within the metadata-database. The search algorithm tests for intersection of the line segments from the bounding polygon of each scene with those from the ROI. A test is also performed to check whether one polygon completely encloses the other (without boundary intersection). The algorithm moves over to the next scene as soon as a positive intersection is found (see Schmidt et al., 2004 for further details). If a ROI is specified, only data for the ROI is processed. Data Selection The overpasses that pass all tests (temporal, satellite overpass, and spatial) are displayed in the users’ browser 1076 S e p t e m b e r 2 0 0 6
together with a list of relevant metadata including a quicklook image and footprint. Figure 5 shows an example of a search output. The user is given the option of inspecting the search results and interactively excluding single passes from the data used for subsequent product generation. The footprintimages are generated as a raster image overlaid with vector data of the spatial extent of the satellite coverage (currently a polygon consisting of eight points) and displayed by the users’ browser. Product Selection Having selected a set of data to process, the user must also specify the processing required to generate the desired product(s) and give the projection system in which the data should be delivered in. The currently supported data projection options are: satellite view, geographic (latitude/longitude), cylindrical equidistant, Mercator, South polar equidistant, and South polar stereographic. A range of different PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
05-045
8/17/06
1:36 PM
Page 1077
Figure 5. Example of a single search result of a database query. (A color version is available on the ASPRS Web Site: www.asprs.org).
products and processing strategies have been implemented to enable as wide a choice as possible. As extreme examples, a user requesting Normalized Difference Vegetation Index (NDVI) data (see Tucker, 1979) could: (a) choose a single satellite overpass in satellite projection with no atmospheric correction; or (b) a 15-year time series of atmospherically and geometrically corrected data in the Mercator projection composited using the maximum value algorithm over 10-day intervals (see Cracknell, 1997 for further reading). The Web-CATS interface and processing system is flexible enough to meet both these and a range of other requirements (Schmidt et al., 2004). Although for many products several options are implemented, in order to allow non-expert users to obtain a standard product, a default option is provided with a set of recommended processing parameters chosen a priori. In addition to the default values, PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
links to online help pages and other background information and documentation are provided. Products derived from the AVHRR data are divided into direct products derived from single bands and synthesized products, which are a combination of single channels and/ or processed data, possibly including auxiliary data (as in the case of atmospherically-corrected data). Table 1 lists the currently available direct products for each raw data channel. Per-pixel solar illumination and satellite view angles are computed from satellite ephemeris data. In addition, the acquisition time of each pixel, particularly useful for thermal studies (see McVicar and Jupp, 1999) can be provided. These three parameters can be exported as optional data layers in the final product. Table 2 shows the synthesized products, where BRDF is the bi-directional reflectance S e p t e m b e r 2 0 0 6 1077
05-045
8/17/06
1:36 PM
Page 1078
TABLE 1: DIRECT DATA PRODUCTS Product
PER AVHRR
SPECTRAL CHANNEL
Ch 1
Ch 2
Ch 3a
Ch 3/3b
Ch 4
Ch 5
X X
X
X X
X X X
X X X
X X X
X X
X X
X X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Digital Number Radiance Brightness temperature Reflectance Normalized reflectance Satellite view angles* Solar illumination angles* Observation time (per pixel)*
*same value for each spectral channel.
TABLE 2: LIST
OF
SYNTHESIZED PRODUCTS (SEE TEXT
FOR
EXPLANATION)
Synthesized Product Vegetation indices (NDVI, etc.) Other indices (NDSI, etc.) Albedo Surface temperature BRDF corrected reflectance Cloud mask Atmospheric correction
the search and product generation parameters for possible re-use.
Backend Processing and Data Delivery The processing software, CAPS, was developed by CSIRO to promote the production of standard AVHRR data products by incorporating best practice methods (Turner et al., 1998; Turner and Davies, 2000). Within CAPS, processing routines are called in a command line or script mode, which allows the list of data files, with the corresponding set of processing parameters, to be sent to a script generating a CAPS processing file. The processing job is then queued into a list of jobs, that are processed in parallel by the LINUX cluster (see King et al., 2004; Schmidt et al., 2004 for further details) When the product generation is complete, the user is sent an email with individual login and password. Their data can then be accessed using file transfer protocol (FTP) and remains on the FTP server for seven days before removal. A file documenting the data selection and processing parameters accompanies the data products. The data are exported in hierarchical data format (HDF), which can be imported into most modern image processing software packages.
Discussion
distribution function, (Roujean et al., 1992); and NDSI is the normalized difference snow index (Romanov et al., 2000). Product Specification For many of the products (both direct and synthesized) there are often a number of choices for algorithms and ancillary data to be used in the processing; an up-to-date list can be obtained at Web-CATS (http://www.eoc.csiro.au/cats). General options for all products are projection, output pixel-size (500 m), data re-sampling method (nearest neighbor or bilinear interpolation). For time series processing the user has the option to choose between different compositing methods: minimum/maximum/mean value composites, minimum view zenith angle composite (MVZAC), and compositing intervals (Table 3). Maximum, minimum, and mean value compositing are described in Cracknell (1997); for the MVZAC, see Lovell et al. (2003). Parameter Check and Logout The final set of product generation parameters is shown in a check-out page with an option to either accept these parameters (invoking the cluster to generate the products), or to alter the setup. Finally the user-approved list of input scenes is recorded in the Web-CATS database together with
Due to recent developments in information technology, the creation of a WWW-based data delivery service providing dynamic and standardized data processing (and product generation) from a large data volume stored on-line is now achievable. The huge advances in Web technologies and the commodification of Web application environments permits the development of sophisticated interfaces that greatly ease user-interaction with data sets. The more advanced Web scripting languages (such as PHP) enable flexible and modular development of these interfaces and greatly facilitate rapid extension to incorporate new functionality. For the first time for the Australasian region AVHRR data from different receiving stations are brought together. The data were substantially improved in quality by stitching the whole archive to generate the best possible data coverage in time and space (King, 2000). Data consistency, dynamic product generation with state-of-the-art algorithms (as implemented in CAPS) can now be applied to this daily, long time-series, continental dataset. The preprocessing is, from an application user perspective, outsourced so that no additional investment in software, hardware, or base processing remote sensing expertise is necessary. These three components usually contribute substantially to the cost of projects. This “outsourcing” (and centralization) of the basic processing has several other advantages:
• • •
TABLE 3:
LIST
OF
COMPOSITING INTERVALS
• Compositing Interval 1 . . . n days (n100) Tri-monthly (10 day) Bi-monthly (15 day) Monthly 2 month 3 month* *for sub-continental search regions only.
1078 S e p t e m b e r 2 0 0 6
• • •
Users need not be concerned with storage media or data format issues; The need for navigation or calibration coefficient files, which would normally hamper the product generation and cause project delays, are dealt with once for all users; Updating of these files (involving download and installation), an activity that is at best tangential to data analysis, is no longer of concern to users; State-of-the-art algorithms are implemented in CAPS as standard routines with strict version control being placed on the pre-processing modules; Due to scientific development these algorithms may improve and users can easily reprocess their required data products; Time series data over a more than 20-year period from the same satellite mission are held in one system; and Every dataset is produced with a common set of methods and can be reproduced identically because of the user management and associated file and processing records held in the user database. PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
05-045
8/17/06
1:36 PM
Page 1079
A potential drawback from the user point of view might be the dependence on an external agency to provide this service. However, by specifying and requesting implementation of new routines and contributing anomaly reports, users can actively participate in improvement and evolution of Web-CATS to meet a wider variety of needs for a large group of potential users. Even so, as new functionality is incorporated into CAPS, it becomes available for all users who wish to do their own processing: executable versions of CAPS are freely available, and can be downloaded by any user. The Web modules can be easily moved to other platforms as the Web-CATS program parts are linked relatively with only one configuration file that needs updating if the pages move to other locations or are used for other satellite sensors.
Conclusions Technical advances in information technology have permitted the establishment of a LINUX cluster and development of a Web-based data and product generation service. A major benefit of Web-CATS is that the longest existing time series of satellite data of the Australasian region can be accessed in near real-time by relatively inexperienced remote sensing users. Information derived from data products, such as drought maps or fire maps can be used to direct decision making. Additionally, Web access to this consistently processed, user-customized database of over 20 years of daily reflective and thermal data provides many opportunities for application-specific research. Web-CATS caters to expert users by providing rich functionality and a wide range of processing choices. However, the needs of non-expert users who simply require standard products are also met through the provision of default parameters. Regardless of the level of knowledge, help buttons and documentation pages throughout the data selection process are provided to assist the user. The list of products implemented in the Web-CATS is under constant update as improvements and new product routines are implemented in CAPS. The Webpages and database architecture are largely transferable to other Earth observation satellite data such as MODIS or the forthcoming NPOESS. Web-CATS is located at the following URL: http://www.eoc.csiro.au/cats. Numerous requests from the Australian environmental remote sensing community have been used to beta test Web-CATS, it will soon be fully operational after more tests have been performed. In the meantime several popular products, such as a NDVI time series processed in a pre-defined manner, will be produced and be available from Web-CATS. The list of pre-processed products, as well as the list of processing options, can and will be expanded as Web-CATS becomes more sophisticated. Due to a common data export format (HDF), Web-CATS can be incorporated in the OGC network.
Acknowledgments This project is funded by contributions from the CSIRO Earth System Science Post-Doctoral Fellowship program, CSIRO Marine and Atmospheric Research and CSIRO Land and Water. Thanks for initial motivational comments for the Web-CATS concept from Professor Henry Nix (“there are oceans of data, rivers of information, trickles of knowledge, and the odd drop of wisdom”), Centre for Resource and Environmental Studies, Australian National University, and Ken Brook, Queensland Department of Natural Resources. Jenny Lovell and Peter Briggs (both from CSIRO Marine and Atmospheric Research) provided comments on an earlier draft of this manuscript. We would like to kindly thank all those who have helped, supported and encouraged us to do this work. The CSIRO AVHRR time series, and the best practice PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
tools to process that data are the result of many years work by past and current colleagues within CSIRO, including: Ian Barton, Peter Briggs, Susan Campbell, Steven Clift, Harvey Davies, Mac Dilley, Mary Edwards, Dean Graetz, Ian Grant, David Jupp, Jeff Kingwell, Jenny Lovell, Ross Mitchell, Denis O’Brien, Alan Pearce, Fred Prata, Chris Rathbone, Michael Raupach, Paul Tildesley, Peter Turner and Murray Wilson. Thanks also to the anonymous reviewers for their comments that improved our paper.
References Alameh, N., 2004. A raster image re-projection web service prototype, Photogrammetric Engineering & Remote Sensing, 70(5): 635–642. Bowman, J., S.L. Emerson, and M. Darnovsky, 1996. The Practical SQL Handbook, Addison-Wesley, Third Edition, 454 p. Bugg, A.L., R.D. Spencer, and A. Lee, 2002. Applying GIS for developing regional forest agreements in Australia, Photogrammetric Engineering & Remote Sensing, 68(3):241–249. Carrara, P., G. Fresta, and A. Rampini, 2003. BANCO: an SVG-based approach to create web sites for the management of remote sensing, spatial and non-spatial data, International Journal of Remote Sensing, 24(20):3903–3916. Cracknell, A.P., 1997. The Advanced Very High Resolution Radiometer, Taylor and Francis, London. Davidson, M.A., and A.H. Miglarese, 2003. Digital coast and The National Map: A seamless cooperative, Photogrammetric Engineering & Remote Sensing, 69(10):1127–1131. Di, L., 2004. Geobrain – A web services based geospatial knowledge building system, E-Proceedings of NASA Earth Science Technology Conference, 22–24 June, Palo Alto, California, 8 p. Eidenshink, J.C., and J.L. Faundeen, 1994. The 1 km AVHRR global land data set – 1st stages in implementation, International Journal of Remote Sensing, 15(17):3443–3462. Gunnlaugsdottir, J., 2003. Seek and you will find, share and you will benefit: Organising knowledge using groupware systems, International Journal of Information Management, 23(5): 363–380. Hicks, B.J., S.J. Culley, R.D. Allen, and G. Mullineux, 2002. A framework for the requirements of capturing, storing and reusing information and knowledge in engineering design, International Journal of Information Management, 22(4): 263–280. Hinke, T.H., J. Rushing, H. Ranganath, and S.J. Graves, 2000. Techniques and experience in mining remotely sensed satellite data, Artificial Intelligence Review, 14(6):503–531. Homer, C., C.Q. Huang, L.M. Yang, B. Wylie, and M. Coan, 2004. Development of a 2001 National Land-Cover Database for the United States, Photogrammetric Engineering & Remote Sensing, 70(7):829–840. Kelmelis, J.A., M.L. DeMulder, C.E. Ogrosky, N.J.V. Driel, and B.J. Ryan, 2003. The National Map - From geography to mapping and back again, Photogrammetric Engineering & Remote Sensing, 69(10):1109–1118. King, E.A., 2000. Stitching the Australian 1-km archive (T.R. McVicar, editor), Proceedings of the 10th Australasian Remote Sensing and Photogrammetry Conference, 25 August, Adelaide, Australia, CSIRO, Canberra, Australia, pp. 41–50, URL: http://www.clw.csiro.au/publications/technical2000/tr36–00.pdf (last date accessed: 15 June 2006). King, E.A., 2003. The Australian AVHRR Data Set at CSIRO/EOC: Origins, Processes, Holdings and Prospects, CSIRO CAR/EOC Report 2003/04, Canberra, Australia. 48 p., URL: http://www. eoc.csiro.au/tech_reps/2003/tr2003_04.pdf (last date accessed: 15 June 2006). King, E.A., M. Schmidt, and T.R. McVicar, 2004. A cost-effective linux cluster for efficient processing of time series continental wide remotely sensed data sets, Proceedings of the 12th Australasian Remote Sensing and Photogrammetry Conference,
S e p t e m b e r 2 0 0 6 1079
05-045
8/17/06
1:36 PM
Page 1080
Fremantle, Australia, Australasian Photogrammetry and Remote Sensing Society. Lewin, T., and K. Spehe, 1996. Information management in federal agencies: The neglected piece of the IRM puzzle, Government Information Quarterly, 13(1):51–63. Lovell, J.L., R.D. Graetz, and E.A. King, 2003. Compositing AVHRR data for the Australian continent: Seeking best practice, Canadian Journal of Remote Sensing, 29(6):770-782. Lu, H., M.R. Raupach, T.R. McVicar, and D.J. Barrett, 2003. Decomposition of vegetation cover into woody and herbaceous components using AVHRR NDVI time series, Remote Sensing of Environment, 86:1–18. McVicar, T.R., P.R. Briggs, E.A. King, and M.R. Raupach, 2003. A review of predictive modelling from a natural resource management perspective: The role of remote sensing of the terrestrial environment, CSIRO Land and Water Client Report to the Bureau of Rural Sciences, Canberra, Australia, 26 p., URL: http://www.clw.csiro.au/publications/consultancy/2003/ Review_of_Remote_Sensing.pdf (last date accessed: 15 June 2006). McVicar, T.R., and D.L.B. Jupp, 1999. Estimating one-time-of-day meteorological data from standard daily data as inputs to thermal remote sensing based energy balance models, Agricultural and Forest Meteorology, 96(4):219–238. McVicar, T.R., and D.L. Jupp, 2002. Using covariates to spatially interpolate moisture availability in the Murray-Darling Basin: A novel use of remotely sensed data, Remote Sensing of Environment, 79:199–212. Ogrosky, C.E. 2003. National mapping examined: An introduction, Photogrammetric Engineering & Remote Sensing, 69(10): 1081–1085. Overton, I., and B. Wragg, 2003. Web-based searching and cataloguing for map collections and the map trade industry, Proceedings of Spatial Science Conference: Spatial Knowledge without Boundaries, 22–26 September, CSIRO, Canberra, Australia. Petiteville, I., C. Caspar, V. Dessard, and O. Barois, 2004. Web mapping techniques applied to the visualization of ESA satellite products, Acta Astronautica, 54:177–181. Romanov, P., G. Gutman, and I. Csiszar, 2000. Automated monitoring of snow cover over North America using multi-
1080 S e p t e m b e r 2 0 0 6
spectral satellite data, Journal of Applied Meteorology, 39:1866–1880. Roujean, J.-L., M. Leroy, and P.-Y. Deschamps, 1992. A bidirectional reflectance model of the Earth’s surface for the correction of remote sensing data, Journal of Geophysical Research, 97(D18): 20,455–20,468. Schmidt, M., E.A. King, and T.R. McVicar, 2004. Development of a Web-based data and product delivery system for the CSIRO AVHRR Time Series (CATS): Architecture, interface and processing descriptions, CSIRO Earth Observation Centre Report 2004/05, 40 p., URL: http://www.eoc.csiro.au.tech_reps/2004/tr2004_05.pdf (last date accessed: 15 June 2006). Townshend, J.R.G., and C.O. Justice, 2002. Towards operational monitoring of terrestrial systems by moderate-resolution remote sensing, Remote Sensing of Environment, 83:351–359. Tucker, C.J., 1979. Red and photographic infrared linear combinations for monitoring Vegetation, Remote Sensing of Environment, 8:127–150. Turner, P.J., and H.L. Davies, 2000. Advances in CAPS (T.R. McVicar, editor), Proceedings of the 10th Australasian Remote Sensing and Photogrammetry Conference, 25 August, Adelaide, Australia, CSIRO, Canberra, pp. 71–80. Turner, P.J., H.L. Davies, P.C. Tildesley, and C.E. Rathbone, 1998. Common AVHRR processing software (CAPS) (T.R. McVicar, editor), Proceedings of the Land AVHRR Workshop, 9th Australasian Remote Sensing and Photogrammetry Conference, 24 July, Sydney, Australia, CSIRO, Canberra, pp. 51–58. Xue, Y., A. Cracknell, and H. Guo, 2002. Telegeoprocessing: The integration of remote sensing, Geographic Information System (GIS), Global Positioning System (GPS) and Telecommunication, International Journal of Remote Sensing, 23(9):1851–1893. Zhou, X., J. Yates, and G. Chen, 2001. Using visual spatial search interface for WWW applications, Information Systems, 26(2): 61–74. Zhou, Z.H., 2003. Three perspectives of data mining, Artificial Intelligence, 143(1):139–146.
(Received 28 March 2005; accepted 07 July 2005; revised 15 August 2005)
PHOTOGRAMMETRIC ENGINEERING & REMOTE SENSING
