From KAREN wiki

• Description

• Contacts

• How does this project use KAREN?

OASIS – Online Archive of Satellite ImageS M.P. Tuohy and M.E. Irwin Institute of Natural Resources, Massey University, Palmerston North, New Zealand Email: m.tuohy@massey.ac.nz Abstract Over the past twenty years a considerable collection of medium to high resolution satellite imagery and aerial photographs has been acquired by the Institute of Natural Resources at Massey University. This archive is a valuable resource for both research and teaching. It includes complete coverage of New Zealand (SPOT, Landsat and ASTER) and several sets of images that cover only part of the country (ALI, Hyperion, Quickbird, digital orthophotos).

Users of this imagery range from remote sensing experts to those who want only a background image to illustrate a paper or presentation. Often the area required is not covered by a single image and some processing is necessary to generate a suitable image; sometimes a researcher wants a sequence of images acquired on different dates; or a lecturer is developing an exercise using imagery of the same location but from various sensors with different spatial resolutions. These situations require the attention of the very few staff with appropriate skills to locate the imagery, process it and deliver the finished product in the correct format.

At the beginning of 2010, the implementation of ERDAS Apollo has made the archive much more accessible to many more users. Much effort has been put into creating a spatial database of the imagery, converting the data to a common format with a common projection and creating mosaics that can be stored and served more speedily as compressed images. All of the imagery, whether it is a composite image or raw data, is now available via the web. The student, teacher or researcher can browse, select an area of interest, choose the type of imagery and then request the data. The imagery is clipped, zipped and shipped automatically.

1 Overview Google Earth is great for viewing imagery and downloading an image that can be used in a presentation or document. But these downloaded images are in .jpg format and cannot be used for further analysis. Furthermore, what if you weren’t interested in the high resolution imagery and wanted to do a regional or national study that required SPOT of even Landsat imagery over an extended time period. At Massey University we have been serving imagery to our students and staff since 2002 (3 years before Google Earth) using both purpose-built and off-the-shelf software. The ATLAS website http://atlas.massey.ac.nz continues to do things a bit like Google Earth except that the user can select the type and date of imagery. A more versatile image portal has now been implemented using ERDAS Apollo. Some of the features available via the new OASIS website http://oasis.massey.ac.nz will be demonstrated in this paper.

2 The ATLAS Website This data portal provides access to imagery and vector layers, but only for New Zealand data. In addition to Landsat, SPOT and ASTER satellite imagery, the user can view and download digital orthophotos and scanned topomaps. A data search facility permits users to search for and download raw data files but this was never fully developed in terms of populating the database with all available imagery. Large mosaics of individual images have been created to allow users to see the complete national coverage by the various satellites. Similarly the higher resolution orthophotos and maps have been put together in mosaics that cover only smaller areas so that data transfer and display times over the internet can be minimised. There is also a link to ArcGISWeb where vector datasets can be displayed at any required resolution, queried and printed. Available datasets include national topographic data (roads, streams, lakes, railways, etc.), New Zealand Land Resource Inventory and Land Cover Database polygons, the FAO Soils of the world and a very detailed soil map of the Massey University farms.

2.1 Satellite imagery on ATLAS The national coverage of SPOT imagery is shown in Figure 1. With tools provided by Image Web Server (now an ERDAS product) you can zoom and pan anywhere within this large mosaic of all the available SPOT imagery. Whatever is displayed can be copied to the clipboard or saved as a .png file for later use. Figure 1. Mosaic of SPOT imagery for New Zealand on the ATLAS website. Figure 2. Part of the ASTER mosaic, showing Mt Ruapehu that has been copied and pasted from the ATLAS website

3 The OASIS Website With the release of ERDAS Apollo in 2009, the stage was set to make significant improvements to the capabilities of our image web server. We wanted a more straightforward method of searching, selecting and downloading georeferenced, multiband satellite imagery. Quickbird imagery was being acquired through the KiwImage project and this was going to add another dimension to our catalogue. While pan-sharpened Quickbird mosaics were available in compressed formats, we also wanted our students to be able to access the mosaics as geotiffs of the original data in the New Zealand Transverse Mercator (NZTM) projection. The same requirement applied to all the other imagery in the archive. Often a project requires several images to cover the study area; in the past, you would have to download all these images separately and then create the mosaic. This is not difficult if you have appropriate image processing software and skills. In Apollo it is simply a matter of creating aggregates of the imagery and then the mosaics, or portions of them, can be downloaded and used immediately. Figures 3 to 7 demonstrate the Apollo server and some of its capabilities. Figure 3. An Aggregate of all the available SPOT5 imagery of New Zealand displayed in ERDAS Apollo. Figure 4. Zoomed display of the SPOT5 aggregate with a box defining the area of interest for which imagery is required. Note the AOI covers several SPOT5 scenes. Figure 5. The geotiff captured from the three SPOT5 scenes. Another extremely useful capability in Apollo is the clip tool. With either an image or an aggregate displayed, you can define a box or indeed a polygon of any shape and then download only that part of the image that falls within the defined area. Figure 6. An odd-shaped area defined on a Quickbird image. Figure 7. The odd-shaped area extracted as a four-band geotiff.

4 Discussion The implementation of ERDAS Apollo on the OASIS website has made the imagery both more accessible and more tailored to the needs of the intended users whether they are first-year students or experienced researchers. Good metadata are essential for accurate cataloguing and searching of the imagery. The Apollo Data Manager makes reading existing metadata a simple operation and then searches by object type, keyword, spatial and temporal extent can be carried out in the Apollo client. There are further enhancements that we will be looking for in the future. These include the provision of better colour-balancing in the mosaics and the ability to download a time sequence of imagery from a single sensor or even several different sensors. The generation of 3-D perspective views using the imagery of your choice over a DEM with a suitable resolution will also be something we hope to have available in the near future.

5 Acknowledgement Thanks to REANNZ (Research and Education Advanced Network New Zealand Ltd), the company that has supported the development of OASIS through its Capability Building Fund to promote KAREN, the network that provides lightning fast, unrestricted broadband for the New Zealand education, research and innovation communities.