Latest Contributions by PeterBecker

When compiling the ISDef the footprints are taken from the shape file not from the RPDefs. The synchronize is primarily to update the RPDefs. A work flow in which you may want to update the RPDefs is if you are importing the RPDefs into a different ISDef. In this case it would be easier to use recompute footprint by related geometry to copy the footprints across. If the service is not compiling properly then is must be another issue. One possibility is that the footprints are very detailed and you have some intersecting vertices. If you have generated the footprints the you may want to recheck this. ... View more

Following Python based code may help you. It takes as input the extents and pixelSize range values and Cols & Rows. You should be able to covnert easily. Ensure you define in the request the Width (Cols) and Height (rows). ... View more

For individual rasters open the table and for selected raster click on  <raster>. Click on Function Tab and look at properties of the specific raster. (A function can have more then one raster) To get the paths for a collection use Datamanagement Toolbox, Raster, RasterCatalog, Export Raster Catalog Paths. This also works on Mosaic Datasets ... View more

There is a direct relation between pixel size and map scale. ScaleFactor = Cell Size x 96/0.0254 Therefore, if the scale is 1:20,000, the cell size is 5.29 meters. Cell Size = ScaleFactor x 0.0254/96 Therefore, if the cell size is 5 meters, the scale is 1:18,898. The scale at which the overviews will be displayed is the MinPS value of the Service overviews * 96/0.0254. (Since the base imagery take priority, it is actually more correct to use the MaxPS of the base imagery, but this is generally set the same as the minPS of the overviews) Note that overviews need not have the same scale at all locations. If the resolution of the base imagery varies then the minPS of the overviews may also be different in different areas. ... View more

Help on 9.3 Image Server can be found at: http://webhelp.esri.com/arcgisdesktop/9.3/index.cfm?TopicName=An_overview_of_the_ArcGIS_Image_Server_extension The similar help at V10 can be found at  http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html#/An_overview_of_ArcGIS_Image_Server/009s0000008w000000/ The Image Server technology is not going away. It is though being fully integrated into ArcGIS. It is therefore a technology not a stand alone product. With ArcGIS10 the on-the-fly processing has become a core component of ArcGIS in the form of functions that can be applied to raster datasets and stored as part of layers. Such on-the-fly processing can be performed both by the desktop and server. The functionality enables you to define processing to be applied to the imagery as it is accessed without the requirement to create new datasets. These functions can be accessed through the new Image Analysis Window. The Image Service Definition (ISDefs) used by Image Server at 9.2 and 9.3 have been replaced by a new Mosaic Dataset that is stored in the geodatabase. The mosaic dataset enables the cataloging of very large collections of images and rasters as well as defining the metadata and processing to be applied to the imagery. Storing them in the geodatabase enables them to be scaled to massive numbers of rasters. As before the pixels can remain in their original format, converted to more optimized formats, or if required stored in the database. No extension is required to create Mosaic Datasets as it is a core part of ArcGIS desktop (Editor or Info). Creating Mosaic Dataset is the recommended method of managing small to vast collections of imagery and rasters. A new set of tools is provided to enable the creation and management of mosaic datasets replacing the need for the Image Service Definition Editor.  The serving of rasterdata sets (optionally including processing) is a core capability of ArcGIS Server Image Services. Any raster dataset (or layer containing a raster datasets and a set of functions) and be optimally served as an Image Service. The Serving of Mosaic Datasets that contain collections or rasters requires the Image Extension. When Mosaic Datasets are served as Image Services they become accessible both as images with the server performing the on-the-fly processing and dynamic mosaicking, but also as catalogs that can be queried for detailed metadata. These changes make imagery very tightly integrated into the whole ArcGIS system and similar tools and models that can be applied to vector datasets and also be applied to imagery. Users that have licensed Image Server in the past will be upgraded to the Image Extension. For existing users that have workflows that use ISDefs the Image Server with its separate Image Service Definition Editor is still available as a separate install, but users are recommended to migrate to using Mosaic Datasets. The migration is simplified by the fact that ISDefs can be directly loaded as a rastertype into a Mosaic Dataset. For more info check out the Help topic: http://help.arcgis.com/en/arcgisdesktop/10.0/help/index.html?TopicName=What_is_raster_data%3f#/ArcGIS_Server_Image_extension_version_10_versus_9_3/009t000001vn000000/ Also check the blog. There are two good items there and some comments on the What�??s new.  http://blogs.esri.com/Dev/blogs/imagery/Default.aspx ... View more

On-the-Fly processing enables the server to perform both geometric and radiometric pixel based processes on the imagery as it is accessed. This processing can be for example applying a LUT (look up table) to change the color (a radiometric process) or re-projecting the data (a geometric process). There are a large range of processes that can be applied. Orthorectification is on example of a geometric process. For an image to be orthorectified a sensor model with orientation parameters and a DEM need to be available. The most common sensor models supported are the Standard Frame (primarily used for frame based aerial cameras such as DSS, UltraCam, DMC) and RPC (used for many satellites such as QuickBird and Ikonos). The frame based sensor models require parameters such as the exterior orientation of the each frame,  focal length, camera distortion and interior orientation. The Digital Elevation Model can be a raster (such as a 32bit TIF or FLT) or another image service serving elevation data. Different raster types can be used to set the ISDef (in V9.3) or MosaicDataset (in V10). These import the orientation data etc and set up the processes. As a users access an area of interest the system determines which images should participate and applies the appropriate process on the fly only for the pixels that need to be viewed. This processing is very fast. For Orthorectification it extracts the required elevation data and then applies the Orthorectification processes based on the provided sensor model and parameters. V10 is very similar but provides more flexibility in defining additional sensor models. In V10 all this capability is also part of the desktop. ... View more

This is an issue that affects a number of users. When ArcMap opens a raster that contains stats it by default applied a 2StdDev stretch. This can be turned off in the ArcMap 'Tools,Option,Raster, Display rasters with contrast stretching' option. This though is only for one machine. When you do a build with compute output properties on an ISDef, the stats for the service get created and so by default when the services is added to ArcMap is applies a stretch. There does not exist a specific tool to remove the stats from a service. The simplest method is to open the xxx.ISDef/ImageService.ISDef file with a text editor. Find and delete the node called <Bands> IE the section <Bands> <ActualValues>true</ActualValues> <Alias>Histogram Generation for ImageService</Alias> ...... </Bands> You can also manually edit the values to specific values if that is requrired/ You will need to recompile the service. No need to build anything. Do not run the Build with the "Compute output properties" else the stats will be regenerated. In V10 this is resolved by a Mosaic Dataset having a specific property "Is Preprocessed Data" that when set to Yes results in no stretch being applied when added to ArcMap. ... View more

This is a know issue with a Beta version of 10. The issues should have been resolved in final. ... View more

I'm not sure which version of Terrashare support what functionality. ISAT project files are a general part of all the Intergraph photogrammetric projects. If your organization is using ISAT for its AT then they will have these ISAT project files, which can be used with or without Terrashare. The ISAT files define the parameters and the location of the files. if the location of the files have changed it can be easily edited by search/replace in ISAT files. Yes the system is very scalable. Certainly more CPU is used in relation to the serving of pre-cached tiles, but the CPU usage is still relatively small as only the requires pixels (1 screen full are processed). Even with orthorectification and pansharpening this is <1sec on a core. IE A single 4 core system could serve easily 24 concurrent users. (Assuming each users zooms about 1 every 6 seconds). This is a relatively conservative number and is dependant on things including the format of the data. Best is to try it out. The system can be easily scaled by increasing the CPU. Areas of higher interest can also be pre-processed. IE you can specify areas of interest that are likely to be hit more often and pre-generate the orthos in these areas, so increasing the performance. It is possible that your organization wishes to continue doing the orthorectification with a specific product that meets their certification requirements. When these orthos are created they can be used in place of the OTF orthos. If mosaicking can color correction is done this can be added to the same service. IE the grade & performance increases over time. If you wish to have a cached background image just for visualization then yes you can use an Image Service in an MXD and cache this. You can cache the complete service or use on-demand-caching. IE cache is created in advance for the most popular locations and then cache is generated as required when the first users pans to a specific location. ... View more

Check the properties of the raster that comes up black in comparion with the other rasters. There is a recent seperate review on a possible issue between some files that are MSBFirst vs LSBFirst. Some of the format readers can get this wrong. As timmerspk  mentions, try converting that one file if necessary. ... View more

You can use an Image Service directly in a geoprocessing tool by first using 'Make Image Server Layer'.This enables you to define various properties such as the mosaic method. The Spatial Analyst tools will use the Raster environment variables that include defintipon of the Cell size. I would recommend you specifically define the cell size such else the system will use the Base Pixel Size which depending on the defined extent may be a very large request. ... View more

Looks like something incorrect in the definition of the top most service overviews. Suggest you select and delete the top (largest LoPS) overviews, then redefine the service overviews and regenerate. May have been caused by an error in the boundary. Re-generate the boundary prior to redefining the service overviews. ... View more

Using DispOrder overrides the mosaic method. Mosaic Method is only applied to records where the dispOrder is 0. To force images in a speciifc order set the disp order to a lower (negative value). IE set -1 for the 2009 and -2 for the 2005 images. The MosaicMethod should also work. For example if you have a field called year, set the be Mosic Method to ByAttribute, set the Field to be Year and the Base value to be say 3000. Images should be ordered closest to year 3000 so long as they are in scale. Double check that the MaxPS values enable the 2009 imagery to be displayed at the smaller scales. ... View more

There is no need to cut the image and craete a new pne. Instead modify the Image Service definition. You can individually edit the footprint of the different rasters to exlude areas that are not to be seen. Alternativly to clip something gloablly edit the Boundary layer and then Recompute Footprint by Clipping to Boundary. Any pixels outside the footprints will be clipped without the need to creating any new dataset. Scuh clipped images work well in ArcGIS applicaitons since transparancy is correctly handled. IE the image can be transitted using JPEG compression, but a mask is also transmitted to cleanly handle the transparent areas. Note that in ArcGIS 10 the desktop can also do this. Any raster can be clipped using a graphic and the result stored as a LYR file that can be used in applications such as ArcGIS Explorer. ... View more

For large service look to using c:\Program Files\ArcGIS\Image Server\Developer Kit\Automation\DeriveImages (avaialble after installing the DevKit). This enables the server to create the Service Overviews and is more performing. Also double check the size of the RPDefs. When service overviews are created the system generates DIDefs that contains all the metadata from the dependant rasters. If the rasters contains a lot of metadta and a large number of rasters are required for a single MD tile, then these DIDefs may get too large and cause such an error. If so consider removing some of the metadata from the RPDefs (using Edit Raster Properties to replace XML nodes) ... View more