03-07-2017 01:44 PM
I have set up several models in ERDAS APOLLO to do geoprocessing of images. Ideally, I could select several images and have them run in batch. However, I can't even enter a filename and have it run correctly unless I type in the entire path to the image as well as the image name.
First, is it possible to link a set of selected images to a model for geoprocessing as a batch?
Second, if this is not possible, is there a way to avoid having to state the file path as well as the image name? That is, is it possible for APOLLO know the path for a file that has been selected in that program?
Thank you for your help.
03-08-2017 11:04 AM
As a follow-up to my previous post, I am increasingly convinced that the problem is that Selection does not work. I would assume that when one Selects an image, the entire file path is preserved, so that one should be able to paste the entire filename+filepath into the input field. If this is so, I would appreciate it if somebody would confirm that. If it's the case, then the problem isn't geoprocessing (that seems to work); it's select -- see another posting about that.
03-09-2017 03:06 AM
It is not possible to run a geoprocessing in a batch using the Geospatial Portal. But you can create you own script using the WPS interface (or create a Imagine batch if you have access to the gmdx...).
You can also build a model that accepts more than one input.
Concerning the file path, in the Portal you can simply select one image and then choose ("add as input"). See screenshots in the other conversation: http://community.hexagongeospatial.com/hmrkh95973/board/message?board.id=Discussions_KS_ERDAS_APOLLO...
03-09-2017 08:38 AM
Thank you for your response. Something still is not working. I have checked over all of the screenshots you included, and I think that I have done everything required. However, when I select an image, the "add as input" field is greyed out.
I did think of one other thing that might be getting in the way. For technical reasons of the camera taking the images, the images in question are JPEGs. Do I need to add JPEGS as a supported data type? If so, how do I do that?
03-11-2017 04:30 PM
Again, the response goes to the point of showing what needs to be done, but not how to get to the point that needs to be done. If it were possible to check the "add as input" field, it would be fine -- that would work. But this is greyed out. What do I need to do in order to be able to "add as input?"
03-13-2017 09:06 AM
The data type in the model is Imagine:File. The data type of the images is JPEG. When I run the model in Imagine and navigate to the directory containing the JPEGs, the model works fine. I do have to change the data type from *.img to *.jpg. The only things I can think of as problems are  the model doesn't know how to access a JPEG as a data type or  it doesn't like spaces in a path. Does either of these make sense to you?
03-14-2017 08:53 AM
As I lay in bed last night obscessing about this problem, I realized a possibility to narrow things considerably.
If you "Select" one or more images, and then click "Download," you can download the images. This suggests that "Select works properly. However, if you click the "Copy to Clipboard" icon, that does not work. Nor is it possible to click on the field you suggested in an earlier posting, as it is grayed out. However, if you put the fully qualified path to the image (e.g. d:\original files\western lake erie\2016-7-15-RA\VIRB0227-1000.jpg), the model works.
This suggests that there's a problem somewhere in the Selection algorithm. Ideally, it would be a permissions issue, but I don't have the slightest idea where one might start. Do you have any ideas?
03-16-2017 07:43 AM
And if you select another raster image, is it possible to choose it as input?
Your JPG are georeferenced or just geotagged? I suspect the Portal to treat the JPG as business objects like documents, geotagged images, that cannot be use as input for a spatial model.
03-16-2017 09:19 AM
I just tried geoprocessing two different images, and it appears to have worked fine. The output does not retain any georeferencing or geotagging information, but that isn't important for the purpose in question. I haven't looked at the outputs, but I have no reason to believe that they are not outputs derived from the inputs (I am at a different location from the server, and I haven't enabled users to get access to the output directory, so I have to go downtown to get local access in Imagine). In any case, two directories were created, dated today, at the times I submitted the inputs, with names that suggested success. If this is not the case, I will update this email.
To be clear, I do not need georeferencing in the output. The inputs are geotagged by Garmin VIRB software, and they show up properly on the map. The proper thumbnails have been created for the images chosen, and the images are properly located on the map. Likewise, when I Select them, the proper stars turn green, and the corresponding downloads are downloaded. The only thing that is not working is that the links to the images -- which are clearly being created by the Select process -- are not being passed to the Geoprocessing models.
If it helps, the information that should be passed for one of these images is D:\original files\Western Lake Erie\2016-7-16-RA\VIRB0228-157.jpg. I can imagine that Apollo is behaving like ArcGIS and barfing with the spaces -- in which case I can easily fix it -- or it might be something else. I find it hard to imagine that geoprocessing is looking for extents in an image properly located by a point, but who knows?
Thanks for your help so far.