Hi, I wonder if you could help me on my learning curve with WebODM?
I have done a lot of research over the weekend, I find conflicting advice and suggestions (I assume this is because some of the information returned is from previous older versions) so I am hoping that I can reach out for some assistance from the community.
I am using a WebODM install on Windows using Git,Python and docker as per instructions.
Issue 1 - How do I increase the HD space available to the default machine?
It has created a 20GB vmdk which I realise cannot be resized, what is the easiest way for a linux noobie like myself to increase this as I find I am filling this up when trying to run my data on the settings I would like to.
Issue 2 - Again disk related, I am trying to redirect the output to my windows environment rather than the results being output in to the virtual space (because of Issue 1 primarily), when I use the ./webodm.sh restart --media-dir /path/to/webodm_results using my drive letter and directory name for the output location, it accepts the command with no errors and restarts but the output is not being delivered to that location. I am sure this is likely something I am missing with my lack of docker and Git knowledge about how this mapping works, if anyone could lend me some advice on this I would appreciate that.
Issue 3 - Unable to produce high resolution ortho
I have tried many of the built in options, many of the tweaks to settings that I have researched but seem unable to produce an ortho of the highest possible resolution.
The image set is made up of 5472 x 3648 resolution images flown at a height that should produce 2.9cm/px, I have tried different settings of -orthophoto-resolution, increased -texturing-nadir-weight, increased -mesh-size and mesh-octree-depth but the output is always a small filesize tiff with low resolution in comparison to what I am expecting from the data, MME for example delivers a near 1GB file, WebODM is 60MB.
On the plus side some of these settings have produced a far better stitch than other platforms have output but just not the resolution.
I have read conflicting information on whether the resize images will effect the final output? If I do not resize the images my machine does not have the horsepower to complete the job, I have thrown a couple of attempts at webodm lightning as well but again I am not getting what I expect.
I am sure this is likely something very simple I am missing but again I would really appreciate some help with this if possible.
I will truly appreciate any advice on offer to help me with this please.
#2 is a problem most likely related to the fact that --media-dir should be a subpath in c:\users\<youruser>\, other paths won’t work with VirtualBox unless you have them shared: Divio Support | Divio
#3 can be solved by turning on --ignore-gsd.
There’s no reason some questions don’t get answered; we’re all volunteering our time here on the forum, if a question is easy to answer people do it, more difficult questions might simply not be known, or a person that knows the answer is not around to answer it.
Some things people can do to make it easier to get answers:
Formulate the questions in simple, concise words
Don’t stack multiple questions in the same thread. One question per thread is better
Use a descriptive title for your thread
Provide screenshots, details and data (logs, task outputs, input images, output results, etc.)
Explain what things you’ve already tried to solve the problem
Also searching the forum can help get answers for questions that have already been asked.
I think the other part here is, once you know (or think you know) the answer to a category of questions and you see that question posted, you are now deputized to answer! So, when you see these questions in the future, feel free to wade in.
Thank you, as you will see from the above I had already answered issue 1 myself as I had found a post to help me with that.
Issue 2 that is very helpful.
Issue 3, I am sure I have that option enabled but as yet I still seem to be getting smaller than expected results, but i am still trying and learning.
Yes, we are all learning here and we are glad you are here. The advice from Piero above is a useful perspective on how to get answers quickly amongst a volunteer community.
You will need to both set ignore gsd and set your target resolution for the orthophoto output. You should be able to verify it has used that resolution by checking it a tool like QGIS or with gdalinfo. I wouldn’t rely on file size as your determinant, as we don’t know which tools are compressing internally.
