Jump to content

cdschroer

Administrators
  • Content Count

    452
  • Joined

  • Last visited

  • Days Won

    78

Everything posted by cdschroer

  1. Thanks Dave - good suggestions! I'll add that the xml file indicates the error is in ptmfitter: DATEF="" DATES="" LEVEL="WARNING" USER="">Fitter returned an error ! Error : -1</xc:event> We see that in the screenshot as well, that ptmfitter is not able to open the file. So, try building an RTI with the hshfitter and see if that works. You have plenty of memory on your computer, and we build files this size with a less beefy machine. However there have been some issues with large files with the ptmfitter on Windows that don't occur on Mac. Also, are you rotating the images before you put them in RTiBuilder? In other words, are the images landscape or portrait, and are they all the same? They really need to be all the same. Hope this helps. Carla
  2. Can you say more about the computer you are using and it's specs. For Example Mac book pro with 16GB RAM and MacOS 10.12.6 Also - you can look in the .xml log file that RTI Builder produces, which will sometimes give you a clue as to what is going on. In this case- from the screen shot it appears that RTiBuilder is successfully making all of your cropped files, and the failure happens within the PTMFitter. The fact that the PTMfitter can't open the file, might indicate a permissions problem. Where did you instal the ptmfitter? Are you able to build HSH based RTIs? the HSHfitter is included with RTIBuilder, so it should just work, without installing extra software. One other thing to try is to make smaller jpegs and see if it works - We regularly make RTIs of this size or larger, but we have fairly "beefy" machines with a lot of RAM (16GB - 128GB) Carla PS: If you don't know what the "fitters" are - you might want to check out the Glossary: http://culturalheritageimaging.org/What_We_Offer/Downloads/Capture/CHI-RTI-Glossary_v1.pdf
  3. cdschroer

    Basic hardware

    I think that John is interested in RTI - which doesn't require as beefy of a system as photogrammetry does. The good news for RTI is that it will work on a broad range of systems, though it will definitely be faster to process on a faster machine. Also, this somewhat depends on the resolution of the camera you will be using, and therefore the resolution of the finished RTIs. If you have 16GB of RAM you will be handle most RTI image sets and also load and view them in RTIViewer. I agree with Leszekp about the graphics card if doing photogrammetry, but not as critical for RTI. If you are doing a lot of photography with high resolution cameras (>25MP) then this will definitely help you. So bottom line is max out the RAM and CPU and graphics card for what you can afford. 16GB is an absolute minimum to work with high resolution images (including RTI) - and 32 if planning for photogrammetry. Pay attention to the the other components too, and not just RAM. Carla
  4. I rebuilt the zip file using a different tool. I tested it - and it seems to be working fine. The original version tried to put it back in my applications folder when I unzipped it. So, hopefully this will fix it - as it seems to work correctly. Carla
  5. In the dropbox are two folders one for Mac and one for Windows. In the Mac folder is a copy of the PTMfitter I have on my mac and run regularly. It doesn't seem to have an extension. I have it in my applications folder. I don't have a disk image or installer for it. I think it should work without an extension. I just tried downloading it and dropbox adds an .html extension to it. When I remove that it still thinks it's an html file when I run it from the terminal. So, I've uploaded a new copy as a zip file. Unzip it - and hopefully that will work. I don't know what the dropbox downloader is doing to the file - but it doesn't seem to like a file with no extension. Note that it's a command line tool, and not designed to work by double clicking on it. You could test it by running it from a terminal window if you are comfortable with that. Or just browsing to wherever you unzipped it from RTIBuilder in the last screen (where you execute the RTI) Carla
  6. I think there was an error in the above recommendations - you absolutely need to align the images with the scale bars in order to use them to set scale for your project. How you placed the scale bars will affect exactly how you work with those images. The Agisoft software will not be confused by some images that have scale bars and some that don't for the same area of your scene - as long as you have good crisp photos with good overlap between photos and plenty of look angles. It's a bit difficult to describe in text, but hopefully this makes sense. Carla
  7. If you shot it properly, all the images should align together in the same chunk. Also for a turntable project, you likely need to mask background images, and then when you align, apply masks to tie points (in the advanced tab) You do not need to mask the scale bar images. You can set the region of the model later in order to remove the scale bars. There are other ways to remove the scale bars as well - but initially, you want them in the main alignment. If they don't all align together that may indicate a different problem (such as blurry photos, monochromatic surface with no texture, etc.). In that case you can align them in separate chunks and try to bring the chunks together. Just understand that at that point you are working around a problem and a properly shot image set should all align together. Hope this helps, Carla
  8. We were working on a version 3 some years ago, and we gave a Beta version to some folks who took classes with us. We were never able to release it due to some bugs that were too difficult to resolve (long story). So, it is possible that you have a 3.0 Beta version. I recommend going back to the 2.0.2 version from the website. The 3.0 Beta is known to not work with higher resolution images, as well as having some other issues. Even though 2.0.2 is marked Beta - we really should have marked it final. It's getting old now, but it's the best option for current use. We are very interested in doing a new version, but it will take some funding, and we haven't been able to make that happen yet. Carla
  9. We recently discovered that the PTM software is no longer available from the HPLabs website. We have placed both the Mac and Windows versions of the PTMFitter in a dropbox along with the license file from HP. You can get the software from here: https://www.dropbox.com/sh/jfsy0lhxu6zv4i4/AADJpq6E_GJmNw_s5C8r94CVa?dl=0 Also - many papers on PTM and RTI are by Tom Malzbender - and you can find them on his personal website: https://sites.google.com/site/tommalzbender/
  10. We recently found out about this. I've set up a dropbox folder with both the Windows and Mac versions of the PTMFitter, along with the license from HP: https://www.dropbox.com/sh/jfsy0lhxu6zv4i4/AADJpq6E_GJmNw_s5C8r94CVa?dl=0 Carla
  11. I just had to do this on Mojave - and it works like a charm. We are looking into getting the proper certificates so that our software can come from a "known developer" which would prevent this problem. The certificates cost money, and it's different for Mac and Windows. That plus work needs to be done on the RTI software. We are aware of this issue though, and we will do what we can to resolve it. We appreciate all of you that support our work with donations and volunteer efforts. It really makes a difference. Carla
  12. We are focusing on good practice for collecting image sets for photogrammetry - which is independent of the software used. We feel that if people collect good image sets that follow the rules (as described in our videos) then they have useful information for now and the future contained in those image sets. Especially if they collect some metadata about what they are doing, their methodology, who was involved, the subjects and locations, etc. We are developing tools for that called the Digital Lab Notebook. While we do use Metashape (aka Photoscan) at this time, the methodology is not dependent on the software. So our freely available public information is not about Metashape. We agree that the Metashape community forums are a great place to go for information about working with that software. We do use an error reduction workflow in Metashape that is a bit different than the process most folks follow. We teach that in our training classes. We spend time looking at metrics and what they mean, and also trying to impart an understanding of the sfm photogrammetry approach, and how that informs the guidance we give on collecting image sets. Our goal is to impart knowledge rather than give a "cookbook" approach, so folks can deal with a variety of situations in their own work, and make the appropriate trade-offs to meet their needs. This kind of material doesn't fit into a forum post. Carla
  13. Thanks for posting this! We recently got a 4K monitor we are using in a Windows environment, and ran into this issue (we are mostly a mac shop). We hadn't tracked down the best way to deal with it yet - and this is very helpful. Note that RTIBuilder (and some of our other tools like inspector and DLN: Capture Context) are written in C++ using the qt libraries for the UI. I'm guessing this is an issue with the qt libraries not handling scaling for hi-res monitors. We are seeing this problem with all of our tools. If you want to check out the Digital Lab Notebook (DLN) tools go here: http://culturalheritageimaging.org/What_We_Offer/Downloads/DLN/index.html Carla
  14. Hmm - I have seen this before - but nothing else is coming to mind at the moment. Is RTIBuilder producing an xml log file? It should be in the ProjectFolder in your example. It should have the name that you gave the project when you started the project. Sometimes there are useful hints in that file. Can you take a look and/or post it here? Carls
  15. Janis - You have to organize your images into a folder called jpeg-exports. The folder should be under a folder for your RTI project. For example a folder called "painting1" with a folder called "jpeg-exports" inside that. When you use RTIBuilder - you tell it to load images from "painting1" (Us the bowser tool to find that folder) Then it will look for a "jpeg-exports" folder within that folder. If you do not have a folder with that name, it will not find any images. This is documented in the RTI Guide to Highlight image processing: http://culturalheritageimaging.org/What_We_Offer/Downloads/Process/index.html I hope this helps. Carla
  16. Hi Josh - I have a couple of ideas. First are you using uppercase JPG for your filenames? There is a bug in the ptmfitter that will not accept files named myimage.JPG but will accept myimage.jpg. Otherwise I wonder if you have some kind of permissions issue. Are you sure that you have full permissions on the full path to where the images are located? I don't have a lot of other ideas off the top of my head, but those two jump to mind. Thank you for providing very detailed information about what you did and what happened. That is very much appreciated. Carla
  17. Alisali - I really don't understand what issue you are having. You are getting an invalid file error - but I don't understand the circumstances of whether it's on a Mac or PC. If this is still an issue - then please provide more details about exactly what you are doing, whether the file is a reasonable size, (i.e. should be MB not K) and whether the files worked in one environment and failed in another. Sorry for my delayed response. I'm not getting notifications like I used to, and just found this. Carla
  18. Great question Jack! First, we think collecting high quality image sets with proper overlap and good geometry is the most critical thing. If you have that and you archive the images (preferably DNGs but could be TIFFs) along with good metadata about your project, you have something really useful for the future. If you want to include models, we would recommend the OBJ or PLY formats. We also recommend exporting masks and camera calibration data if possible. These could be useful in better understanding your outputs. Masks especially can take a lot of work and you don't want to lose that. Software like Metashape (AKA PhotoScan) can easily export these things. Other software may or may not be able to do this. As part of our Digital Lab Notebook tools we have been working on some recommendations for archiving. This is still an early draft and we are getting some feedback from some advisors - so stay tuned. Carla
  19. Somehow I didn't get notified of this whole thread. So sorry that I missed it until now. When we updated the forum software it changed the notification mechanisms. Anyway here are a few thoughts. There are not hard limits for processing RTIs - though it will depend on how beefy your machine is. We use a Canon 5DSR and can process the data with no problems on Mac laptops with 16GB RAM. I seem to recall that there were some issues with the old ptmfitter with large images on Windows, but that shouldn't affect the HSHFitter. AS for the image shown here - it's a bit hard to tell from a single image but I have 2 guesses: 1. Using a zoom lens the lens can slip when pointed down. Normally you would see images go out of focus when this happens. You can use tools like Lens Band to stop this kind of creeping, or ideally, use a fixed focal length lens. 2. If the individual images look sharp, and the RTI is sharp in some light directions, and goes out of focus in other light directions something likely moved in your setup. It could be the camera, or it could be the subject. With a 50MP camera very small movements can move several pixels creating this kind of effect. One way to look for this is inside RTiBuilder after the spheres are detected - try choosing the first image in the image set, and then the last image. You can see if the sphere moved because the red ring showing the sphere detection will be in a different place on the sphere. You can frequently go through the images and find the image where things shifted. Then you could process just the images before or after the move (whichever is a bigger set). We refer to this kind of RTI as containing "a mover". Carla
  20. The format specifications for both RTI and PTM are available - and there is an item in the FAQ with a link for PTM and the file for RTI.
  21. Jack - It would help to see some screen shots of what is happening. Things with narrow edges are difficult to get alignment of the 2 sides. I can't tell from your post, whether you are shooting circuits with the pieces upright, or shooting each side and turning over. That would affect the approach to solving this. Also since you say that images of the "bottom" don't align - I'm wondering if they would align to themselves in a separate chunk - then you might be able to bring the chunks together. That is not the optimum way to do it, but it could be a interesting test. I think you know that the biggest issue with relatively flat subjects, is having enough data across the thin edges to get the sides to be in the right relationship to each other. Also, if you shoot each side, you can put each side on a different background, so that the software doesn't match the backgrounds - this can create errors in alignment. That kind of problem could be solved by masking - but it's easier to just use different backgrounds. Carla
  22. Glad you were able to sort it out yourself. Well done! Carla
  23. You do not say what platform you are on. If you are on the Mac - then you must make sure that you installed the software by dragging it from the .dmg (disk image) onto the applications folder. If you try to run the software from the disk image (dmg) then these kinds of problems can occur. If this does not solve the problem, please give more details about the environment where you are running the software. Carla
  24. Hello! You want even illumination across the surface of your subject - and from an appropriate standoff distance (2 times the diameter of the area you are imaging is the minimum) We use a wide beam spread for things like a speedlite. So ours are set to 24mm for the old Canon 580ExII model. Some of the newer models have an even wider spread. You want to make sure you are pointing the light at the center of the subject, and keep pointing it there for each light angle. In other words, you don't want some light pointing at the center, and others at the edges. All lights should point to the center. If you don't have enough light standoff (the distance to the front of the light from the center of the subject) you will not get good quality normals as you move away from the center of your subject. This is usually visible in the normals visualization, and can often be seen in the specular enhancement mode, especially with no color. Carla
  25. Thanks for reporting this! I wasn't aware of it, and don't think it has been reported before. We don't have any plans for updates to the viewer in the short term, but we will put this on the list for the next time we do updates. Carla
×
×
  • Create New...