Advice for ImageGroup Import
Image groups are bundles of images that can be uploaded onto the 3DUserNet Platform. These
are visually displayed in three formats:
1 2 3
- Orientated Panoramics
- Orientated Images
- Non-oriented Images
For Orientated Images and Panoramics (1, 2), an orientation of each image must be provided as well as position. Furthermore, a focal length is required for adding orientated images (2) to a scene. For non-orientated Images (3), only a position is required, when a user goes to a view in this dataset, the images are positioned in a horizontal orientation and will not align with the data. Please note that Panoramics (1) must be full 360 by 180 equirectangular projection images. When bringing your data into VISION it can extract information from two sources:
(a) EXIF and XMP data
Our software reads the XMP data from images and takes either DJI format or Microdrones Format to read Course, Pitch, Roll. It will also extract the focal length either from the XMP data or from 35mmFocalLength tag in the EXIF data. Latitude, Longitude and Altitude is also extracted from the EXIF data.
(b) CSV files (required for panoramas)
If your image files do not have anything stored in the EXIF data, or you are importing panoramas, then a CSV file is required. The CSV can be either space, comma or tab delimited. Headers are optional as the user can select to ignore a number of lines before the main data columns. The columns can take the following information:
- Filename: name of image (with or without extension-case sensitive)
- Positional: x, y, z or latitude, longitude, altitude
- Rotational: Course, roll, pitch, or quaternion (q1, q2, q3, q4) or vector angle (ax, ay, az, angle)
- Focal length (as 35mm equivalent)
- Pixel length
- Pixel height
(c) combination of the two
The software will always try to orientate Images, therefore if not all the information is in the CSV file it will try and extract any missing information from the EXIF/XMP data. This can be useful as not all software will export the focal lengths.
Uploading to 3DUserNet
In the Project page of the dashboard, select upload Dataset for your project. Then select the Imagery button followed by the Image Groups button. On the next menu select ‘Proceed’ to get to the options menu.
Select Panoramic if the images are panoramas, select the color and add an optional ‘Date of Capture’. For Panoramics and Images where you have a CSV file, select the Text/CSV option, this will open up an area for uploading your CSV file.
Once it has been uploaded, you will be taken straight to the following menu:
The textbox at the bottom is to help the user understand the file that they have loaded.
At the top, you can select a Preset from known manufacturer export formats. If you have generated your CSV through another package, then you will have to fill the form manually.
- Lines till Data: number of lines the program can ignore before the data columns.
- Delimiter: The type of character that separates the columns.
- Data Order: Here you define the column order, please note a single dropdown can cover multiple columns, i.e. columns X, Y and Z are added when you select ‘XYZ’ from the dropdown.
Once this table is complete, click on ‘Proceed’ to take you back to the previous menu.
Once you are happy with your options, select ‘Proceed’ again to go onto the Upload page.
Here you can either upload your images by selecting them all and dropping them in, or you can compress them into a .zip file and upload the compressed file.
Please make sure that your images are in the root of the zip file and that you have no folders in the zip file.
Examples of CSV file for Image Import:
Menu Options:
Examples of CSV file for Panorama Import:
!