How to create georeferenced maps in RealityCapture - workshop

[@danielkuzev4992]

What a wonderful tutorial. Thank you!

[@progamers2607]

It was a superb presentation, I want to know how a raw image gets height, as in photogrammetry we need to overlap for 60% for images in order to get ariel triangulation, Also how we capture 3d data to make a 3d shapefile

[@CapturingReality]

Hi, thank you for your kind words. Coordinates (in WGS84) are written to the exif data of the images during the flight. Alternatively, you can obtain a flight log with the image names and coordinates and load it into the application. Unfortunately, it is not possible to create shapefiles inside RealityCapture.

[@ignacio_the_coralbiologist]

Hello, i am currently working on a underwater model in RC to hopefully create surface complexity values using ArcGIS pro. Do you recommend anyone who i could message to ask a few questions?

[@ziminger]

Very usefull video. What kind of distortion model do you recomand for this type of project? Taking pictures by drone perpendicular to the ground.

[@CapturingReality]

Hello Michal, you can switch between models, e.g., starting with a simpler default Brown 3 model first, and later change it to Brown 3 with tangential2 to optimize the data.

[@ziminger]

@@CapturingReality ok, thank you for your answer :)

[@KungFuLama]

I have a lot of very big projects on high resolution. This means the jpeg or tiff outputs exceed the maximum limits (2 gigapixel or higher). Is it possible to do accurate and automatic tiling of the orthophoto so we can later load the multiple tiles in a cad system?

[@BowenLiLevi]

I am confused about how to construct the ground control point txt. document. What does the coordinates in it come from? Need I measure them in the location where I took the pictures?

[@CapturingReality]

Hi, there are more options how to create GCP file. But basically yes, they need to be measured on the capturing place. For not so precise georeferencing you can use also information taken from the online map for example.

[@bertrand5115]

Hi, thank you for the nice tutorials. Do you have any image set to try it out ourselves ? Cheers,

[@CapturingReality]

Hello, you can download a sample dataset from our website, e.g. Drone imagery with ground control points: www.capturingreality.com/SampleDatasets

[@suhasm8779]

I dont have map wizard icon in my software .. can u please help me out

[@TheYagga]

Hi Suhas, please let me know when you figure it out.

[@CapturingReality]

Hi, please excuse our late response. It should be located in the Workflow tab. Which version of RealityCapture do you use? You can check it in the About dialog once you click blue question mar in the upper-right corner of the application.

[@Abrimaal]

I tried this and every time I got a warning message "the images do not cover 75 percent of the area". I want to skip this and create a map anyway.

[@CapturingReality]

You can create a map also without Map Wizzard. But you need to do a whole process: Alignment, Model creation, Texture, Ortho creation. In which step did you get this warning?

[@Abrimaal]

@@CapturingReality I get this message after importing large components, starting map wizard, selecting area manually on map, then Start. There may be missing parts between imported components, but I add them later, as they are done. I have done whole streets, fragments of streets and architectural objects. I want to add them gradually to a map, to keep the same scale for all models. Is this the correct method?

[@CapturingReality]

@@Abrimaal There could be a problem with missing parts. Are you using georeferenced data? Is it possible to merge these components before using Map Wizard?

[@Abrimaal]

@@CapturingReality the most of images are video frames, without GPS data, but there are always photos taken with the same camera that contain GPS coords. I don't need high quality models now, and processing from video already creates large components even without control points. I want to merge as many as possible, but it takes whole nights to rebuild and recolorize a single model made from 2000 or 3000 images. I just need a "playground", a planar map of the whole town, where I could place objects and build a virtual town. For various purposes, documentary, preservation of moments and maybe a racing track for games. Reached the expert level in making photo panoramas, I try one level up, into 3D. This is the same sequence: images, control points and the panorama projection plane is actually also a 3D model, where the images are projected. Time will tell.

[@jamesgjurovich7893]

please no uuhs very distracting and annoying

[@kusonsin]

good point for her improvement (while at it Wizard, not Wizzard) , but nonessential/nonconsequential to what was presented. Good video it was