Computer Vision, ReCap, and Photoshop

Google generates their surface for Google Earth using a concept called computer vision.  Those of us not on the infinite-internet-funded budgets don’t have the pleasure of using that costly equipment but essentially the computers semi-automate the construction of a 3D surfaces, or meshes from flights.  The tech has advanced enough that mapping materials and textures has become more streamlined as well.  The rest of are stuck with a more budget-friendly (and labor intensive) approach to generating a 3D surface for our modeling and presentation purposes.   ReCap is software developed by Autodesk that can help to generate computer models from aerial photography, especially useful for someone with a a drone and the right permits.  Permitting for urban areas of major cities can be delayed by a year or more though, so many of us resort to manually modeling our surfaces the best we can in time constraints using whatever methods we can. One such method is using Recap and Google Earth.  By recording a 360-degree video rotating a surface in google earth we can convert those video frames to images and upload them to ReCap for processing.  Manually modeling high-priority buildings is best, but for low levels of detail (outlining areas and background), this has potential.  During my escapades I found Photoshop opens videos, allows me to crop them however I wish, and then render the frames to images. There are other methods but the simplicity in cropping and converting has made this my favorite.





Leave a Reply