Have you ever wondered how does google 3D map work? Or, have you ever observed 3D objects in games or movies that look so real, it is impossible for you to believe that it was modeled and rendered artificially?
Well, the answer might be Photogrammetry. Photogrammetry is often defined as the science of taking measurements from photographs. In a simpler term, you can scan an object and create your own 3D model through a free software just by taking a few overlapping pictures of the model. Need fancy cameras? No. You can even use your cellphone camera and take some pictures by hand. Benefit of using photogrammetry over laser scanning is that photogrammetry can provide a vrml file format, that means that it would also include the surface colors of the scanned object. Whereas, laser scanning fails to capture the colors of the surfaces and can only provide an stl file with no color data.
Photogrammetry can be classified into two basic types, depending upon the object size and camera location:
Aerial Photogrammetry
This method is generally adapted to capture topographical maps or terrains of a desired location. The camera is usually mounted under an aircraft or drones that would be pointed towards the ground. Multiple overlapping photos of the area are taken which would be later used to generate point clouds in a software.
Close range Photogrammetry
Also known as image based modelling, the camera is located much closer to the object. Basic cameras can be used to model objects such as artefacts, live humans, furniture and even buildings.
Softwares such as Autodesk Recap, Agisoft photoscan, 3DF Zephyr and many more are designed for this purpose. By scanning a series of overlapping images, the computer software can stitch the images together to get a 3D model. The software generates point clouds by calculating depths from various overlapping pictures and then converts them into a 3D mesh object, depending upon which software you are using. My personal recommendation would be to use Autodesk Recap. It is available for free to students and requires the least amount of time and effort compared to other softwares. This software also uses cloud computing, that means that you do not have to rely on the hardware of your local computer. Instead, the pictures get uploaded online and the processing takes place on higher grade servers online, which automatically downloads the model to your computer once ready.
Although the final generated 3D models are proportionate, they are not scaled. Hence, they need to be scaled appropriately with reference to one side. The final scaled model can then be used to measure lengths, areas and volumes and can even be used for fun in VR as virtual models.
To learn more about creating your own 3D models, please refer to the attached workflow of ‘Processing photogrammetry images using Autodesk Recap’, courtesy of Stephanie O'Malley from Emerging Technologies Group, Duderstadt Center, University of Michigan. You can learn more about the program and get a chance to use fancier equipment setup at Duderstadt Center for free by contacting Stephanie at somalley@umich.edu