Tech Insights
Kavinda Abeyrathne
February 18, 2020

Photogrammetry- The Art of Making 3D models from pictures

Photogrammetry- The Art of Making 3D models from pictures

Since the firstphotograph was taken by Joseph Nicéphore Niépce in 1826 photography and camerashave changed dramatically over the years. Now taking a picture is as simple aspressing or tapping a button on a mobile phone. But there is a whole differentside to photography most of us don’t know: Photogrammetry. What’s this word?What does it mean? Is it the same as photography? Who uses this? Well, let’sstart at the beginning.

What’s this word? What does itmean?

The word“Photogrammetry” is derived from three Greek words: phos or phot,meaning light; gramma, which means “letter” or something drawn and metrein,the noun of measure. This word was first introduced in 1867 by Germangeographer Otto Kersten. The word “Photogrammetry” means the process of makingsurveys and maps using photographs. It is the art, the science, the technologyof obtaining reliable information and measurements about the properties ofsurfaces and objects through photographs.

Is it the same as photography?

Photogrammetry is notphotography. Photography uses cameras to get the final output. Photogrammetryuses photographs to get the final output. With that said, comes the question of“what is photogrammetry then?” Photogrammetry has many techniques, but the keymethods are optical and projective geography methods.

Projective Geography Method

In projective geography,we use special software to detect similar points in photographs and align them.This software extracts three-dimensional data from two-dimensional images. Thedistance between two points on a plane parallel to the image plane can bedetermined by measuring their distance on the image. Also, it extracts colourdata from the images representing diffuse, specular, ambient occlusion andmetallicity so that the re-creation can be textured as an ultra-realisticmodel. Imagine you need to model an old statue in your city centre. To do this,first you have to take a couple of dozen overlapping photographs in an orbitalmanner either from top to bottom or circling the statue with the camera indifferent heights. Thus, you get rings of photographs. By taking overlappingphotographs, it’s easier to triangulate a specific point and calculate thedistance between objects in the photos.


Optical Method

The other method is theuse of optical scanners. A 3D scanner is a device that uses infra-red lightbeams to construct metric data from objects it is scanning. Laser scannersgather plane data using a laser beam pointed at the surface and then convertingthat data into a point cloud. These point clouds are made with software like Agisoft Metashape, Reality Capture and Autodesk ReCap tomake life easier through the pipeline.

The outcome of laserscans is more accurate than photographs. This is because sometimes inphotogrammetry reconstructing smooth, transparent or reflective objects is abit tricky. It’s hard to identify details of those kinds of objects sincethere’s no point to grab on to. This is the only known practical drawback ofphotogrammetry. This can easily be fixed by messing up the surface you’rephotographing, giving it points to hold on to when aligning. The drawback oflaser scanning is that devices (Handheld IR scanners, Lidar scanners) areridiculously expensive.

Using Reality Capture torecreate a statue from photographs - Who uses this?

Today photogrammetry hasmany applications. Aerial Surveying and Mapping, Engineering, Manufacturing,Cultural and Heritage preservation, Police Investigations, Geology,Archaeology, Architecture, you name it. Other than these fields, the mostcommon field is the gaming industry. With photogrammetry, more complex modelscan be created in hours, which may take days or weeks to complete withmodeling. And the outcome is absolutely astonishing. Ultra-realistic modelsthat are used in most video games today are there because of this technology.Battlefield 1, The Vanishing of Ethan Carter, Star Wars Battlefront,PlayerUnknown’s Battlegrounds and Red Dead Redemption 2 are a few of many gamesthat use photogrammetry.

Surveying and Mappingwith drones is the next industry that uses photogrammetry heavily. Creatingmaps for larger fields with drones and photogrammetry is becoming a more commonway of mapping day by day. Images that are taken with a drone are thenrectified with defined control points to get a more precise map with somefitting, trimming and some tweaking.

The first evidence ofphotogrammetry is also related to surveying and mapping a building in 1502 bynone other than Leonardo Da Vinci. He developed the concept ofperspective and projective geometry around 1480 and through that he developedthe map of the town of Imola in Italy for a conqueror named Cesare Borgia.Borgia didn’t have a geological idea about the town he conquered and by employingDa Vinci to make the map, he used his artistic skills for the cutting-edge,groundbreaking creation of surveying history.


Since Da Vinci publishedthe concept of perspective and projective geometry, many mathematicians andscientists have developed and helped photogrammetry to evolve over the years.Even as we speak, this technology is getting better, adapting to serve morefields than it did before.