Camera Obscura Images Can be Collected From Any Windowed Room

The camera obscura has been around for a long time (Middle Ages long) and typically consisted of a box or room with a hole in one side through which an image of its surroundings could be formed. As you can see from the example above, any room — in this case a bathroom — can be turned into a camera obscura given a small enough “aperture.” Unfortunately, most rooms have big, blaring windows that let in too much light, and the only image formed on the opposite wall is a shadowy blob.

In the name of forensics, however, Antonio Torralba and William Freeman from MIT have discovered a technique by which they can turn any windowed room into a camera obscura, using a couple of stills of the room to magically gather an image of the outside world.

You’ve probably seen similar things done in TV crime dramas before (enhance!), but this is one case where CGI may have to make room for reality. All they need is two images of that blob of shadows on the wall, one of which has an object placed in front of the window to block out some of the light. After that they can use software to analyze the images and significantly sharpen the projection into a distinguishable picture of the world outside that window.

The forensic implications are vast, and could even include helping to locate hostages using two stills from a video of the victims. But whatever the uses end up being, it looks like the camera obscura is making a comeback.

(via New Scientist via Engadget)

Image credit: Camera obscura by 1banaan

  • VelazquezRosalie

    my roomate’s st ep-mot her brought ho me $1 3342 the previous mo nth. s he g ets pa id on the computer and bought a $369300 condo. All she did was get lucky and profit by the advice uncove red on this web page

    ⇛⇛⇛⇛► (Click On My Name For Link)

  • stanimir stoyanov

    I don’t get it–how do they recover/sharpen details from the dispersed light (“shadowy blob”)?

  • GlossGreen

    That is very cool. Another example of science in fiction that is now science in fact.

  • DL Cade

    The software compares the two images, and removes the light rays from the image with the object in the way. That way it takes away many of the erroneous, unnecessary light rays and sharpens blob into a distinguishable image.

    The New Scientist source link at the bottom goes into more detail, hope this helps! 

  • Stephan Zielinski

    Folks wanting more details, there’s already a PDF of the paper up: Accidental pinhole and pinspeck cameras: revealing the scene outside the picture.

  • Medevac71

    Well, sure you can make any windowed room a camera obscura, but you need a bright wall color/ best: white surface. It really makes a different

  • 9inchnail

    I still don’t get how you’re gonna actually apply this in a scenario like mentioned in the article. Let’s take this hostage situation: You get a video or a photo of a hostage being held somewhere. Your goal would be to get a vision of the world outside that room to find the location. BUT you need a second photo with an object blocking the light… where are you gonna get that? Politeley ask the captors if they could make one for you? One of them would have to walk past the window while the video is taken to actually use this. And if you held someone hostage somewhere, wouldn’t you close the curtains? 

  • Guest

    did that in the dorm lounge a few years back. the image wasn’t crisp per se, but I wouldn’t call it a shadowy blob.

    to make it less of a shadowy blob, wouldn’t you just.. you know, cover most of the window except for the aperture hole?