Tech To Turn Smartphone Camera Into 3D Scanner Developed

Three dimensional scanning can potentially change the applications of archaeology, architecture, sculpture, fashion, virtual training, security control, environment and what not. Even after having such a wide scope, 3D scanning technology has lagged behind its cousin, 3D printing if we put the recent trends in research under the microscope. However, Taubin's research group at Brown University, have been giving considerable effort to make sure that 3D scanning becomes more and more efficient in the near future.


The existing 3D scanning techniques require synchronization between a camera and a projector. The newly proposed method eliminates the use of a projector, thereby making it as simple as using a smart phone camera to accurately scan 3D objects.

According to Gabriel Taubin, a professor in Brown's School of engineering, one of the things his team has its focus on is producing 3D image capture from low cost components. The technology available in the market today is either too expensive or fails to produce a high resolution image with adequate precision.

The currently available technology uses a method called 'structured light' which is created when a projector and a camera are synchronized with the least possible error. The projector casts a series of light patterns on the objects, and a camera captures several photos of the object which later constructs a 3D image of that image. But this technique has a major constraint. 3D scanning requires both the camera and the projector to be in precise synchronization with each other, something that demands sophisticated instrumentation leading to a high cost.

Taubin's group have created an algorithm which removes the projector from the scene and uses any camera (without very high specifications) with untethered structured light flash. The camera just needs to possess the ability to capture uncompressed images in burst mode (several successive frames per second), which many DSLR cameras and smartphones can do with ease. Their research has been presented at the SIGGRAPH Asia computer graphics conference last month.

The problem in the newly customized mechanism is that projector could switch from one pattern to the other, while the first image is still exposed. This leads to the captured image being distorted, with two or more patterns being present in it. Another problem is that most digital cameras use a rolling shutter mechanism. Rather than capturing the whole image in one snapshot, it captures one pixel row at a time leading to the same problems – a mixed pattern on the resultant image.

Graduate student Daniel Moreno has referred to the above problem as the biggest loophole in this technique. To overcome it, they have written an algorithm which synthesises images - one for every pattern projected, which gives an appearance like that of synchronised capturing.

After the camera captures the bursts of images, an algorithm searches for the timing of the image sequence using the binary information embedded in the projected pattern. Thereafter, the algorithm scans pixel by pixel to assemble a new sequence of images, which captures the entire pattern. Finally, the standard structured light 3D reconstruction algorithm is used to reconstruct the 3D image of the object.

In the SIGGRAPH paper, the team has stated that during testing, the researchers had used a calibrated structured light projector, but results have shown that the process could have been easily implemented using standard digital cameras or even smartphone cameras. Taubin explained that he believes, that this is a significant step towards making photographic instrumentation more economical.

Source: #-Link-Snipped-#| #-Link-Snipped-#