LIDAR - Rethinking how cameras work

One cool feature of Apple's new iPad Pro tablets is that they include a new 3d sensor in the camera that utilizes LIDAR. The LIDAR sensor allows the camera to compute depth information for an image in addition to the normal 2D RGB pixel data that you get from a digital camera.

So how does LIDAR work? LIDAR stands for Light Detection and Ranging. It uses a pulsed infrared laser to sweep a spot of light across the area that the camera is looking at. The pulsed laser light is used to measure range (the variable distance for a point in the captured 2D image). Think of it this way. Your typical digital camera gives you a 2-dimensional pixel image of a 3-dimensional space. LIDAR provides a way to get the 3-dimensional distance information for that pixel. So the digital photo is capturing and recording the depth information for the pixels of the image. Your camera is now a 3D scanner as well as a 2D imaging system.

Of course it's more complicated then that, since the camera's LIDAR resolution is not the same as the pixel resolution. You can think of the LIDAR image as a series of small light dots that are super fast swept across the scene you are photographing. How fast? Crazy fast. Hundreds of picoseconds.

IFixit has an interesting iPad Pro teardown where they dive into what the LIDAR sensor looks like, and what it's resolution actually is. And it's a very course resolution. The infrared footage of LILDAR scanner in action at 1:43 in the movie shows you exactly how coarse the pulsed laser dot depth scanning actually is.

Any one familiar with the Microsoft Kinect systems they were selling as an attachment to Xboxes will realize that this all sounds very familiar. And digital artists loved Kinect and what they could potentially do with it.

We're very interested in AR (Augmented Reality) applications here at HTC. Consumer tablets and cameras that actually let you directly work with depth information in digital camera photos is the new wild west. Very exciting.

We're looking forward to providing easy to use interfaces and information on for working with this new technology in our HTC member toolkit.

Comments

Post a Comment

Popular posts from this blog

CycleGAN: a GAN architecture for learning unpaired image to image transformations

Image
 Like the Doublemint Twins touting the joys of Doublemint gum, 2 GANs are surely better than 1 GAN.  Especially if we package them together inside of one meta GAN module.  And this is exactly what the CycleGAN architecture does. Have you ever harbored dark secrets of turning a horse into a zebra?  The CycleGAN was developed to do just that. Learn how to turn a horse into a zebra.  And more. Now right away you can notice a difference between the image to image transformation GAN architectures we've been discussing over the last few posts.  Those last few posts described systems that learn from a database of matched input-output image pairs. And if your goal is to turn an edge representation into a nicely filled in continuous tone image, it's easy to build your database of matched input-output image pairs that your GAN system can then learn off of.  Take a continuous tone image (which will be the output of the database pair entry), then run it through an edge detector algorithm. 
Read more

Pix2Pix: a GAN architecture for image to image transformation

Image
 I thought following up yesterday's TraVelGAN post with a Pix2Pix GAN post would be useful to compare what is going on in the 2 architectures.  Two different approaches to the same problem. I stole this Pix2Pix Overview slide below from an excellent deeplearning.ai GAN course (note that they borrowed it from the original paper) because it gives you a good feel for what is going on inside of the Pix2Pix architecture.    Note how the Generator part is very much like an auto-encoder architecture, but rebuilt using the U-Net architecture features (based on skip-connections) that fastai has been discussing in their courses for several years before it became more widely known to the deep learning community at large  (and which originally came from an obscure medical image segmentation paper) . So the Generator in this Pix2Pix GAN is really pretty sophisticated, consisting of a whole image to image auto-encoder network with U-Net skip connections to generate better image quality at highe
Read more

Smart Fabrics

Image
"What makes smart fabrics revolutionary is that they have the ability to do many things that traditional fabrics cannot, including communicate, transform, conduct energy and even grow."  This is a quote from Rebeccah Pailes-Friedman, who is a professor at Pratt with a research focus on developing wearable technology and smart textiles. Enhancing textiles can be viewed through different lens.  Enhancements can be aesthetic, or performance boosting, or durability boosting. 
Read more