Caltech Develops Camera Chip To Turn Your Smartphone Into A Powerful 3D Scanner
Smartphones are capable of so many wonderful things that go beyond the realm of simply making phone calls. They're essentially handheld PCs, and one day soon, they'll also be companions to your 3D printer -- you'll be able to snap a picture of an object using your phone's special 3D lens, which will capture the necessary information for your 3D printer to make a real-life replica.
This is what researchers at Caltech are working on. It's a new high-resolution 3D image sensor that can make an exact copy of an object for a 3D printer. When you snap a photo of an object you want to create, the sensor produces a high resolution scan that measures the object's height, width, and depth. Here's a look at a 3D image of a penny taken with the new sensor:
Image Source: Ali Hajimiri/Caltech
Scanning objects is already possible and has been for some time now, but the equipment necessary is too large and expensive for smartphone integration. That's where a new device called a nonophotonic coherent imager (NCI) comes in to play. It's relatively inexpensive, compact (less than a millimeter squared), and highly accurate.
The camera in whatever smartphone you're wielding right now records pixels that represent the intensity of light received from a specific point in the image, whether it's near or far. There's no information about the relative distance to the sensor. Caltech's NCI, on the other hand, provides both the intensity and distance information of each pixel.
It's not voodoo magic, just fancy technology at play. The chip uses an existing detection and ranging technology called LIDAR, which entails lighting up an image with scanning laser beams. Measurements are then taken of the light that reflects off the object and analyzed based on wavelength to detect its size and distance.
So, when you can expect this to become a mainstream thing? On the bright side, there already exists a proof of concept of the NCI. Unfortunately, it's only 16 coherent pixels, which means that the 3D images it produces can only be 16 pixels at any given distance. Ali Hajimii, one of the researchers working on the project, says it can scale to hundreds of thousands of pixels, but didn't say how long it will take.
This is what researchers at Caltech are working on. It's a new high-resolution 3D image sensor that can make an exact copy of an object for a 3D printer. When you snap a photo of an object you want to create, the sensor produces a high resolution scan that measures the object's height, width, and depth. Here's a look at a 3D image of a penny taken with the new sensor:
Image Source: Ali Hajimiri/Caltech
Scanning objects is already possible and has been for some time now, but the equipment necessary is too large and expensive for smartphone integration. That's where a new device called a nonophotonic coherent imager (NCI) comes in to play. It's relatively inexpensive, compact (less than a millimeter squared), and highly accurate.
The camera in whatever smartphone you're wielding right now records pixels that represent the intensity of light received from a specific point in the image, whether it's near or far. There's no information about the relative distance to the sensor. Caltech's NCI, on the other hand, provides both the intensity and distance information of each pixel.
It's not voodoo magic, just fancy technology at play. The chip uses an existing detection and ranging technology called LIDAR, which entails lighting up an image with scanning laser beams. Measurements are then taken of the light that reflects off the object and analyzed based on wavelength to detect its size and distance.
So, when you can expect this to become a mainstream thing? On the bright side, there already exists a proof of concept of the NCI. Unfortunately, it's only 16 coherent pixels, which means that the 3D images it produces can only be 16 pixels at any given distance. Ali Hajimii, one of the researchers working on the project, says it can scale to hundreds of thousands of pixels, but didn't say how long it will take.
