Canon Is Developing Dual Pixel CMOS 3D Photo Tech
Canon is developing 3D photo technology that combines Canon’s Dual Pixel CMOS image sensors with bespoke software to make 3D data creation more accessible and much more straightforward.
The tentatively named 3D Photo technology relies upon Canon’s unique Dual Pixel CMOS image sensors. While Dual Pixel CMOS most often comes up concerning the Dual Pixel CMOS AF phase-detect autofocus system in many Canon cameras, the technology splits each pixel on the image sensor into two effective photodiodes. Now, for autofocus, this ensures every pixel can perform both autofocus phase detection and image capture.
However, as Canon demonstrates through its original EOS R5 camera and its Dual Pixel RAW mode, these split pixels can perform other tasks. In the case of Dual Pixel RAW mode, the camera captures two separate images simultaneously using the two separate photodiodes. Since they occupy different positions on the sensor, these photodiodes capture slightly different pictures of the same scene.
With Dual Pixel RAW enabled, the camera captures a 3D depth map, which can be manipulated using Canon’s Digital Photo Professional software to perform portrait re-lighting, adjust background clarity, slightly adjust the maximum point of sharpness, shift bokeh, and even correct some optical aberrations like ghosting and flare.
Enter Canon 3D Photo, which takes the idea of a 3D depth map much further. With the Dual Pixel image capture and specialized software, the camera can generate a 3D image with a single shot. This supports viewing from directions close to the camera and works well for near objects. Canon admits it has difficulties with mirror surfaces and transparent objects.
It can do much more when the camera captures multiple images and the subject moves. At CP+, Canon demonstrated 3D Photo using objects on a turntable. Canon’s 3D app on a connected computer automatically controls object rotation and shooting, so the user can just hit a button and let the camera do the rest. 3D data creation starts automatically and finishes once each necessary viewpoint has been captured. With this method, creating a full 3D model of an object is possible, as shown in the video below Canon sent to PetaPixel.
Canon is no stranger to 3D image capture technology. The company has various lenses that can capture stereoscopic 3D photos and videos, including the RF 5.2mm f/2.8 L Dual Fisheye. However, these are specialized lenses, and in the case of the RF 5.2mm f/2.8, they are expensive. The new 3D Photo system does not require a specialized lens. Canon has demonstrated using the RF 50mm f/1.8 STM lens, a $200 prime costing $1,800 less than the RF 5.2mm f/2.8 L Dual Fisheye.
The accessibility of this system could make it especially useful for photographers who need 3D models of products for a small business or online marketplace. Imagine allowing prospective customers to zoom in on and rotate an image of something like a jacket, handbag, or vase.
As of now, Canon’s 3D Photo technology remains in development, and no specific release date is planned. However, the reception at CP+ was positive, which always bodes well for tech demonstrations like this becoming commercial products.
Image credits: Canon
