Canon unveils a sensor that could revolutionize digital photography and autonomous driving

Canon developed the world’s first Spad sensor with one million pixels. This definition will make it possible to take pictures without noise, measure the distance of objects, and will have applications for cameras, robots, and self-driving cars.

You will also be interested

[EN VIDÉO] Underwater art competition
The prestigious 2020 competition brought together photographers from more than 80 countries who competed for $ 45,000 in prizes in nine categories (wide angle, macro, animal behavior, portrait, cold water, dark water, angled worms, marine protection, and art).

Last year, Canon unveiled the first SPAD sensor (Single photon Avalanche dual vacume Or in French, avalanche binaries in Photon Single) from a megapixel, in collaboration with Swiss researchers. In one New articleAnd the The company provides more information on this unique breach that could have repercussions in several sectors.

This is new The sensor Primarily designed for use as a monochrome photo and video camera. The CMOS sensorsIt is usually used in cameras Digital And our smartphones, measure how much Light in a way Analog, Then converting the result into a digital signal, which creates a lot of “noise”. The Spad works by counting every photon, thus obtaining an image with unparalleled clarity.

Eyes for robots

To achieve the definition of one million pixels without losing sensitivity, Canon has pushed the limits of current Spad sensors. The company has succeeded in creating square pixels and removing pixel spaces to capture all the light. This technology is extremely fast, and makes it possible to reduce the exposure time to just 3.8 nanoseconds, and thus shoot scenes at a rate of 24,000 frames per second. Hence, Canon was able to create a slow motion clip for the Beam of light It moves through smoke.

READ  24 million images over 37 years, Google Earth updated

Finally, this sensor can also measure The time of flight of photons, Such as ToF sensors in smartphones and Lidar From’Iphone. This additional information makes it possible to view the scene in three dimensions with much higher accuracy than current sensors. This is used in particular for near-instant autofocus, augmented reality and Autonomous cars. Canon places particular emphasis on the use of robots that can be seen in 3D. Curiously, Canon chose to publish more details a year after the initial announcement. Maybe we can hope to see this sensor integrate a car or robot soon?

This ultra-fast camera capable of photographing photons is impressive

Thanks to a collaboration with Canon, EPFL researchers have created a camera capable of capturing 3D images with an extension Speed And an unparalleled definition. MegaX is so fast that it can capture the path of individual photons!

Posted on 04/25/2020 by Edward Buck

Researchers from Swiss Federal Institute of Technology in Lausanne (EPFL), in cooperation with Canon, has developed a new type of camera. Thanks to single photon diodes (SPAD), Mega Capable of capturing the path of individual photons with a definition of one million pixels.

MegaX can record up to 24,000 photos per second, and at the same time capture very dark and very bright objects. “We can increase the dynamics dramatically, which is not possible with a high-resolution camera,” the professor noted. Eduardo Charcoal.

3D images of multiple objects simultaneously

The device can also measure Duration The path of the photon between the object and the camera. This means that it acts as a sensor for Flight time (ToF) for 3D photos, with a special app in augmented reality. In addition, MegaX can detect multiple objects simultaneously. At the same time, the researchers were able to measure the distance of the ball and the transparent plate placed in front, but it was not visible in a two-dimensional image.

READ  Destiny 2: Symbiote season starts May 11

At present, MegaX is not intended to replace the image modules in our smartphones. The sensor Much larger, each pixel measures 9 ΜmCompared to 0.9 μm in a conventional camera. “ My research goal is not necessarily to adapt the MegaX model to a conventional camera, but to create a 4D (3D plus time) camera with as many pixels as possible in order to obtain Precision More important, ‘declared Professor Eduardo coal.

Interested in what you just read?

Leave a Reply

Your email address will not be published. Required fields are marked *