This sensor takes color photos in the dark

Photography is above all a story of light. When there is little or none, it becomes more difficult for optical sensors. This does not prevent the various optical specialists from floor for a long time on taking snapshots in darkness almost total. Taking pictures in the dark is what people do with difficulty. astrophotographers. It is necessary to have a long exposure time to allow sufficient light to penetrate, or to use an emitter infrared with a rather poor result on the colors.

Cannon has just made a contribution to the building by developing a new sensor with shooting capabilities nocturnal exceptional. The brand has improved a sensor technology called Spad (single photo avalanche diode Where avalanche diode at photon unique). Instead of recording the amount of photons that pixels of the sensor can absorb over a defined period of time, it can multiply the effects of the detection of a single photon. This is why the phenomenon gate the name of waterfall or avalanche. With this process, the sensitivity is ultimately more than 10 times greater than that of a traditional sensor. Canon’s sensor is 3.2 megapixels. This has theair ridiculous when compared to what can be placed on a smartphone, but for taking pictures in the dark, it is the densest sensor in the world.

See colors in the dark

In addition to improving light sensitivity, it can also recognize objects by measuring the time it takes for light particles to reach the sensor. This makes it possible to take three-dimensional object captures, even under conditions of brightness very weak and this in color. This sensor has another advantage: an ultra-fast shutter frequency. On a sensor CMOS of reflex, theshutter activates the pixels one after the other. With the Spad, there is a global shutter that can activate them all at once. This has the effect of reducing the exposure time to 3.8 nanoseconds. Record time for a sensor designed for taking pictures in the dark. With this process, the shooting frequency can peak at 24,000 shots per second!

In terms of applications, this type of sensor will initially be intended for specific areas, such as the video surveillance sector, augmented reality. It could well also be used for self-driving cars. Ultimately, the sensor should integrate a next version of the EOS Ra, a reflex camera appreciated by astrophotographers.

In any case, this sensor is no longer at the prototype level since, in its press release, Canon indicates that these new sensors should be produced in mass from next year. Canon is not alone in working on such a night sensor. This is also the case with Sony and Panasonic who are also looking to produce sensors that could be used by autonomous cars and specialized robots.

Canon unveils sensor that could revolutionize digital photography and autonomous driving

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

Article by Edward Back, published on 1^er June 2021

Last year, Canon unveiled the first Spad sensor (Single Photon Avalanche Diode or in French, single-photon avalanche diodes) of a megapixel, in association with Swiss researchers. In one new article, the firm provides more information on this unique breakthrough which could have repercussions in many sectors.

This new sensor is primarily intended for use as a monochrome photo and video camera. The CMOS sensors, usually used in digital cameras and our smartphones, measure the amount of light in a analog, then convert the result to a digital signal, which creates a lot of “noise”. The Spad works by counting each photon, thus obtaining an image of unparalleled clarity.

Eyes for robots

To achieve a definition of one million pixels without losing sensitivity, Canon has succeeded in exceeding the limits of current Spad sensors. The firm has succeeded in creating square pixels and removing the spaces between the pixels in order to capture all the light. This technology is extremely fast, and makes it possible to reduce the exposure time to only 3.8 nanoseconds and thus to film scenes at 24,000 frames per second. Canon was thus able to create a slow motion clip of a beam of light that moves through smoke.

Finally, this sensor can also measure the photon time of flight, such as ToF sensors in smartphones and the Lidar of the iPhone. This additional information makes it possible to see the scene in three dimensions with a much higher definition than current sensors. This is used in particular to have an almost instantaneous autofocus, augmented reality and autonomous cars. Canon particularly emphasizes the use for robots which could thus see in 3D. What is curious is that Canon chose to release more details a year after the initial announcement. Maybe we can hope to see this sensor integrate a car or a 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 taking 3D photos with a speed and unmatched definition. MegaX is so fast that it can capture the path of individual photons!

Posted on 04/25/2020 by Edward Back

Researchers from the Swiss Federal Institute of Technology in Lausanne (EPFL), in association with Canon, have developed a new kind of camera. Thanks to single-photon avalanche diodes (SPAD), MegaX is able to capture the path of individual photons with a definition of one million pixels.

MegaX can record up to 24,000 images per second, and simultaneously capture very dark and very bright objects. “We can increase the dynamics enormously, which is not possible with a high definition camera” indicated the professor. Edoardo Charcoal.

3D images of several objects simultaneously

The device can also measure the duration of the path of a photon between the object and the camera. This means that it works as a sensor of flight time (ToF) for 3D images, with an application in particular in augmented reality. In addition, MegaX can detect several objects simultaneously. The researchers were able to measure at the same time the distance of a sphere and a transparent plate placed in front, however invisible on a 2D image.

MegaX is not intended for the moment to replace the photo modules in our smartphones. The sensor is much larger, each pixel measuring 9 micrometers, compared to 0.9 micrometers in a traditional camera. “The goal of my research is not necessarily to adapt the MegaX model to a conventional camera, but to create a 4D camera (3 dimensions plus time) with as many pixels as possible in order to obtain a resolution more important ”declared Professor Edoardo Coal.