Interesting Nikon patents (curved image sensor and more)

Nikon's curved image sensor patent (Google translation) will improve the sharpness in the corners of the image:

  • Patent publication number 2010-109096
    • Release date 2010/05/13
    • Filing Date 2008/10/29
  • Using a lens with field curvature aberration
  • The image plane conversion element, sandwiched between the lens and image sensor
    • Each pixel has a single optical transmission pipe
    • Element corresponds to the surface curvature aberration
  • As a countermeasure light falloff at edges, with a filter that only the central dimming

Nikon patent application 20100150539 is for a focus adjustment solution (see wikipedia's explanation of the two different passive focus solutions: phase detection and contrast measurement). Different implementations are shown in this patent application: the first drawing is a camera with a half mirror, the second is mirrorless (EVIL) camera, the third implementations is without a prism (Pellix camera?) and in the last drawing the sensor is located under the mirror (instead of behind the mirror):

"Field of the invention: The present invention relates to a device that adjusts the focus of an imaging optical system, an imaging device equipped with the focus adjustment device and a focus adjustment method.

Description of the related art: There is a digital camera known in the related art (see Japanese Laid Open Patent Publication No. 2001-281530) that includes a focus detection sensor adopting a phase difference detection method and an image sensor. This digital camera first drives the imaging optical system to a point near the focus match position based upon the results of detection executed by the focus detection sensor, then detects the focus match position through a contrast detection method by using the output from the image sensor and executes fine adjustment to the focus match position. However, the contrast detection method and the phase difference detection method adopted in the camera in the related art described above are not perfectly compatible with each other since their focus detection principles are different and, for this reason, the correspondence of the focus detection results obtained through the two methods tends to be poor. For instance, there may be a photographic subject for which the focus can be detected successfully through the phase difference detection method but the focus detection cannot be executed successfully through the contrast detection method. In addition, while the focus may be detected with a high level of accuracy through the phase difference detection method for a given photographic subject, the accuracy of the focus detection for the same photographic subject executed through the contrast detection method may be poor. Under such circumstances, the imaging optical system may hunt for, and fail to reach the focus match point, the imaging optical system may be driven to an unexpected position or the imaging optical system may not move at all when the camera attempts to fine-adjust the position of the imaging optical system through the contrast detection method after driving the lens to a point near the focus match position through the phase difference detection method."

The half mirror drawing is shown also in patent application US20100149361.

Patent Application 20100149649 is for a zoom eyepiece lens system.

Patent applications are not guarantee for upcoming products.

This entry was posted in Nikon Patents. Bookmark the permalink. Trackbacks are closed, but you can post a comment.
  • aetas

    A Lens with even less light falloff would be great. I like hearing about nikon glass getting even better.

  • Very interesting stuff.

  • Discontinued

    It’s about time for the curved image sensor. Same for a focus adjustment system, that would spare us things like AF-Fine-Tuning. What I don’t get is, that it took 20 years to solve “all” the problems, that came with inventions like AF and digital. Pros could choose between not less than 7 (!) brands in the past, all capable of meeting professional requirements for a pro 35mm body and for serious wide angles, teles, shift-lenses and such. Ever since we are waiting for less and less companies to provide us with what used to be common for manually focused shots on film. We’ve been waiting for lenses, for actually working AF, for less noise, more DR, more RES and for sharp corners – and since there are no longer 7 brands around, that could be taken seriously, we’ve all turned into stupid Nikon vs. Canon debaters. I guess that’s called progress. So be it.

    At least nobody needs to mess with chemicals anymore and I never found much joy in scanning either. So don’t get me wrong. I don’t cry a single tear for the “good old times”. What bothers me is, that the good new times always get announced ages before they actually truly arrive. I guess that’s called marketing. So be it.

    • Eric Pepin

      I wish i needed to mess with chemicals. Ill be building a darkroom once my new basement bathroom is finished. Well ill be using the bathroom as my darkroom 😛 Also, film > digital, period.

      • Great

        digital has more DR than film for a while now. AF has been working for ages too. can’t comment on noise etc, never had ISO3200 film before.

        • Victor Hassleblood


          how does a shift lens and a curved sensor go together? Think again.

          never had a T-MAX? Too young for analog I guess. Anyway, I think I’m getting what Discontinued was trying to say. How long did we have to wait for FF and other solutions? It took ages for digital systems to become as universal as analog photography used to be.

    • PHB

      Using a curved image sensor is obvious.

      How to make one reliably is far from obvious. In fact that is very difficult indeed and I would be very interested to see how it was to be done…

      Only way I can think of doing it would be to machine the silicon and grind it…

      This does not sound to me like something for a mass market camera. Not unless there is some serious production engineering as well..

    • gssse


  • Seppl

    Ahm, well interessting, the curved sensor.
    BUT: All Nikon DSLR lenses + all 3rd party lenses are designed for flat sensors. So that means, the old glas is useless for cameras with this new sensor design.

    • donde?

      Which is probably what Nikon wants: Their old stuff becoming out of date because they’ve got something that is even better.

      But maybe it works, it seems there’s a lens before the sensor that will project the image on the curved field.

      • Discontinued

        the single lens-element symbolizes a lens and the curved sensor is drawn exaggerated. It’s meant for existing lens designs and their problems with flat sensors. At least this is how I understand it.

    • litebyte

      It totally depends on how the data from the sensor is handled, nothnig to do with the lens itself.

  • Martijn Bouius

    it was filed in 2008, and still not used to date?
    why wait?

    • zack

      you’re joking right? It’s NIkon, 2008 is like two days ago for them

  • zzddrr

    Very interesting patents. It seems none of them has mechanical shutter.

    On the other hand, it wouldn’t be bad if Nikon would actually deliver something from these great patents because they do look great on paper…

  • thinmac

    Curved sensors sound awesome and all, but what I want to see is the patent on how to fab curved integrated circuits. ICs (like digital camera sensors) are printed onto big flat silicon disks. Nikon can come up with neat ideas like curved sensors, but they’ll have to come up with the actual implementation of silicon lithography onto a curved surface before it will do anyone any good.

  • R

    Don’t expect to see this before 2016. Barely working in university labs right now, and not economic until there’s a breakthrough in printed semiconductors.

  • enesunkie

    Oh, an ad for a Canon 7D. How appropriate. 🙂

  • Johan Krüger-Haglert

    How can they accept a patent for a curved sensor? It’s and obvious idea and we are plenty who have thought about it earlier. The problem isn’t imagine it, the problem is to manufacture it. It’s at that stage there can happen some innovation.

  • Hendog

    The patent is not for an actual curved sensor chip, but a FLAT sensor surface with a curved “transmission” element which is probably made of bundles of optic fibres (corresponding to number of pixels) to transmit the light falling on the curved surface of the optic fibre bundle to the flat sensor.

    I imagine that this sort of technology would be implemented in compact digital cameras not dslrs. I heard of a similar design being developed for mobile phone cameras about a year ago, but that was for an actual curved sensor chip which would surely involve much higher production costs. Nikon’s idea seems like a better solution.

    It’s about time a curved image plane was seriously considered, as it would simplify lens design and improve quality. An example of technology inspired by nature – the human eye’s image plane is curved.

  • Alain2x

    What a stupid joke !

    And how about curved screens ?

    Or curved print papers ?

    A flat object gives a flat image, with a normal lens.

    Any other stuff is junk, that’s all. No need to create expensive screens or sensors to solve basic optical problems.

    I’ve never read worse news, please keep this for april’s fool day.

    • Hendog

      Alain, you’re sadly mistaken and ignorant to the fact that producing a FLAT projection to a surface (image sensor) by a lens is NOT a natural tendancy for a lens element. A flat image from a flat subject requires a hell of a lot of optical trickery and complexity, and even then we still we have edge softness, chromatic aberrations etc…

      Look at nature – the ultimate image forming device we know of – the human eye (curved sensor!). It is very simple and it doesn’t need 10+ optical elements to produce a high quality image. A single optical element will want to produce a curved image. This is what it’s about.

      The huge gains in optical performance and simplicity of lens designs would be profound. But it’s unlikely you’ll see a curved sensor in a DSLR due to the redesign of lenses required.

    • plug

      The Earth is flat, right? It makes fall off at the edges very dangerous!

  • Nicole

    I’m sure that somebody will correct me if I am wrong, (They probably will even if Im right too! 🙂 ) but wouldn’t the curvature need to be different for different focal lengths? This would surely limit the practical uses of this design, as it would not be suitable for zoom lenses. Otherwise, it’s a neat solution to an age old problem. 🙂

    • rhodium

      That was exactly what I thought. Different lenses have different curvature of fields, so how would the sensor change to cope?

  • CEMZ

    We could hace a curved sensor with a **fixed** curvature field, so all the lenses will have to be redesigned to match that curvature field, or…

    We could have an **adaptive curvature field** sensor that will change its curvature depending on conditions and lens design. It could adapt its curvature to existing lenses, but no one will want to use the old lenses because…

    We will have **adaptive lens** designs that will also adapt their curvature to dynamic conditions and will drive the changeable curvature of the sensor. That’s really how the eye works, so no moving parts will be required for focusing.

    And think about the endless possibilities of defining how we want the images: sharp from corner to corner, selective fuzzing of corners…

    And then adapt the processing software to behave more like the brain, for example, different ISO sensitivity in different areas, to compensate for highlights or shadows, a non-square image (just how the eye sees) and…

    The final product could also be a curved print, rigid non deformable, or even better a curved screen of a 3D projection with two of those curved sensors…

  • Ray

    Isn’t the M9 already using a similar form of technology to be able to make a FF and use them with the M lenses?

    Not really a curved sensor but microlenses on top of the sensor set in a curved manner or something like that if I am not mistaken.

  • Alan

    The first camera I ever used was my parents’ Kodak Brownie that took 120 negative film [exposure was too hit and miss for transparency film – besides as a young child I hadn’t even heard of it!]. Interestingly, the film plane was curved in one direction to help out a very basic one element lens.
    Going back even further, was the eye ever patented?

  • Wierdo

    This maybe the sign of NIKON D5″C”

    new sign for the flagship camera in 2014 🙂

    if they need new lens mount and type, maybe that’s why they been pushing hard on the lenses nowadays, with new stuff comming out and future release such as 85 f1.4 or 35 f1.4 and a bunch of VRII tele zoom just to get your money now and crushes you again when it comes out 🙂

    or the sensor takes the pic with the current lens and anti distort them automatically with only nikon lenses as they put all the distortion adjustment for only nikon lenses, and with new one comes out the camera software will get update frequently and it connects to your wireless Internet for fast access 🙂

    I should work as a idea person… He he he

    • Wierdo

      This way the can monopolize the lens market, so third party like sigma and tamron have no way to get their product use unless they pay royalties to nikon to put their data inside their software update! 🙂

      the EVIL will be their camera for video instead of pushing it for the apc-s or fx camera so this way they get more consumer attraction for their EVIL, it is a great idea since it will be mirrorless and it have interchangeable lenses, beside 3/4 is enough for a 1920 res movie.

  • semaj

    Maybe a sign of digital rangefinder, hopefully.

  • Huggs

    The curved sensor makes sense. The mechanics of photography mimic the human eye.

    Now, would it be silly to ask for a round sensor?

    • CEMZ

      An elliptic sensor would emulate even better the feel of human seeing. Or stereo round sensors.

  • Leica M8 had curved sensor (really slighlty inclined photodiodes) and worked like a charm with legacy lenses.

  • jr

    CMIIW but isn’t the curvature between lenses are different ? e.g. I think wide angle lenses has more curvature than telephoto lenses. Will there be sensor curvature adjustments when we change from telephoto to wide angle? And how about wide to telezooms?

  • Ok, before everybody goes all off on the brilliance of curved sensors, here’s the biggest hurdle: current lens design is for flat sensors.

    Curved sensors are only practical with lenses designed for them. Maybe such lenses will be cheaper to produce. But it’s water that has been largely unexplored. What that means is much higher R&D costs to start with.

    We are likely to see such things in compact cameras before we see them in DSLR type cameras. Reasons being: that’s where the big money is and small sensors of any design are cheaper to produce.

    Curved is revolutionary and the wave of the future, but they aren’t going to appear on the scene tomorrow.

    There was a good article about them in either Science or Nature within the past couple years. Sensor design is still in it’s infancy.

  • Roger

    I’m still waiting on a patent for a sensor that has perfect QE and zero read noise. Come on, Nikon!

  • Back to top