Sony Patents a Method to Make Cameras Focus and Expose More Accurately Using Two Screen Patterns
Getting a camera to see the world accurately requires a careful setup process called calibration, and Sony thinks flashing two different dot patterns on a screen one after another, then combining them, beats the old way of doing it.
What Sony's time-split calibration grid actually does
A factory technician points a camera at a flat screen covered in a grid of dots. The camera captures the grid, and software uses it to figure out how the lens bends and distorts light, correcting those flaws forever after. If the grid doesn't have enough dots, or they're arranged poorly, the correction is less accurate. That's the everyday problem this patent tries to fix.
Sony's idea is to show two different dot patterns on the screen one at a time, a fraction of a second apart, then stitch the reference points from both frames together as if they all appeared at once. You end up with a much denser, richer map of reference points than a single static grid could give you, even though your screen never changed size.
The screen is also tilted at an angle to the camera on purpose. That angle, combined with the combined point map, gives the calibration software more geometry to work with, which translates into more accurate lens correction at the end of the process.
… generates integrated feature point information by integrating a feature point at a first time of a chart pattern at the first time and a feature point at a second time of a chart pattern at the second time on a single plane …
Translation: It combines marker points from two different screen displays into one unified flat view.
How the system merges two pattern frames into one reference map
Camera calibration works by showing the camera a known reference pattern, usually a checkerboard or a grid of dots, and then doing the math to figure out how the camera's lens distorts the image. The more reference points you have, and the better they're spread across the frame, the more accurate the final correction.
Sony's patent describes a system that breaks the reference grid into two separate patterns and flashes them on a display screen one after the other (a technique called time-division display). Each pattern has its feature points (the recognizable dots or corners the software locks onto) arranged differently. The software then overlays both sets of feature points onto a single virtual plane, as if both patterns existed simultaneously.
The display panel is deliberately tilted at an angle relative to the camera's line of sight. This tilt matters because it forces the calibration math to account for perspective, giving the software more information to pin down the lens geometry. Flat-on grids can leave certain distortion parameters underdetermined.
- Image acquisition unit: grabs each video frame as the screen switches patterns.
- Feature-point-information generation unit: identifies the key reference points in each frame and merges them into one combined reference map.
- Parameter acquisition unit: runs the calibration math on that combined map to produce the final correction values for the camera.
… a chart pattern displayed in a time division manner on a display panel inclined with respect to an optical axis direction of the camera …
Translation: The screen shows changing test patterns while tilted at an angle to the camera lens.
What this means for cameras that need precise factory tuning
Lens calibration is invisible to most people but affects every camera that ships in a phone, a car, a security system, or a medical device. A poorly calibrated lens means straight lines look curved, distances look wrong, and any software that depends on the camera's geometry (like a depth sensor or an autonomous-driving vision system) gets fed bad data from day one.
By pulling more reference points from a single screen without needing a physically larger chart, Sony's approach could make factory calibration faster and more precise, especially for cameras going into tight spaces where a big printed chart won't fit. Sony's pattern of camera-accuracy filings suggests the company sees calibration as a meaningful competitive edge in its imaging hardware lines. For the end user, the payoff is a camera that simply gets the geometry right from the moment it leaves the factory.
Sony's latest display application is the 79th we've tracked since May, adding to a set that includes daisy-chained display chips and gaze-directed remote cameras in our Display patent coverage.
Lens calibration goes wrong. A doorbell camera that stretches doorframes, a car's parking assist that misjudges curb distance by a finger's width: these failures trace back to a reference image that didn't give the camera enough information to learn from.
Sony's approach here splits a dense grid of reference points across two back-to-back frames, then stitches the results together as if both sets appeared at once. The camera gets twice the reference data without needing a physically larger or more expensive calibration screen.
For someone buying a car with self-parking sensors or a surgical imaging system, this matters because it removes a source of accumulated error before the product ever ships. You would never see the correction working, only the absence of the small, costly mistakes it prevents.
Get our take in your Top Stories
Liked this breakdown? Add Patentlyze as a preferred source on Google, and our plain-English take shows up more often in your Top Stories the next time Sony patent news breaks.
There are more where this came from
We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.
The drawings
26 drawing sheets from US 2026/0260381 A1 · click any drawing to enlarge
Show the other 18 sheets
Want this weekly breakdown for a company we don't cover? Patentlyze Pro →
Be the first to weigh in