patent.fyi
Applications  /  US 2026/0355855
United States Patent Application Publication

US 2026/0355855 A1

Published
Nov 12, 2026
Filed
May 7, 2026
Earliest date on the face
May 8, 2025
Classification
B23K 26/032
First published here
Nov 19, 2026

Generated automatically. Not reviewed by a person. Not legal advice or an opinion.

Section 1 · machine-generated

What if?

Questions about how individual parts of claim 1 might be changed. A question may describe something different from what the claim language recites, or a version of it; no view is offered on which.

1.1Single photodiode monitors emission

What ifWhat if a single photodiode measures emission from the machining area?
Claim language“an image sensor for capturing an image”

A collection lens directs light from the machining area through a bandpass filter onto a silicon photodiode. A transimpedance amplifier converts the photocurrent to voltage, and the controller flags sustained drops below a calibrated threshold.

1.1Sequential full field filtering

What ifWhat if a tunable filter applies one passband across the entire sensor during each exposure?
Claim language“a mosaic filter”

An acousto-optic tunable filter steps through programmed passbands while a monochrome sensor captures one frame per passband. A controller associates each frame with its passband and computes band ratios at corresponding image coordinates.

1.2Acoustic monitoring from fixture

What ifWhat if the controller monitors the process using acoustic emission from the workpiece fixture?
Claim language“based on sensor data from said image sensor”

A piezoelectric transducer coupled to the fixture converts elastic waves into a voltage signal. The controller bandpass filters the signal and flags bursts whose energy exceeds a threshold calibrated for the machining operation.

1.3Automatic tracking of machining zone

What ifWhat if a live image determines the selected hyperspectral pixels automatically?
Claim language“based on a user input”

A co-aligned monochrome camera captures the machining zone, and an image processor locates the centroid of a bright region near the beam axis. A calibrated coordinate transform maps the centroid to the hyperspectral sensor, and the controller selects the groups within a fixed radius.

1.3Readout of two bands

What ifWhat if the sensor reads only two detector sites in each selected hyperspectral pixel?
Claim language“read out all pixels”

A randomly addressable sensor reads the two sites whose mosaic passbands match the monitoring bands. The controller subtracts dark offsets, divides one corrected signal by the other when the denominator exceeds a noise threshold, and compares the ratio with a calibrated range.

Section 2 · machine-generated

Variations

Further arrangements of the same subject matter, described here as of the publication date at the foot of this page.

Other field of use

Thermal monitoring of glass edge polishing

A mosaic-filtered image sensor views the edge of a glass panel through a side-mounted lens while a laser polishes the edge. The operator marks a strip around the heated edge on a setup image, and the controller maps the strip to complete hyperspectral pixels. The controller reads every detector pixel in the selected groups and compares the calibrated ratio of two thermal emission bands with a reference range measured on reference panels.

Component substitution

Dielectric filters over the detector array

A patterned array of dielectric interference filters serves as the mosaic over a monochrome detector array. Alternating layers of silicon dioxide and tantalum pentoxide form the filter stacks, with a different cavity thickness for each passband in a repeating group. The controller maps the operator's selection to complete groups, reads every detector pixel in those groups, and applies a gain correction measured separately for each passband.

Parameter or scale variation

Two monitoring windows with unequal dimensions

A 2048 × 2048 detector uses a repeating 4 × 4 filter mosaic and supports two rectangular readout windows in each frame. The operator places a 64 × 64 detector-pixel window around the weld pool and a 128 × 32 detector-pixel window along the cooling track. The controller snaps both window origins to mosaic group boundaries, reads every detector pixel in both windows at 500 frames per second, and calculates separate band averages for the pool and track.

Combination with existing systems

Spectral monitoring with conveyor encoder tracking

A mosaic-filtered camera views the laser weld zone on a conveyor equipped with a rotary encoder and a part-entry photoelectric sensor. The operator selects the weld zone in the camera image and enters the distance from the photoelectric sensor to the weld zone. The controller triggers readout of every detector pixel in the selected hyperspectral groups at encoder intervals and stores the spectral measurements against the part identifier and distance along the weld.

Spectral monitoring paired with surface profilometry

A mosaic-filtered camera views the weld pool, and a trailing laser line profilometer measures the surface of the solidified bead. The operator marks the pool region on the camera image, and the controller reads every detector pixel in the hyperspectral groups that intersect it. The controller tags both data streams with motion encoder positions and compensates for the measured spacing between the viewing locations to pair each spectral measurement with the corresponding bead height and width.

Automated or data-driven variant

Thermal image tracking of the melt pool

A beam splitter directs light from the machining area to a mosaic-filtered camera and a monochrome near-infrared tracking camera. The operator marks a search window in the tracking image, and software locates the melt pool centroid within that window by thresholding the image and measuring the largest connected region. A coordinate transform measured with a calibration target maps the centroid to the mosaic sensor, and the controller reads every detector pixel in the hyperspectral groups around the updated location.

Specific version of a part

Record data for later review

A data recorder stores each readout with a timestamp and the corresponding machine position. After the cycle, a workstation aligns the recorded samples with the tool path for manual review.

Sixteen band filter mosaic

A repeating four-by-four mosaic of thin-film bandpass filters covers a monochrome pixel array. Each complete tile forms one hyperspectral pixel, with one detector site assigned to each of sixteen spectral bands.

Band ratio threshold monitoring

A digital controller subtracts stored dark offsets from two spectral channels and computes their intensity ratio when both corrected signals exceed the noise floor. The controller flags a process excursion when the ratio remains outside a calibrated interval for five consecutive frames.

Tile aligned rectangular readout

An operator draws a rectangle on a preview, and the controller translates its edges into sensor coordinates and expands them to mosaic tile boundaries. The sensor reads every detector site in the resulting window, and the controller computes the mean intensity per band across the selected tiles and compares each mean with its process limits.

Section 3 · quoted from USPTO records

Other work

Passages reproduced word for word from two earlier-published U.S. patent documents, placed beside the language of claim 1. No characterization is offered. Each passage is marked A or B for the document it comes from; ¶ numbers are the paragraph numbers in the USPTO’s published full text of that document.

Document AUS 2008/0123097 A1
Published
May 29, 2008
Filed
Oct 25, 2005
Class
G01J 3/51
Document BUS 2014/0088380 A1
Published
Mar 27, 2014
Filed
Jul 10, 2013
Class
A61B 5/445
Claim 1
Document A
Document B
1.1an image sensor for capturing an image of a machining area of the laser machining process, the image sensor comprises an array of X × Y pixels and a mosaic filter, wherein the X × Y pixels are grouped into N hyperspectral pixels by the mosaic filter, and
The orientation and dimensions of colour mosaics may in a simple embodiment be free, but it is desirable to achieve pixel-wise matching between the a colour mosaic and the image sensor array, and to achieve element-wise matching between a colour mosaic and another one with elements that cover exactly 2×2, 2×3, 3×3 or other number of elements of the first one (the higher resolution mosaic), i.e. the two colour mosaics are exactly matched.
A · US 2008/0123097 A1 · ¶ 136
In one embodiment, the use of the mosaic filter facilitates construction of an image cube from the data collected by the sensor. Construction of the image cube can be performed by iterating over rows and columns of the sensor data and reorganizing the groups of four filters that make up a single pixel.
B · US 2014/0088380 A1 · ¶ 78
1.2a control device for monitoring the laser machining process based on sensor data from said image sensor,
A method for non-invasive detection, identification or characterization of erythema or bruising of skin beyond visual inspection with the unaided eye, the method comprising the steps of: transmitting light at a different wavelengths within a plurality of narrow spectral regions to the skin to illuminate the skin; capturing a plurality of separate spectral images of the skin using a multi-spectral filter array directly mounted to a single sensor; and processing the images to f
B · US 2014/0088380 A1 · claim 1
1.3wherein the control device is configured to: select at least one hyperspectral pixel based on a user input for at least one selected image area, read out all pixels of the at least one selected hyperspectral pixel from the image sensor as sensor data, and monitor and/or control the laser machining process based on the read-out sensor data.
The ROI was then located using the four marks that circumscribed the area of tissue to which the ischemic load was applied. The pixels comprising the ROI were hand selected from the central area defined by these marks.
B · US 2014/0088380 A1 · ¶ 89

An empty cell means only that no passage was reproduced there.

Statements in these documents
A disadvantage associated with the need of scanning procedures, is that the resulting image is sensitive to changes in the scene faster than the scanning time.
A · US 2008/0123097 A1 · ¶ 10
The present invention aims at capturing instantaneous multi- and hyperspectral 2D-images when no filter interchange or exchange is performed, i.e. no spectral scanning of the scene is required.
A · US 2008/0123097 A1 · ¶ 10
More specifically, the present invention is directed to systems and methods for real time multispectral imaging of surfaces and subsurfaces.
B · US 2014/0088380 A1 · ¶ 3
However, skin pigmentation (due to the presence of melanin) can mask the indicators of erythema visible to the unaided eye, thereby hindering clinical assessment of pressure ulcers in people with darkly pigmented skin.
B · US 2014/0088380 A1 · ¶ 7
Cite aspatent.fyi, US 2026/0355855 A1, claim 1, first published 2026-11-19T16:20:06-05:00, https://patent.fyi/US20260355855/, SHA-256 67375a9a4690b9a108a1e6495e04434c59bf22cb9dd0c2eb230201dd6e17c95b
First published
Permanent addresshttps://patent.fyi/US20260355855/
Content fingerprint (SHA-256)67375a9a4690b9a108a1e6495e04434c59bf22cb9dd0c2eb230201dd6e17c95b
Methodv2026-11-18.2 · automated, not reviewed by a person
Correctionscorrections@patent.fyi