Granted Patent B2

US 12,272,190 · App. 17/632,472 · Granted Apr 8, 2025

Inventors: Josselin Xavier Coupard (Moissy-Cramayel, FR); Alrick Patrick Michel Jacques Verrier (Moissy-Cramayel, FR)

Assignee: SAFRAN AIRCRAFT ENGINES

G07C5/0825B64D45/00G07C5/0808B64D2045/0085B64D47/08

Abstract

Claims (36)

1. A method of monitoring an aircraft apparatus, comprising:

receiving, from a video acquisition device, a video stream representing at least part of the apparatus;

performing a diagnostic, based on the video stream, of the video acquisition device, the performing the diagnostic of the video acquisition device comprises:

selecting at least two control images from the video stream, separated in time by a predetermined period, and

evaluating, from a result of a comparison of the at least two control images, at least one of a level of quality and a fault of the video acquisition device;

providing a sensor associated with the apparatus, the sensor being configured to generate, over time, a measurement signal representing an operation of the apparatus, the video stream and the measurement signal being synchronized;

in case of detection of fulfillment of a predetermined condition, extracting part of the video stream to form a sequence of interest, the predetermined condition being fulfilled if an anomaly is detected in the measurement signal generated by the sensor;

identifying, in the extracted sequence of interest, at least one image of interest;

comparing a signature, obtained by processing the at least one image of interest by at least one of filtering, adjusting contrast, edge detection, and adjusting colors based on the diagnostic of the video acquisition device, with at least one predetermined signature from a library of known signatures, the library of known signatures comprising a library of faults updated by learning at a time of occurrence of an event considered to be anomalous; and

generating an anomaly detection signal if the result of the comparison meets a predetermined criterion.

2. The monitoring method according to claim 1 , wherein the identifying step comprises pre-processing provided to reject extrapolated images from the video stream.

3. The monitoring method according to claim 1 , wherein the library of signatures is updated using a neural network.

4. A non-transitory computer readable medium comprising program code instructions which, when run by a computer, implement a method comprising

receiving, from a video acquisition device, a video stream representing at least part of the apparatus;

performing a diagnostic, based on the video stream, of the video acquisition device, the performing the diagnostic of the video acquisition device comprises:

selecting at least two control images from the video stream, separated in time by a predetermined period, and

evaluating, from a result of a comparison of the at least two control images, at least one of a level of quality and a fault of the video acquisition device:

in case of detection of fulfillment of a predetermined condition, extracting part of the video stream to form a sequence of interest, the predetermined condition being fulfilled if an anomaly is detected in a measurement signal generated by a sensor associated with the apparatus, the sensor being configured to generate, over time, a measurement signal representing an operation of the apparatus, the video stream and the measurement signal being synchronized;

identifying, in the extracted sequence of interest, at least one image of interest;

comparing a signature, obtained by processing the at least one image of interest by at least one of filtering, adjusting contrast, edge detection, and adjusting colors based on the diagnostic of the video acquisition device, with at least one predetermined signature from a library of known signatures, the library of known signatures comprising a library of faults updated by learning at a time of occurrence of an event considered to be anomalous; and

generating an anomaly detection signal if the result of the comparison meets a predetermined criterion.

5. A monitoring system comprising:

a video acquisition device;

a processing device; and

a sensor, wherein

the video acquisition device is configured for generating a video stream representing at least part of an apparatus to monitor of an aircraft,

the sensor is associated with the apparatus, the sensor is configured to generate, over time, a measurement signal representing an operation of the apparatus, the video stream and the measurement signal being synchronized, and

the processing device is configured for:

extracting part of the video stream to form a sequence of interest, in case of detection of the fulfillment of a predetermined condition, the predetermined condition being fulfilled if an anomaly is detected in the measurement signal generated by the sensor;

performing a diagnostic, based on the video stream, of the video acquisition device, the performing the diagnostic of the video acquisition device comprises:

selecting at least two control images from the video stream, separated in time by a predetermined period, and

evaluating, from a result of a comparison of the at least two control images, at least one of a level of quality and a fault of the video acquisition device;

identifying, in the extracted sequence of interest, at least one image of interest;

comparing a signature, obtained by processing the at least one image of interest by at least one of filtering, adjusting contrast, edge detection, and adjusting colors based on the diagnostic of the video acquisition device, with at least one predetermined signature from a library of known signatures, the library of known signatures comprising a library of faults updated by learning at a time of occurrence of an event considered to be anomalous; and

generating an anomaly detection signal if the result of the comparison meets a predetermined criterion.

6. The monitoring system according to claim 5 , wherein the monitoring system is provided in an aircraft.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (24)