MWIR sensor for flame detection

Patent metadata

Field	Value
Publication number	US-9250135-B2
Application number	US-201213419410-A
Country	US
Kind code	B2
Filing date	Mar 13, 2012
Priority date	Mar 16, 2011
Publication date	Feb 2, 2016
Grant date	Feb 2, 2016

Abstract

A system for detecting a flame. The system may discriminate between a detected hot object and flame. The system may be a camera-like structure incorporating an infrared sensor, a lens, and an element that could filter out some of the long-wave infrared radiation. The sensor may receive radiation of a scene which forms images on the sensor. The images may be provided to a processor that incorporates one or more modules to determine whether a flame is present in the scene.

First claim

What is claimed is:

1. A flame detector comprising:
an infrared sensor having a pixel array with a plurality of pixels;
a lens situated at a first distance from the infrared sensor, where light passes through the lens to the pixel array; and
a long wave infrared filter spaced from the infrared sensor, wherein the long wave infrared filter is situated a second distance from the infrared sensor that is different than the first distance and is configured to filter long wave infrared radiation from substantially all of the light passing through the lens to the plurality of pixels of the pixel array; and
wherein:
infrared radiation from a scene promulgates through the lens and the filter, in either order, to the infrared sensor; and
wherein long wave infrared radiation of the infrared radiation from the scene between the filter and the infrared sensor has a magnitude less than the magnitude of mid wave infrared radiation of the infrared radiation from the scene between the filter and the infrared sensor.

2. The detector of claim 1 , further comprising a camera housing holding the infrared sensor, the filter and the lens in place relative to one another.

3. The detector of claim 1 , wherein the filter permits primarily infrared radiation having less than an eight micron wavelength to promulgate from the scene to the sensor to restrict the spectral range of detection of the infrared sensor to mid wave infrared radiation.

4. The detector of claim 1 , wherein:
the sensor is a bolometer; and
the filter is selected from a group consisting of a sapphire window, AlON material, and a dual band CO 2 and H 2 O filter.

5. The detector of claim 1 , further comprising a processor connected to the sensor for receiving an image from the sensor of the scene.

6. The detector of claim 5 , wherein the processor comprises:
a module for seeking one or more regions of interest in the image that exceed a threshold temperature;
a module for extracting spatial and/or temporal features from each region of interest; and
a module for classifying the spatial and/or temporal features to determine a class of a region of interest; and
wherein a class indicates a likelihood of the region of interest being a flame.

7. The detector of claim 6 , wherein:
spatial and/or temporal features from each region of interest comprise:
flicker rate of the region, movements of the region, growth rate of the region, and/or other spatial and temporal characteristics of the region; and

the processor further comprises a module for assessing the spatial and/or temporal features against similar features of a reference flame.

8. The detector of claim 6 , wherein the processor further comprises:
a module for analyzing a context of the scene by outlining areas of the scene in the image, a module for ranking the areas of the scene in the image according to an amount of risk in having a hazardous flame, and/or a module for confirming a hazardous flame.

9. The detector of claim 6 , wherein extracting features is based on:
histograms of x, y and/or temporal gradients of the image, Z-scores of area variation in time of the region of interest, Z-scores of intensity variation in time of the region of interest, and/or a power spectral density estimate of an intensity in time of the region of interest.

10. The detector of claim 6 , wherein classifying the features is based on:
one or more features extracted from each region of interest, a statistical classification, trained with flame samples and non-flame samples, to determine a class of the region of interest, and/or a support vector machine determining a class of the region of interest.

11. The detector of claim 1 , wherein pixels of the pixel array have an absorber coating applied thereto to facilitate absorption of mid wave infrared radiation.

12. The detector of claim 1 , wherein the second distance is configured to allow the filter to filter light passing to each pixel of the pixel array.

13. A method for detecting a flame, comprising:
capturing a scene as images from light passing onto a bolometer sensitive to mid-wave infrared radiation, wherein the bolometer includes a pixel array;
filtering out at least a portion of existing long wave infrared radiation from substantially all light passing to the pixel array before the radiation reaches the bolometer; and
processing the images to determine whether the images contain a flame; and
wherein the filtering is accomplished with an item selected from a group consisting of a sapphire window, AlON material, and a dual band CO 2 and H 2 O filter.

14. The method of claim 13 , wherein the processing comprises:
confirming whether the images contain a hazardous flame; and
wherein confirming whether the images contain a hazardous flame, is based on:
seeking one or more regions of interest in the image that exceed a predetermined threshold temperature, classifying the spatial and/or temporal features from each region of interest to determine a class of a region of interest, and/or analyzing a context of the scene by outlining areas of the scene in the images; and
ranking the areas according to an amount of risk in having a hazardous flame.

15. The method claim 14 , wherein confirming that the images contain a hazardous flame results in a hazardous flame alarm.

16. A flame detection system comprising:
a camera comprising:
an uncooled infrared sensor;
a lens for focusing an image of a scene onto the sensor; and
an infrared filter, for filtering out long wave infrared radiation exceeding a predetermined magnitude, situated between the scene and the sensor; and

a processor connected to the sensor, the processor comprising:
a flame detection module connected to the sensor and configured to detect and identify a plurality of regions of interest each at different locations in an image.

17. The system of claim 16 , wherein the camera comprises:
a housing having an opening for admission of radiation from the scene; and
wherein the lens, window and/or filter provides a covering of the opening.

18. The system of claim 16 , wherein the processor comprises:
a region of area sequence generator connected to the flame detection module, the region of area sequence generator being configured to group two or more detected regions of interests with detected regions of interest at similar locations in one or more images to form regions of interest sequences; and
a flame dynamic analysis module connected to the generator, the flame dynamic analysis module being configured to differentiate between regions of interest sequences with a flame and regions of interest sequences without a flame.

19. The system of claim 18 , wherein the flame dynamic analysis module comprises:
a spatial and temporal feature extraction module connected to the generator; and
a selected feature classification module connected to the spatial and temporal feature extraction module.

20. The system of claim 18 , wherein the processor further comprises:
a potential hazard map module; and
a hazardous flame confirmation module connected to the potential hazard map module and the flame dynamic analysis module.

21. The system of claim 20 , wherein the processor further comprises:
a distortion removal module connected between the sensor and the hot region of interest detection module; and
a scene context analysis module connected to the distortion removal module and the potential hazard map module.

Assignees

• Cole Barrett E
• Cox James Allen
• Au Kwong
• Larsen Christopher
• Honeywell Int Inc

Inventors

• Cole Barrett E
• Cox James Allen
• Au Kwong
• Larsen Christopher

Classifications

• G01J5/0816

  using attenuators · CPC title

• G01J5/026

  Control of working procedures of a pyrometer, other than calibration; Bandwidth calculation; Gain control · CPC title

• G01J5/0014

  for sensing the radiation from gases, flames · CPC title

• G01J5/0862

  Physics · mapped topic

• G01J5/046 Primary

  Materials; Selection of thermal materials · CPC title