Atmospheric chemical species detection using multispectral imaging

Patent metadata

Field	Value
Publication number	US-12217502-B2
Application number	US-202217804815-A
Country	US
Kind code	B2
Filing date	May 31, 2022
Priority date	Mar 30, 2022
Publication date	Feb 4, 2025
Grant date	Feb 4, 2025

Abstract

Techniques for optically detecting a subject chemical species within an atmospheric environment are disclosed. Image data is obtained representing multispectral imagery of a geographic region captured through the atmospheric environment. The image data includes an array of band-specific intensity values for each of a plurality of spectral bands, including a sample spectral band having increased sensitivity to the subject chemical species as compared to a plurality of reference spectral bands. A background reflectance map is generated that includes an array of inter-band intensity values in which each inter-band intensity value represents a filtered combination of band-specific intensity values of albedo-normalized arrays for a grouped subset of the plurality of reference spectral bands. The albedo-normalized array of band-specific intensity values for the sample spectral band is compared to the background reflectance map to obtain an index array of intensity variance values for the subject chemical species.

First claim

The invention claimed is:

1. A method performed by a computing system for optically detecting a subject chemical species within an atmospheric environment, the method comprising:
obtaining image data representing multi spectral imagery of a geographic region captured through the atmospheric environment via an aeronautical vehicle,
the image data including an array of band-specific intensity values for each of a plurality of spectral bands, including a plurality of reference spectral bands and a sample spectral band having increased sensitivity to the subject chemical species as compared to the plurality of reference spectral bands;

removing albedo background from the image data for each of the plurality of spectral bands to obtain a normalized array of band-specific intensity values for each of the plurality of spectral bands;
generating a background reflectance map that includes an array of inter-band intensity values in which each inter-band intensity value represents a filtered combination of band-specific intensity values of the normalized arrays for a grouped subset of the plurality of reference spectral bands;
comparing the normalized array of band-specific intensity values for the sample spectral band to the background reflectance map to obtain an index array of intensity variance values; and
outputting the index array for the subject chemical species.

2. The method of claim 1 , wherein the subject chemical species includes methane; and
wherein the sample spectral band has a central wavelength within a range of 2093.2 nm-2289.9 nm.

3. The method of claim 1 , wherein for each inter-band intensity value, the filtered combination is an average of the band-specific intensity values of the grouped subset of the plurality of reference spectral bands.

4. The method of claim 1 , further comprising:
for each inter-band intensity value of the background reflectance map, classifying the plurality of reference spectral bands into the grouped subset using a clustering algorithm.

5. The method of claim 4 , wherein the clustering algorithm is a non-parametric, unsupervised K-NN algorithm; and wherein the method further comprising:
training the non-parametric, unsupervised K-NN algorithm with the normalized arrays of the plurality of spectral bands.

6. The method of claim 1 , wherein the plurality of spectral bands further includes a quasi-sensitive spectral band having increased sensitivity to the subject chemical species as compared to the plurality of reference spectral bands; and
wherein the background reflectance map is not based on the normalized array of band-specific intensity values for the quasi-sensitive spectral band.

7. The method of claim 1 , wherein the image data is obtained as spatially aliased image data; and
wherein the method further comprises, prior to removing the albedo background, performing spatial anti-aliasing of the spatially aliased image data.

8. The method of claim 7 , wherein performing the spatial anti-aliasing includes applying a Gaussian filter to the aliased image data.

9. The method of claim 1 , wherein removing the albedo background from the image data includes, for each of the plurality of spectral bands:
detecting the albedo background for the spectral band; and
determining a difference between the albedo background for the spectral band and the array of band-specific intensity values for the spectral band;
wherein the normalized array of band-specific intensity values for the spectral band is the difference between the albedo background for the spectral band and the array of band-specific intensity values for the spectral band.

10. The method of claim 1 , further comprising:
comparing each intensity variance value of the index array to a variance threshold to obtain an image mask that distinguishes intensity variance values that are less than the variance threshold from other intensity variance values that are greater than the variance threshold; and
outputting the image mask.

11. A computing system for optically detecting a subject chemical species within an atmospheric environment, the computing system comprising:
a logic machine; and
a storage machine having instructions stored thereon executable by the logic machine to:
obtain image data representing multispectral imagery of a geographic region captured through the atmospheric environment via an aeronautical vehicle,
the image data including an array of band-specific intensity values for each of a plurality of spectral bands, including a plurality of reference spectral bands and a sample spectral band having increased sensitivity to the subject chemical species as compared to the plurality of reference spectral bands;

remove albedo background from the image data for each of the plurality of spectral bands to obtain a normalized array of band-specific intensity values for each of the plurality of spectral bands;
generate a background reflectance map that includes an array of inter-band intensity values in which each inter-band intensity value represents a filtered combination of band-specific intensity values of the normalized arrays for a grouped subset of the plurality of reference spectral bands;
compare the normalized array of band-specific intensity values for the sample spectral band to the background reflectance map to obtain an index array of intensity variance values; and
output the index array for the subject chemical species.

12. The computing system of claim 11 , wherein for each inter-band intensity value, the filtered combination is an average of the band-specific intensity values of the grouped subset of the plurality of reference spectral bands.

13. The computing system of claim 11 , wherein the instructions are further executable by the logic machine to:
for each inter-band intensity value of the background reflectance map, classify the plurality of reference spectral bands into the grouped subset using a clustering algorithm.

14. The computing system of claim 13 , wherein the clustering algorithm is a non-parametric, unsupervised K-NN algorithm; and
wherein the instructions are further executable by the logic machine to train the non-parametric, unsupervised K-NN algorithm with the normalized arrays of the plurality of spectral bands.

15. The computing system of claim 11 , wherein the plurality of spectral bands further includes a quasi-sensitive spectral band having increased sensitivity to the subject chemical species as compared to the plurality of reference spectral bands; and
wherein the background reflectance map is not based on the normalized array of band-specific intensity values for the quasi-sensitive spectral band.

16. The computing system of claim 11 , wherein the image data is obtained as spatially aliased image data; and
wherein the instructions are further executable by the logic machine to:
prior to removing the albedo background, perform spatial anti-aliasing of the spatially aliased image data.

17. The computing system of claim 11 , wherein the instructions are further executable by the logic machine to remove the albedo background from the image data by, for each of the plurality of spectral bands:
detecting the albedo background for the spectral band; and
determining a difference between the albedo background for the spectral band and the array of band-specific intensity values for the spectral band;
wherein the normalized array of band-specific intensity values for the spectral band is the difference between the albedo background for the spectral band and the array of band-specific intensity values for t

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Malvar Maua Sara
• de Oliveira Nunes Leonardo
• De Moura Estevao Filho Roberto
• Oruganti Yagna Deepika
• Badam Anirudh

Classifications

• G06V10/58

  relating to hyperspectral data · CPC title

• G06V10/60

  relating to illumination properties, e.g. using a reflectance or lighting model · CPC title

• G06N3/08

  Learning methods · CPC title

• G06N20/00

  Machine learning · CPC title

• G06V20/194 Primary

  using hyperspectral data, i.e. more or other wavelengths than RGB · CPC title