Thermal wake/vessel detection technique

1. A computer-implemented method for detecting a vessel in water based on an image of a portion of Earth, the method comprising:

generating a thermal anomaly mask flagging each pixel of the image initially deemed to be a wake pixel based on a comparison of a thermal value of each pixel against other thermal values of other pixels localized about each pixel;

grouping contiguous pixels flagged by the thermal anomaly mask into pixel clusters;

analyzing a shape of each of the pixel clusters to determine whether each of the pixel clusters represents a possible vessel detection event; and

visually representing any possible vessel detection event within the image.

2. The computer-implemented method of claim 1 , wherein generating the thermal anomaly mask comprises:

separating the image into groups, wherein the groups comprise a plurality of overlapping groups of pixels;

generating statistics for each of the groups; and

for each of the pixels of the image:

comparing the thermal value of a given pixel to the statistics for each of the groups of which the given pixel is a member; and

flagging the given pixel based on a result of the comparing.

3. The computer-implemented method of claim 2 ,

wherein generating the statistics for each of the groups comprises:

generating a mean thermal value (MTV) for each of the groups; and

generating a thermal standard deviation value (TSDV) for each of the groups,

wherein comparing the thermal value of the given pixel to the statistics for each of the groups of which the given pixel is the member comprises comparing the thermal value of the given pixel to a combination of the MTV and the TSDV for each of the groups of which the given pixel is the member, and

wherein flagging the given pixel based on the result of the comparing comprises flagging the given pixel based on the result of the comparing for a threshold percentage of the groups including the given pixel.

4. The computer-implemented method of claim 3 , wherein the combination of the MTV and the TSDV comprises either [(the MTV)+(n*(the TSDV))] or [(the MTV)−(n*(the TSDV))], where n is a constant multiplier.

5. The computer-implemented method of claim 1 , wherein the thermal anomaly mask comprises a cool thermal anomaly mask flagging potential cool wake pixels, the computer-automated method further comprising:

generating a warm thermal anomaly mask flagging potential warm wake pixels within the image; and

generating a hot thermal anomaly mask flagging potential hot ship pixels within the image.

6. The computer-implemented method of claim 5 , wherein the contiguous pixels comprise first contiguous pixels and the pixel clusters comprise cool wake pixel clusters, the computer-automated method further comprising:

grouping second contiguous pixels flagged by the warm thermal anomaly mask into warm wake pixel clusters; and

grouping third contiguous pixels flagged by the hot thermal anomaly mask into hot ship pixel clusters.

7. The computer-implemented method of claim 1 , further comprising performing a minimum size cluster test on each of the pixel clusters to filter a false vessel detection event.

8. The computer-implemented method of claim 1 , further comprising performing a thermal gradient test on each of the pixel clusters to filter a false vessel detection event, the thermal gradient test comprising:

inspecting at least first and second thermal values within the image on either side of a given pixel cluster; and

if the first and second thermal values deviate greater than a threshold gradient value, then rejecting the given pixel cluster as a false vessel detection event.

9. The computer-implemented method of claim 1 , wherein analyzing the shape of each of the pixel clusters comprises:

performing a linear shape analysis on a given pixel cluster to generate a linear shape correlation value; and

comparing the linear shape correlation value to a threshold linear shape correlation value to determine if the given pixel cluster has a sufficiently elongated shape to represent a wake shape.

10. The computer-implemented method of claim 1 , wherein analyzing the shape of each of the pixel clusters comprises:

measuring a width of a given pixel cluster; and

comparing the width to a threshold wake width to determine if the given pixel cluster has a sufficiently narrow width to represent a wake shape.

11. The computer-implemented method of claim 1 , wherein analyzing the shape of each of the pixel clusters comprises:

determining a best fit line through a given pixel cluster;

determining a length of the best fit line through the given pixel cluster;

determining a width of the given pixel cluster measured normal to the best fit line; and

comparing a ratio of the length to the width to a threshold aspect ratio to determine if the given pixel cluster has a sufficiently slender aspect ratio to represent a wake shape.

12. The computer-implemented method of claim 1 , further comprising:

assigning one of at least two confidence levels to each of the pixel clusters determined to represent the possible vessel detection event; and

visually indicating the confidence levels associated with each of the pixel clusters determined to represent the possible vessel detection event.

13. The computer-implemented method of claim 1 , wherein the image comprises an aerial perspective image and the thermal value of each pixel comprises a brightness temperature calculated according to Plank's Law.

14. A non-transitory computer-readable storage medium that provides instructions that, if executed by a computer, will cause the computer to perform operations comprising:

generating a thermal anomaly mask flagging each pixel of an image of a portion of Earth initially deemed to be a wake pixel based on a comparison of a thermal value of each pixel against other thermal values of other pixels localized about each pixel;

grouping contiguous pixels flagged by the thermal anomaly mask into pixel clusters;

analyzing a shape of each of the pixel clusters to determine whether each of the pixel clusters represents a possible vessel detection event; and

visually representing any possible vessel detection event within the image.

15. The computer-readable storage medium of claim 14 , wherein generating the thermal anomaly mask comprises:

separating the image into groups, wherein the groups comprise a plurality of overlapping groups of pixels;

generating a mean thermal value (MTV) for each of the groups;

generating a thermal standard deviation value (TSDV) for each of the groups; and

for each of the pixels of the image:

comparing the thermal value of a given pixel to a combination of the MTV and the TSDV for each of the groups of which the given pixel is a member; and

flagging the given pixel based on a result of the comparing for a threshold percentage of the groups including the given pixel.

16. The computer-readable storage medium of claim 15 , wherein the combination of the MTV and the TSDV comprises either [(the MTV)+(n*(the TSDV))] or [(the MTV)−(n*(the TSDV))], where n is a constant multiplier.

17. The computer-readable storage medium of claim 14 , wherein the thermal anomaly mask comprises a cool thermal anomaly mask flagging potential cool wake pixels, wherein the computer-readable storage medium further provides instructions that, if executed by the computer, will cause the computer to perform further operations, comprising:

generating a warm thermal anomaly mask flagging potential warm wake pixels within the image; and

generating a hot thermal anomaly mask flagging potential hot ship pixels within the image.

18. The computer-readable storage medium of claim 17 , wherein the contiguous pixels comprise first contiguous pixels and the pixel clusters comprise cool wake pixel clusters, wherein the computer-readable storage medium further provides instructions that, if executed by the computer, will cause the computer to perform further operations, comprising:

grouping second contiguous pixels flagged by the warm thermal anomaly mask into warm wake pixel clusters; and

grouping third contiguous pixels flagged by the hot thermal anomaly mask into hot ship pixel clusters.

19. The computer-readable storage medium of claim 14 , further providing instructions that, if executed by the computer, will cause the computer to perform further operations, comprising:

performing a minimum size cluster test on each of the pixel clusters to filter false vessel detection event.

20. The computer-readable storage medium of claim 14 , further providing instructions that, if executed by the computer, will cause the computer to perform further operations, comprising:

performing a thermal gradient test on each of the pixel clusters to filter a false vessel detection event, the thermal gradient test comprising:

inspecting at least first and second thermal values within the image on either side of a given pixel cluster; and

if the first and second thermal values deviate greater than a threshold gradient value, then rejecting the given pixel cluster as a false vessel detection event.

21. The computer-readable storage medium of claim 14 , wherein analyzing the shape of each of the pixel clusters comprises:

performing a linear shape analysis on a given pixel cluster to generate a linear shape correlation value; and

comparing the linear shape correlation value to a threshold linear shape correlation value to determine if the given pixel cluster has a sufficiently elongated shape to represent a wake shape.

22. The computer-readable storage medium of claim 14 , wherein analyzing the shape of each of the pixel clusters comprises:

measuring a width of a given pixel cluster; and

comparing the width to a threshold wake width to determine if the given pixel cluster has a sufficiently narrow width to represent a wake shape.

23. The computer-readable storage medium of claim 14 , wherein analyzing the shape of each of the pixel clusters comprises:

determining a best fit line through a given pixel cluster;

determining a length of the best fit line through the given pixel cluster;

determining a width of the given pixel cluster measured normal to the best fit line; and

comparing a ratio of the length to the width to a threshold aspect ratio to determine if the given pixel cluster has a sufficiently slender aspect ratio to represent a wake shape.

24. The computer-readable storage medium of claim 14 , further providing instructions that, if executed by the computer, will cause the computer to perform further operations, comprising:

assigning one of at least two confidence levels to each of the pixel clusters determined to represent the possible vessel detection event; and

visually indicating the confidence levels associated with each of the pixel clusters determined to represent the possible vessel detection event.

25. The computer-readable storage medium of claim 14 , wherein the image comprises an aerial perspective image and the thermal value of each pixel comprises a brightness temperature calculated according to Plank's Law.