Image-based visibility measurement

• US 7,899,207 B2
• Filed: 10/12/2009
• Issued: 03/01/2011
• Est. Priority Date: 06/08/2004
• Status: Active Grant

• Current Assignee: International Electronic Machines Corp.
• Original Assignee: International Electronic Machines Corp.
• Inventors: vonAchen, William G., Mian, Zahid F.
• Primary Examiner(s): Johns; Andrew W
• Application Number:
  US12/577,701
• Publication Number:
  US 20100027838A1
• Time in Patent Office: 505 Days
• Field of Search: 382/100, 382/286, 356/432, 356/433, 356/434, 356/436, 356/437, 340/601, 702/3, 731/701.6, 731/702.7
• US Class Current: 382/100
• CPC Class Codes: G01N 21/538 for determining atmospheric attenuation and visibility
• CPC Class Codes: G06T 7/00 Image analysis

Abstract

The invention provides an image-based visibility measurement solution in which an image is used to calculate a visibility (visual range). In one embodiment, a lighting condition for the image is determined and the visibility calculation is adjusted based on the lighting condition. Further, the invention can obtain image data for a set of portions of the image and estimate a visual range based on each portion. The estimated visual ranges can be combined to calculate the visibility for the image. Still further, multiple metrics can be calculated, each of which is used to estimate a visual range. Subsequently, the visual ranges can be used to calculate the visibility for the image. Even further, configuration data that is based on a set of training images can be used to calculate the visibility for a new image. To this extent, the invention can incorporate the lighting condition, portions of the image having differing features, multiple metrics, and/or feedback through training images to accurately measure visibility based on an image.

Claims

1. A non-transitory computer-readable medium comprising computer program code configured to enable a computer system to measure visibility, the non-transitory computer-readable medium comprising computer program code for performing a method comprising:

calculating a visibility based on an image and a lighting condition for the image, wherein the calculating includes;

selecting any one of;

a set of visibility calculations for a daytime lighting condition, a set of visibility calculations for a nighttime lighting condition, or a set of visibility calculations comprising at least one calculation for a daytime lighting condition and at least one calculation for a nighttime lighting condition according to the lighting condition; and

calculating the visibility using the selected set of visibility calculations.

2. The non-transitory computer-readable medium of claim 1, the method further comprising determining a lighting condition for the image.

3. The non-transitory computer-readable medium of claim 1, the method further comprising:

selecting a field of view for the image; and

obtaining a captured image having the field of view.

4. The non-transitory computer-readable medium of claim 3, the method further comprising filtering undesired wavelengths of light from the image prior to the calculating.

5. The non-transitory computer-readable medium of claim 1, the method further comprising analyzing an image attribute of at least one portion of the image.

6. The non-transitory computer-readable medium of claim 1, the calculating further including:

estimating a plurality of visual ranges, wherein each estimated visual range is based on image data for one of a plurality of portions of the image; and

calculating the visibility based on the plurality of visual ranges.

7. The non-transitory computer-readable medium of claim 6, the estimating including:

calculating a plurality of metrics based on the image data for the one of the plurality of portions;

estimating a plurality of metric-based visual ranges, wherein each of the plurality of metric-based visual ranges is based on one of the plurality of metrics; and

estimating the visual range for the one of the plurality of portions based on the plurality of metric-based visual ranges.

8. A method comprising:

obtaining image data for a plurality of portions of an image; and

estimating a unique visual range for each of the plurality of portions of the image, wherein each visual range is estimated based on a set of features detected in the image data for at least one of;

a set of target objects or a set of light sources in a unique one of the plurality of portions.

9. The method of claim 8, further comprising calculating a visibility for the image based on a combination of the plurality of visual ranges.

10. The method of claim 8, further comprising:

identifying a plurality of unique portions in the image; and

selecting a subset of the plurality of unique portions as the plurality of portions based on a lighting condition for the image.

11. The method of claim 8, wherein the estimating includes:

calculating a plurality of metrics based on the image data for the one of the plurality of portions;

estimating a plurality of metric-based visual ranges, wherein each of the plurality of metric-based visual ranges is based on one of the plurality of metrics; and

estimating the visual range for the one of the plurality of portions based on the plurality of metric-based visual ranges.

12. A non-transitory computer-readable medium comprising computer program code configured to enable a computer system to perform a method comprising:

obtaining image data for a plurality of portions of an image; and

estimating a unique visual range for each of the plurality of portions of the image, wherein each visual range is estimated based on a set of features detected in the image data for at least one of;

a set of target objects or a set of light sources in a unique one of the plurality of portions.

13. The non-transitory computer-readable medium of claim 12, the method further comprising calculating a visibility for the image based on a combination of the plurality of visual ranges.

14. The non-transitory computer-readable medium of claim 12, the method further comprising:

identifying a plurality of unique portions in the image; and

selecting a subset of the plurality of unique portions as the plurality of portions based on a lighting condition for the image.

15. The non-transitory computer-readable medium of claim 12, wherein the estimating includes:

calculating a plurality of metrics based on the image data for the one of the plurality of portions;

estimating a plurality of metric-based visual ranges, wherein each of the plurality of metric-based visual ranges is based on one of the plurality of metrics; and

estimating the visual range for the one of the plurality of portions based on the plurality of metric-based visual ranges.

16. The non-transitory computer-readable medium of claim 15, wherein the plurality of metrics include at least one metric based on a pixel intensity for at least a portion of the image.

17. The non-transitory computer-readable medium of claim 15, wherein the plurality of metrics include at least one metric based on a contrast between a target object in the image and a background.

18. The non-transitory computer-readable medium of claim 15, wherein the plurality of metrics include at least one metric based on edges detected in at least a portion of the image.

19. The non-transitory computer-readable medium of claim 15, wherein the plurality of metrics include at least one metric based on a wavelet transform of image data for at least a portion of the image.

20. The non-transitory computer-readable medium of claim 15, wherein the plurality of metrics include at least one metric based on a number of light sources in at least a portion of the image.

21. The non-transitory computer-readable medium of claim 15, wherein the plurality of metrics include at least one metric based on a color intensity for at least a portion of the image.