Systems and methods for identifying roads in images

• US 9,547,805 B1
• Filed: 01/22/2013
• Issued: 01/17/2017
• Est. Priority Date: 01/22/2013
• Status: Active Grant

• Current Assignee: The Boeing Co.
• Original Assignee: The Boeing Co.
• Inventors: Erignac, Charles A.
• Primary Examiner(s): Park, Chan
• Assistant Examiner(s): Thomas, Mia M
• Application Number:
  US13/747,007
• Time in Patent Office: 1,456 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06V 10/426 Graphical representations
• CPC Class Codes: G06V 20/182 Network patterns, e.g. roads or rivers
• CPC Class Codes: G06V 20/188 Vegetation

Abstract

Methods and systems described herein enable self-supervised road detection in images. The method includes receiving an image, segmenting the image into at least one fragment based at least in part on at least one pixel feature, determining, using a processor, a road likeness score for the at least one fragment based at least in part on a medial radius, and identifying roads based at least in part on the road likeness score.

Claims

1. A computer-implemented method comprising:

receiving an image formed from at least one of aerial or satellite imagery data;

segmenting, via a subdivision engine, the image into at least one fragment having a medial axis and medial radiuses, wherein the medial radiuses correspond to about half the width of the at least one fragment, and wherein the segmenting is based at least in part on at least one pixel feature;

for each of the at least one fragments, computing a mean medial radius and a standard deviation for the corresponding fragment based on the corresponding medial radiuses;

determining, using a processor, a road likeness score for each of the at least one fragments based at least in part on the medial radiuses of the corresponding fragment and the computed mean medial radius of the corresponding fragment; and

identifying roads in the image based at least in part on the road likeness score.

2. A computer-implemented method in accordance with claim 1, wherein segmenting the image comprises clustering pixels into segments based at least in part on at least one pixel feature.

3. A computer-implemented method in accordance with claim 2, wherein the at least one pixel feature is one of hue, saturation, and normalized difference vegetation index.

4. A computer-implemented method in accordance with claim 2, wherein segmenting the image further comprises subdividing segments into fragments.

5. A computer-implemented method in accordance with claim 4, wherein subdividing segments into fragments comprises using an oriented bounding box.

6. A computer-implemented method in accordance with claim 1, further comprising labeling each fragment as one of unknown, road, and non-road.

7. A computer-implemented method in accordance with claim 6, further comprising training a classifier using at least one labeled fragment.

8. A computer-implemented method in accordance with claim 1, further comprising outputting a binary road mask that identifies pixels in the image associated with an identified road.

9. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon, wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

receive an image formed from at least one of aerial or satellite imagery data;

segment the image into at least one fragment having a medial axis and medial radiuses, wherein the medial radiuses correspond to about half the width of the at least one fragment, and wherein the image is segmented based at least in part on at least one pixel feature;

for each of the at least one fragments, compute a mean medial radius and a standard deviation for the corresponding fragment based on the corresponding medial radiuses;

determine a road likeness score for each of the at least one fragments based at least in part on the medial radiuses of the corresponding fragment and the computed mean medial radius of the corresponding fragment; and

identify roads in the image based at least in part on the road likeness score.

10. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon according to claim 9 wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

cluster pixels into segments based at least in part on at least one pixel feature.

11. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon according to claim 9 wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

segment the image into at least one fragment based at least in part on at least one pixel feature, wherein the at least one pixel feature is one of hue, saturation, and normalized difference vegetation index.

12. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon according to claim 9 wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

subdivide segments into fragments.

13. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon according to claim 9 wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

subdivide segments into fragments using an oriented bounding box.

14. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon according to claim 9 wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

label each fragment as one of unknown, road, and non-road.

15. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon according to claim 14 wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

train a classifier using at least one labeled fragment.

16. One or more non-transitory computer-readable storage media having computer-executable instructions embodied thereon according to claim 9 wherein when executed by at least one processor, the computer-executable instructions cause the at least one processor to:

output a binary road mask that identifies pixels in the image associated with an identified road.

17. A system for analyzing images, said system comprising:

a computing device comprising;

a segmentation engine configured to;

receive an image formed from at least one of aerial or satellite imagery data; and

segment the image into at least one fragment having a medial axis and medial radiuses, wherein the medial radiuses correspond to about half the width of the at least one fragment, and wherein the image is segmented based at least in part on at least one pixel feature; and

a classifier configured to;

for each of the at least one fragments, computing a mean medial radius and a standard deviation for the corresponding fragment based on the corresponding medial radiuses;

determine a road likeness score for each of the at least one fragments based at least in part on the medial radiuses of the corresponding fragment and the computed mean medial radius of the corresponding fragment; and

identify roads in the image based at least in part on the road likeness score.

18. A system in accordance with claim 17, wherein said segmentation engine is further configured to subdivide segments into fragments.

19. A system in accordance with claim 17, wherein said segmentation engine is further configured to subdivide segments into fragments using an oriented bounding box.

20. A system in accordance with claim 17, wherein said classifier is further configured to label each fragment as one of unknown, road, and non-road.