Detecting debris in a vehicle path

1. A system comprising:

one or more processors; and

one or more non-transitory computer-readable media including executable instructions, which, when executed by the one or more processors, configure the one or more processors to perform operations comprising:

receiving at least one image of a road;

determining at least one candidate group of pixels in the image as potentially corresponding to debris on the road;

determining at least two height-based features for the candidate group of pixels, the at least two height-based features including a maximum height associated with the candidate group of pixels relative to a surface of the road and an average height associated with the candidate group of pixels relative to the surface of the road;

determining at least one weighting factor based on comparing the at least two height-based features to respective thresholds;

determining one or more slope-based features associated with the candidate group of pixels;

determining at least one weight value for the one or more slope-based features;

determining an aggregated weight based on the at least one weight value and the weighting factor; and

determining whether the group of pixels corresponds to debris based at least on the aggregated weight.

2. The system as recited in claim 1 , the operations further comprising determining a validity of the pixels in the group of pixels based at least on determining whether individual pixels in the group of pixels represent the surface of the road.

3. The system as recited in claim 2 , wherein the determining the validity of the pixels is further based on determining, for the individual pixels in the group of pixels at least one of:

an upward slope associated with an individual pixel relative to neighboring pixels;

a downward slope associated with the individual pixel relative to the neighboring pixels; or

a height associated with the individual pixel relative to neighboring pixels.

4. The system as recited in claim 2 , the operations further comprising determining a road plane estimation of the surface of the road prior to determining the validity of the pixels in the group of pixels.

5. The system as recited in claim 1 , the operations further comprising:

determining a region of interest in the image based at least in part on lane information determined from the image; and

determining the at least one candidate group of pixels in the image based at least on determining that the at least one candidate group of pixels is located within the region of interest.

6. The system as recited in claim 1 , wherein two images are received from a stereo camera, the operations further comprising:

determining a disparity image from the received images;

determining edge information for at least one of the received images;

determining a modified disparity image based on the disparity image and the edge information; and

determining the at least one candidate group of pixels from the modified disparity image.

7. The system as recited in claim 1 , wherein the one or more slope-based features include at least one of:

an average slope associated with individual pixels relative to neighboring pixels;

a maximum slope associated with individual pixels; or

a percentage of pixels in the group of pixels associated with a slope that exceeds a specified value.

8. The system as recited in claim 1 , the operations further comprising sending at least one control signal based on determining that the group of pixels corresponds to debris, the at least one control signal including at least one of:

an instruction for controlling a vehicle to cause the vehicle to avoid the debris;

an instruction to cause an alert;

an instruction to cause presentation of information related to the debris on a display.

9. The system as recited in claim 1 , the operations further comprising:

based on determining that the group of pixels corresponds to debris, sending debris information related to the debris to a computing device over a network, to cause, at least in part, the computing device to store the debris information in a data structure; and

receiving additional debris information from the computing device, the debris information indicating a location of additional debris in another roadway; and

determining a route to a destination based at least on the additional debris information.

10. A method comprising:

receiving, by one or more processors, at least one image of a road;

determining at least one candidate group of pixels in the image as potentially corresponding to debris on the road;

determining at least two height-based features for the candidate group of pixels, the at least two height based features including a maximum height associated with the candidate group of pixels relative to a surface of the road and an average height associated the candidate group of pixels relative to the surface of the road;

determining at least one weighting factor based on comparing the at least two height-based features to respective thresholds; and

determining whether the group of pixels corresponds to debris based at least on the comparing.

11. The method as recited in claim 10 , further comprising determining one or more slope-based features associated with the candidate group of pixels, the one or more slope-based features comprising at least one of:

an average slope associated with individual pixels relative to neighboring pixels;

a maximum slope associated with individual pixels; or

a percentage of pixels in the group of pixels associated with a slope that exceeds a specified value.

12. The method as recited in claim 11 , further comprising

applying the at least one weighting factor to at least one weight value for at least one of the one or more slope-based features to determine an aggregated weight; and

determining whether the group of pixels corresponds to debris based at least on the aggregated weight.

13. The method as recited in claim 10 , further comprising determining a validity of the pixels in the group of pixels based at least on determining whether individual pixels in the group of pixels represent the surface of the road.

14. The method as recited in claim 10 , further comprising

determining a region of interest in the image based at least in part on lane information determined from the image; and

determining the at least one candidate group of pixels in the image based at least on determining that the at least one candidate group of pixels is located within the region of interest.

15. The method as recited in claim 10 , wherein two images are received from a stereo camera, the method further comprising

determining a disparity image from the received images;

determining edge information for at least one of the received images;

determining a modified disparity image based on the disparity image and the edge information; and

determining the at least one candidate group of pixels from the modified disparity image.

16. A system comprising:

a one or more processors; and

one or more non-transitory computer-readable media including executable instructions, which, when executed by the one or more processors, configure the one or more processors to perform operations comprising:

receiving at least one image of a road traversed by a vehicle;

determining at least one candidate group of pixels in the image as potentially corresponding to debris on the road;

determining one or more height-based features for the candidate group of pixels;

determining a weighting factor based on comparing the one or more height-based features to a respective threshold;

determining a plurality of slope-based features associated with the candidate group of pixels, the slope-based features including at least one of:

an average slope associated with individual pixels relative to neighboring pixels;

a maximum slope associated with the individual pixels; or

a percentage of pixels in the group of pixels associated with a slope that exceeds a specified value;

applying the weighting factor to at least one weight value for at least one of the plurality of slope-based features to determine an aggregated weight; and

determining whether the group of pixels corresponds to debris based at least on the aggregated weight.

17. The system as recited in claim 16 , the operations further comprising determining a validity of the pixels in the group of pixels based at least on determining whether individual pixels in the group of pixels represent the surface of the road.

18. The system as recited in claim 16 , the operations further comprising:

determining one or more height-based features for the candidate group of pixels, the one or more height based features including at least one of:

a maximum height associated with the candidate group of pixels relative to a surface of the road, or

an average height associated with the candidate group of pixels relative to the surface of the road;

determining the at least one weighting factor based on comparing at least one of the height-based features to a threshold.

19. The system as recited in claim 16 , the operations further comprising:

determining a region of interest in the image based at least in part on lane information determined from the image; and

determining the at least one candidate group of pixels in the image based at least on determining that the at least one candidate group of pixels is located within the region of interest.

20. The system as recited in claim 16 , wherein two images are received from a stereo camera, the operations further comprising:

determining a disparity image from the received images;

determining edge information for at least one of the received images;

determining a modified disparity image based on the disparity image and the edge information; and

determining the at least one candidate group of pixels from the modified disparity image.