SYSTEMS AND METHODS FOR NON-OBSTACLE AREA DETECTION

• US 20190333229A1
• Filed: 07/11/2019
• Published: 10/31/2019
• Est. Priority Date: 09/18/2015
• Status: Abandoned Application

• Current Assignee: Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: DANE, Gokce, BHASKARAN, Vasudev
• Application Number:
  US16/509,043
• Publication Number:
  US 20190333229A1
• Time in Patent Office: Days
• CPC Class Codes: G06F 18/254 of classification results, e.g. of results related to same input data
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 2207/30256 Lane; Road marking
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/20 Analysis of motion motion estimation for coding, decoding, compressing or decompressing digital video signals H04N19/43, H04N19/51
• CPC Class Codes: G06T 7/50 Depth or shape recovery
• CPC Class Codes: G06T 7/507 from shading G06T7/586 takes precedence
• CPC Class Codes: G06V 20/56 exterior to a vehicle by using sensors mounted on the vehicle
• CPC Class Codes: G06V 20/588 Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road

Abstract

A method performed by an electronic device is described. The method includes generating a depth map of a scene external to a vehicle. The method also includes performing first processing in a first direction of a depth map to determine a first non-obstacle estimation of the scene. The method also includes performing second processing in a second direction of the depth map to determine a second non-obstacle estimation of the scene. The method further includes combining the first non-obstacle estimation and the second non-obstacle estimation to determine a non-obstacle map of the scene. The combining includes combining comprises selectively using a first reliability map of the first processing and/or a second reliability map of the second processing The method additionally includes navigating the vehicle using the non-obstacle map.

Claims

1. An electronic device for processing images, comprising:

a memory; and

a processor coupled to the memory configured to;

obtain an image;

generate a depth map based on the image;

map the depth map to linear model parameters based on a depth of a portion of the image;

determine a fitting is achieved based on comparing the linear model parameters to pre-determined model parameters; and

classify a road in the image based on determining that the fitting is achieved.

2. The electronic device of claim 1, wherein the processor is further configured to determine a type of non-obstacle area in the image based on classifying the road in the image.

3. The electronic device of claim 2, wherein the processor is further configured to generate a non-obstacle map comprising an indication of the non-obstacle area.

4. The electronic device of claim 3, wherein, to map the depth map to the linear model parameters based on the depth of the portion of the image, the processor is further configured to:

process the depth map in a first direction to determine linear model parameters based on the first direction; and

process the depth map in a second direction that is different from the first direction to determine linear model parameters based on the second direction, wherein the processor is further configured to combine the results of the processing in the first and second directions to obtain the non-obstacle map.

5. The electronic device of claim 4, wherein the combining is based on a first reliability value associated with the first direction and a second reliability value associated with the second direction.

6. The electronic device of claim 4, wherein the first direction is horizontal and the second direction is vertical.

7. The electronic device of claim 4, wherein processing the depth map in a first direction comprises dividing the depth map into linear segments in the first direction.

8. The electronic device of claim 7, wherein the linear segments in the first direction are rows of pixels of the depth map.

9. The electronic device of claim 7, wherein the linear segments in the first direction are columns of pixels of the depth map.

10. The electronic device of claim 2, wherein the type of non-obstacle area in the image comprises one or more of a flat plane, a slope, a valley, or an irregular road.

11. The electronic device of claim 1, wherein the image comprises a stereo image pair.

12. The electronic device of claim 1, wherein the image was captured from a mono camera and the depth map is generated using a structure from motion process.

13. The electronic device of claim 1, wherein the depth map comprises a depth map of a scene external to a vehicle.

14. The electronic device of claim 13, wherein the processor is further configured to navigate the vehicle based on classifying the road in the image.

15. The electronic device of claim 14, wherein navigating the vehicle comprises using an advanced driver assistance system (ADAS) to regulate at least one of speed or steering of the vehicle based on classifying the road in the image.

16. The electronic device of claim 15, wherein navigating the vehicle comprises identifying a region of interest that is used by at least one of an object detection algorithm or a lane detection algorithm.

17. A method for processing images, comprising:

obtaining an image;

generating a depth map based on the image;

mapping the depth map to linear model parameters based on a depth of a portion of the image;

determining a fitting is achieved based on comparing the linear model parameters to pre-determined model parameters; and

classifying a road in the image based on determining that the fitting is achieved.

18. The method of claim 17, further comprising determining a type of non-obstacle area in the image based on classifying the road in the image.

19. The method of claim 18, further comprising generating a non-obstacle map comprising an indication of the non-obstacle area.

20. The method of claim 19, wherein mapping the depth map to the linear model parameters based on the depth of the portion of the image comprises:

processing the depth map in a first direction to determine linear model parameters based on the first direction; and

processing the depth map in a second direction that is different from the first direction to determine linear model parameters based on the second direction, the method further comprising;

combining the results of the processing in the first and second directions to obtain the non-obstacle map.

21. The method of claim 20, wherein the first direction is horizontal and the second direction is vertical.

22. The method of claim 18, wherein the type of non-obstacle area in the image comprises one or more of a flat plane, a slope, a valley, or an irregular road.

23. The method of claim 17, wherein the depth map comprises a depth map of a scene external to a vehicle.

24. The method of claim 23, further comprising navigating the vehicle based on classifying the road in the image.

25. The method of claim 24, wherein navigating the vehicle comprises using an advanced driver assistance system (ADAS) to regulate at least one of speed or steering of the vehicle based on classifying the road in the image.

26. The method of claim 25, wherein navigating the vehicle comprises identifying a region of interest that is used by at least one of an object detection algorithm or a lane detection algorithm.

27. A non-transitory computer-readable medium for processing images, the non-transitory computer-readable medium storing a program containing instructions that, when executed by a processor of a device, cause the device to perform a method comprising:

obtaining an image;

generating a depth map based on the image;

mapping the depth map to linear model parameters based on a depth of a portion of the image;

determining a fitting is achieved based on comparing the linear model parameters to pre-determined model parameters; and

classifying a road in the image based on determining that the fitting is achieved.