System and method for detecting obstacles using a single camera

• US 9,798,936 B2
• Filed: 07/31/2012
• Issued: 10/24/2017
• Est. Priority Date: 07/31/2012
• Status: Active Grant

• Current Assignee: Harman International Industries Incorporated (Samsung Electronics Co. Ltd.)
• Original Assignee: Harman International Industries Incorporated (Samsung Electronics Co. Ltd.)
• Inventors: Jiang, Ruyi, Zhang, Yankun, Hong, Chuyang, Weyrich, Norman
• Primary Examiner(s): Moyer, Andrew
• Assistant Examiner(s): Rosario, Dennis
• Application Number:
  US14/391,937
• Publication Number:
  US 20150078619A1
• Time in Patent Office: 1,911 Days
• CPC Class Codes: B60Q 9/005 using a video camera
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/30261 Obstacle
• CPC Class Codes: G06T 5/80 Geometric correction
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
• CPC Class Codes: G06V 20/586 of parking space
• CPC Class Codes: G08G 1/168 Driving aids for parking, e.g. acoustic or visual feedback on parking space

Abstract

The present application provides an obstacle detection system and method thereof. The obstacle detection method comprises: obtaining a first image captured by a camera at a first time point; identifying a vertical edge candidate in the first image, and measuring a first length of the vertical edge candidate based on the first image; obtaining a second image captured by the camera at a second time point; measuring a second length of the vertical edge candidate based on the second image; calculating a difference between the first length and the second length; and comparing the difference with a predetermined length difference threshold, if the difference is greater than the length difference threshold, outputting a message that an obstacle is found.

Claims

1. An obstacle detection method comprising:

obtaining a first image captured by a camera at a first time point;

generating a first transformed image based on the first image;

identifying a vertical edge candidate in the first transformed image, and measuring a first length of the vertical edge candidate based on the first transformed image;

obtaining a second image captured by the camera at a second time point;

generating a second transformed image based on the second image;

measuring a second length of the vertical edge candidate based on the second transformed image;

measuring an angular deviation between the vertical edge candidate and a focal line in the first transformed image or the second transformed image;

comparing the measured angular deviation with a predetermined angular deviation threshold to filter out a false vertical edge candidate;

calculating a difference between the first length and the second length; and

comparing the difference with a predetermined length difference threshold, if the difference is greater than the predetermined length difference threshold, outputting a message that an obstacle is found;

wherein the vertical edge candidate comprises a straight line segment in the first and second transformed images.

2. The method of claim 1, wherein the camera is a visual spectrum camera able to capture color images and/or greyscale images.

3. The method of claim 1,

wherein the first transformed image comprises a first top view image;

wherein the second transformed image comprises a second top view image;

wherein generating the first top view image is based on the first image using an inverse perspective mapping method; and

wherein generating the second top view image is based on the second image using the inverse perspective mapping method; and

further comprising

measuring a length of the vertical edge candidate in the first top view image as the first length; and

measuring a length of the vertical edge candidate in the second top view image as the second length.

4. The method of claim 3, further comprising:

performing distortion correction on the first image to obtain a first corrected image;

generating the first top view image using the first corrected image;

performing distortion correction on the second image to obtain a second corrected image; and

generating the second top view image using the second corrected image.

5. The method of claim 3,

wherein comparing the measured angular deviation with the predetermined angular deviation threshold to filter out the false vertical edge candidate includes filtering out the vertical edge candidate if the measured angular deviation is greater than the predetermined angular deviation threshold, and

wherein the focal line comprises a line passing through a projected focal point of the camera in the first top view image.

6. The method of claim 3,

wherein measuring the angular deviation includes measuring the angular deviation between the vertical edge candidate and the focal line in the first top view image or the second top view image;

wherein comparing the measured angular deviation with the predetermined angular deviation threshold to filter out the false vertical edge candidate includes filtering out the vertical edge candidate if the measured angular deviation is greater than the predetermined angular deviation threshold; and

wherein the focal line comprises a line passing through a projected focal point of the camera in the second top view image.

7. The method of claim 5, wherein the focal line further passes through a point on the vertical edge candidate.

8. The method of claim 1, where the method is a parking assist method, to identify and remind a driver whether there is any obstacle behind a vehicle when the driver is parking the vehicle, and wherein the vertical edge candidate is identified using a fast line segment method.

9. The method of claim 1, wherein the first image is captured in response to the vertical edge candidate entering into a tracking region, and the second image is captured in response to the vertical edge candidate entering into an alert region which is smaller than the tracking region and falls in the tracking region, the method further comprising determining if the vertical edge candidate within the alert region is an obstacle, and outputting an alert via a speaker responsive to determining that the vertical edge candidate within the alert region is an obstacle.

10. The method of claim 9, wherein the tracking region is configured such that a difference between the first length and the second length of the vertical edge candidate when the vertical edge candidate is identified as a true vertical edge is large enough to differentiate the true vertical edge from false candidates.

11. An obstacle detection system comprising:

a processing device configured to;

obtain a first image captured by a camera at a first time point;

generate a first top view image based on the first image;

identify a vertical edge candidate based on the first top view image, and measure a first length of the vertical edge candidate based on the first top view image;

obtain a second image captured by the camera at a second time point;

generate a second top view image based on the second image;

measure a second length of the vertical edge candidate based on the second top view image;

measure an angular deviation between the vertical edge candidate and a focal line in the first top view image or the second top view image;

compare the measured angular deviation with a predetermined angular deviation threshold to filter out a false vertical edge candidate;

calculate a difference between the first length and the second length; and

compare the difference with a predetermined length difference threshold, if the difference is greater than the predetermined length difference threshold, instruct an output device to output a message that an obstacle is found;

wherein the vertical edge candidate comprises a straight line segment in the first and second top view images.

12. The obstacle detection system of claim 11, wherein the camera is a visual spectrum camera configured to capture color images or greyscale images.

13. The obstacle detection system of claim 11, further comprising a memory device for storing a status of vertical edge candidates, and wherein the vertical edge candidate is identified using a fast line segment method.

14. The obstacle detection system of claim 11, wherein the processing device is further configured to:

measure a length of the vertical edge candidate in the first top view image as the first length; and

measure a length of the vertical edge candidate in the second top view image as the second length;

wherein generating the first top view image based on the first image is performed using an inverse perspective mapping method; and

wherein generating the second top view image based on the second image is performed using the inverse perspective mapping method.

15. The obstacle detection system of claim 14, wherein the processing device is further configured to:

perform distortion correction on the first image to obtain a first corrected image;

generate the first top view image using the first corrected image;

perform distortion correction on the second image to obtain a second corrected image; and

generate the second top view image using the second corrected image.

16. The obstacle detection system of claim 14, wherein

comparing the measured angular deviation with the predetermined angular deviation threshold includes filtering out the vertical edge candidate if the measured angular deviation is greater than the predetermined angular deviation threshold, and

wherein the focal line comprises a line passing through a projected focal point of the camera in the first top view image.

17. The obstacle detection system of claim 14, wherein

comparing the measured angular deviation with the predetermined angular deviation threshold includes filtering out the vertical edge candidate if the measured angular deviation is greater than the predetermined angular deviation threshold, and

wherein the focal line comprises a line passing through a projected focal point of the camera in the second top view image.

18. The obstacle detection system of claim 16, wherein the focal line further passes through a point on the vertical edge candidate.

19. The obstacle detection system of claim 11, where the system is a vehicle parking assist system to identify and remind a driver whether there is any obstacle behind a vehicle when the driver is parking the vehicle.

20. The obstacle detection system of claim 11, wherein the first image is captured when the vertical edge candidate enters into a tracking region, and the second image is captured when the vertical edge candidate enters into an alert region which is smaller than the tracking region and falls in the tracking region.