Parked vehicle detection based on edge detection

• US 8,923,565 B1
• Filed: 01/05/2014
• Issued: 12/30/2014
• Est. Priority Date: 09/26/2013
• Status: Expired due to Fees

• Current Assignee: ChengDu HaiCun IP Technology LLC, Guobiao Zhang
• Original Assignee: ChengDu HaiCun IP Technology LLC, Guobiao Zhang
• Inventors: Zhang, Guobiao, Feng, Yaofei, Wang, Bruce Bing
• Primary Examiner(s): VILLECCO, JOHN M
• Application Number:
  US14/147,578
• Time in Patent Office: 359 Days
• Field of Search: 382/104, 340/932.2, 340/937, 340/938, 348/113, 348/118, 348/116, 701/117, 701/118
• US Class Current: 382/104
• CPC Class Codes: G06V 20/586 of parking space
• CPC Class Codes: G08G 1/04 using optical or ultrasonic detectors
• CPC Class Codes: G08G 1/14 indicating individual free spaces in parking areas
• CPC Class Codes: G08G 1/146 where the parking area is a limited parking space, e.g. parking garage, restricted space
• CPC Class Codes: G08G 1/147 where the parking area is within an open public zone, e.g. city centre

Abstract

The present invention discloses a method for detecting parked vehicles based on edge detection. For each parking space, its boundary comprises an exposed edge, which is not occluded by any parked vehicle. Its primary detected edges are the detected edges that are substantially parallel to the exposed edge. A parking space is detected as occupied if its primary detected edges satisfy at least one of these conditions: (1) their total number is more than a pre-determined minimum number; and/or, (2) their total length is more than a pre-determined minimum length.

Claims

1. A method for detecting parked vehicles, comprising the steps of:

capturing an image of a plurality of parking spaces, the boundary of each parking space comprising an exposed edge, wherein said exposed edge is not occluded by any parked vehicles;

detecting edges within a region of interest (ROI) associated with each parking space, wherein a substantial portion of said ROI is formed by scanning at least a portion of said exposed edge upward;

selecting primary detected edges from detected edges for each parking space, wherein said primary detected edges are substantially parallel to the exposed edge of each parking space;

determining the parking state of each parking space, wherein a parking space is detected as occupied if the primary detected edges for said parking space satisfy a first pre-determined condition.

2. The method according to claim 1, wherein said first pre-determined condition is that the total number of said primary detected edges is more than a first pre-determined minimum number.

3. The method according to claim 1, wherein said first pre-determined condition is that the total length of said primary detected edges is more than a first pre-determined minimum length.

4. The method according to claim 1, wherein the tangential angles along each of said primary detected edges are within a first pre-determined range from the directional angle of said exposed edge.

5. The method according to claim 1, wherein the directional angle of the best-fit line for each of said primary detected edges is within a second pre-determined range from the directional angle of said exposed edge.

6. The method according to claim 1, wherein each pixel on each of said primary detected edges is within a first pre-determined distance from a shifted line parallel to said exposed edge.

7. The method according to claim 1, wherein:

the boundary of each parking space comprises at least a secondary edge, wherein said secondary edge is not parallel to said exposed edge;

a parking space is detected as vacant if the secondary detected edges for said parking space satisfy a second pre-determined condition, wherein said secondary detected edges are substantially parallel to the secondary edge of said parking space.

8. The method according to claim 1, wherein vehicles in said plurality of parking spaces are parked in parallel.

9. The method according to claim 8, wherein each of said primary detected edges is at least a portion of an edge of a rear part of a vehicle.

10. The method according to claim 8, wherein each of said primary detected edges is at least a portion of an edge of a front part of a vehicle.

11. The method according to claim 1, wherein vehicles in said plurality of parking spaces are parked in series.

12. The method according to claim 11, wherein each of said primary detected edges is at least a portion of an edge of a side part of a vehicle.

13. A method for detecting parked vehicles, comprising the steps of:

capturing an image of a plurality of parking spaces, the boundary of each parking space comprising an exposed edge and a secondary edge, wherein said exposed edge is not occluded by any parked vehicles, and said secondary edge is not parallel to said exposed edge;

detecting edges within a region of interest (ROI) associated with each parking space, wherein a substantial portion of said ROI is formed by scanning at least a portion of said exposed edge upward;

selecting secondary detected edges from detected edges for each parking space, wherein said secondary detected edges are substantially parallel to the secondary edge of each parking space;

determining the parking state of each parking space, wherein a parking spaces is detected as vacant if the secondary detected edges for said parking space satisfy a second pre-determined condition.

14. The method according to claim 13, wherein said second pre-determined condition is that the total number of said secondary detected edges is more than a second pre-determined minimum number.

15. The method according to claim 13, wherein said second pre-determined condition is that the total length of said secondary detected edges is more than a second pre-determined minimum length.

16. The method according to claim 13, wherein the tangential angles along each of said secondary detected edges are within a third pre-determined range from the directional angle of said secondary edge.

17. The method according to claim 13, wherein the directional angle of the best-fit line for each of said secondary detected edges is within a fourth pre-determined range from the directional angle of said secondary edge.

18. The method according to claim 13, wherein each pixel on each of said secondary detected edges is within a second pre-determined distance from a shifted line parallel to said secondary edge.

19. The method according to claim 13, wherein a parking space is detected as occupied if the primary detected edges for said parking space satisfy a first pre-determined condition, said primary detected edges being substantially parallel to the exposed edge of said parking space.

20. The method according to claim 13, wherein vehicles in said plurality of parking spaces are parked in parallel or in series.