Vehicle dimension estimation from vehicle images

• US 9,396,553 B2
• Filed: 04/16/2014
• Issued: 07/19/2016
• Est. Priority Date: 04/16/2014
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Fan, Zhigang, Wu, Hao, Wang, Yao Rong
• Primary Examiner(s): YANG, WEI WEN
• Application Number:
  US14/254,360
• Publication Number:
  US 20150302611A1
• Time in Patent Office: 825 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20021 Dividing image into blocks, subimages or windows
• CPC Class Codes: G06T 2207/30236 Traffic on road, railway or crossing
• CPC Class Codes: G06T 2207/30248 Vehicle exterior or interior
• CPC Class Codes: G06T 2210/12 Bounding box
• CPC Class Codes: G06T 7/60 Analysis of geometric attributes
• CPC Class Codes: G06T 7/73 using feature-based methods
• CPC Class Codes: G06V 10/42 Global feature extraction by analysis of the whole pattern, e.g. using frequency domain transformations or autocorrelation
• CPC Class Codes: G06V 20/56 exterior to a vehicle by using sensors mounted on the vehicle

Abstract

A method and system for determining one or more dimension estimations for a vehicle. A sequence of image frames of a vehicle is received, and a digital representation of the vehicle is extracted from each image in the sequence of image frames. A bounding box is determined for the vehicle based upon the extracted digital representation for each digital representation of the vehicle and at least one dimension of the vehicle is estimated based upon the determined bounding box. An indication of the at least one estimated dimension is output. The indication of at least one estimated dimension is transferred as an input to a comparison technique for further processing the indication of the at least one estimated dimension. The comparison technique can include a parking determination process configured to determine a recommended spot for the vehicle based upon the indication of the at least one estimated dimension.

Claims

1. A method of determining one or more dimension estimations for a vehicle, the method comprising:

receiving, by a processor, a sequence of image frames of a vehicle;

extracting, by the processor, a digital representation of the vehicle from each image in the sequence of image frames;

identifying a moving axis of the vehicle based on at least two image frames in the sequence;

extracting two or more points where two or more tires of a side of the vehicle contact a supporting surface, wherein the extracting the two or more points comprises detecting an edge or subtracting a background of each image, and wherein the two or more points form a line that runs along the identified moving axis of the vehicle;

determining, by the processor, a bounding box for the vehicle based at least partially upon the two or more extracted points for the digital representation of the vehicle;

estimating, by the processor, at least one dimension for the vehicle based upon the determined bounding box; and

outputting, by the processor, an indication of the at least one estimated dimension.

2. The method of claim 1, wherein outputting the indication comprises transferring the indication of the at least one estimated dimension as an input to a comparison technique for further processing the indication of the at least one estimated dimension.

3. The method of claim 2, wherein the comparison technique comprises a parking determination process configured to determine a recommended spot for the vehicle based upon the indication of the at least one estimated dimension.

4. The method of claim 1, wherein determining the bounding box further comprises identifying a line associated with the point where the vehicle contacts the supporting surface.

5. The method of claim 4, wherein identifying the line associated with the point where the vehicle contacts the supporting surface comprises:

determining two or more points where each tire of a near side of the vehicle contacts the supporting surface; and

drawing a line that includes a segment that extends between each of the points.

6. The method of claim 4, wherein determining the bounding box further comprises identifying a far right point of the digital representation of the vehicle and a far left point of the digital representation of the vehicle.

7. The method of claim 1, wherein estimating at least one dimension of the vehicle comprises estimating a length and width of the vehicle based upon a bottom edge of the boundary box.

8. The method of claim 7, wherein estimating at least one dimension of the vehicle comprises estimating a height of the vehicle based upon a distance measured between a rear vertex of the bounding box and an upper vertex of the bounding box.

9. The method of claim 1, wherein extracting the digital representation of the vehicle comprises at least one of example background subtraction, particle filtering, and a shadow removal algorithm.

21. The method of claim 1, wherein extracting the two or more points comprises analyzing contrast ratios at areas where the two or more tires are most likely to be placed.

24. The method of claim 3, wherein determining the recommended spot comprises comparing the indication of the at least one estimated dimension against a listing of dimensions of available parking spaces.

10. A system for determining one or more dimension estimations for a vehicle, the system comprising:

a processor configured to perform various functions and operations; and

a non-transitory computer readable medium in communication with the processor, the computer readable medium comprising one or more programming instructions for causing the processor to;

receive a sequence of image frames of a vehicle,

extract a digital representation of the vehicle from each image in the sequence of image frames,

identify a moving axis of the vehicle based on at least two image frames in the sequence,

extract two or more points where two or more tires of a side of the vehicle contact a supporting surface, wherein the extracting the two or more points comprises detecting an edge or subtracting a background of each image, and wherein the two or more points form a line that runs along the identified moving axis of the vehicle,

determine a bounding box for the vehicle based at least partially upon the two or more extracted points for the digital representation of the vehicle,

estimate at least one dimension for the vehicle based upon the determined bounding box, and

output an indication of the at least one estimated dimension.

11. The system of claim 10, wherein the one or more instructions for causing the processor to output the indication comprise one or more additional instructions for causing the processor to transfer the indication of the at least one estimated dimension as an input to a comparison technique for further processing the indication of the at least one estimated dimension.

12. The system, of claim 11, wherein the comparison technique comprises a parking determination process configured to determine a recommended spot for the vehicle based upon the indication of the at least one estimated dimension.

13. The system of claim 10, wherein the one or more instructions for causing the processor to determine the bounding box further comprise one or more additional instructions for causing the processor to identify a line associated with the point where the vehicle contacts the supporting surface.

14. The system of claim 13, wherein the one or more instructions for causing the processor to identify the line associated with the point where the vehicle contacts the supporting surface comprise one or more additional instructions for causing the processor to:

determine two or more points where each tire of a near side of the vehicle contacts the supporting surface; and

draw a line that includes a segment that extends between each of the points.

15. The system of claim 13, wherein the one or more instructions for causing the processor to determine the bounding box further comprise one or more additional instructions for causing the processor to identify a far right point of the digital representation of the vehicle and a far left point of the digital representation of the vehicle.

16. The system of claim 10, wherein the one or more instructions for causing the processor to estimate at least one dimension of the vehicle comprise one or more additional instructions for causing the processor to estimate length and width of the vehicle based upon a bottom edge of the boundary box.

17. The system of claim 16, wherein the one or more instructions for causing the processor to estimate at least one dimension of the vehicle comprise one or more additional instructions for causing the processor to estimate a height of the vehicle based upon a distance measured between a rear vertex of the bounding box and an upper vertex of the bounding box.

18. The system of claim 10, wherein the one or more instructions for causing the processor to extract the digital representation of the vehicle comprise one or more additional instructions for causing the processor to perform at least one of example background subtraction, particle filtering, and a shadow removal algorithm.

22. The system of claim 10, wherein the instructions for extracting the two or more points comprise additional instructions that will analyze contrast ratios at areas where the two or more tires are most likely to be placed.

25. The system of claim 12, wherein the instructions for determining the recommended spot comprises additional instructions that will compare the indication of the at least one estimated dimension against a listing of dimensions of available parking spaces.

19. A method of determining one or more dimension estimations for a vehicle, the method comprising:

receiving, by a processor, a sequence of image frames of a vehicle;

extracting, by the processor, a plurality of digital representations of the vehicle from each image in the sequence of image frames, wherein each of the digital representations is associated with one image frame;

identifying a moving axis of the vehicle based on at least two image frames in the sequence;

determining, by the processor, a bounding box in each of the plurality of digital representations for the vehicle, wherein determining the bounding box comprises;

identifying a first line based at least partially upon extracting two or more points where two or more tires of a side of the vehicle contact a supporting surface, wherein the extracting of the two or more points comprises detecting an edge or subtracting a background of each image, wherein the first line runs along the identified moving axis of the vehicle,

identifying a second line based upon an identified far right point of the vehicle as determined from the extracted plurality of digital representations, wherein the second line is drawn perpendicular to the supporting surface and includes the far right point,

identifying a third line based upon an identified far left point of the vehicle as determined from the extracted plurality of digital representations, wherein the third line is drawn perpendicular to the supporting surface and includes the far left point, and

identifying a fourth line connecting the second line and the third line and draw essentially parallel to a top surface of the vehicle;

estimating, by the processor, at least one dimension for the vehicle based upon the determined bounding box; and

outputting, by the processor, an indication of the at least one estimated dimension.

20. The method of claim 19, wherein outputting the indication comprises transferring the indication of the at least one estimated dimension as an input to a comparison technique for further processing the indication of the at least one estimated dimension.

23. The method of claim 19, wherein extracting the two or more points comprises analyzing contrast ratios at areas where the two or more tires are most likely to be placed.