System and method for automatically measuring the dimensions of and identifying the type of exterior siding

• US 10,074,171 B1
• Filed: 09/06/2017
• Issued: 09/11/2018
• Est. Priority Date: 03/28/2014
• Status: Active Grant

• Current Assignee: State Farm Mutual Automobile Insurance Company (State Farm)
• Original Assignee: State Farm Mutual Automobile Insurance Company (State Farm)
• Inventors: Maurer, Jonathan D., Christopulos, Nicholas U.
• Primary Examiner(s): Hung, Yubin
• Application Number:
  US15/696,589
• Time in Patent Office: 370 Days
• CPC Class Codes: G06T 2207/10004 Still image; Photographic image
• CPC Class Codes: G06T 2207/20048 Transform domain processing
• CPC Class Codes: G06T 2207/20061 Hough transform
• CPC Class Codes: G06T 2207/30108 Industrial image inspection
• CPC Class Codes: G06T 7/001 using an image reference approach
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06T 7/60 Analysis of geometric attributes
• CPC Class Codes: G06T 7/62 of area, perimeter, diameter or volume
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• CPC Class Codes: G06V 10/245 by locating a pattern; Special marks for positioning

Abstract

Methods, systems, and computer readable media are disclosed for determining a pixel-to-length ratio between a number of pixels disposed over a predetermined length of a reference object within an image of a siding sample and the predetermined length of the reference object. A first and second distance between respective first and second pairs of points within the image corresponding to respective first and second length measurements of the siding sample are determined, as well as a first and second number of pixels disposed between the first and second pair of points, respectively. Furthermore, the method, system, and computer readable medium disclose determining the first length measurement based on the pixel-to-length ratio and the first number of pixels, determining the second length measurement based on the pixel-to-length ratio and the second number of pixels, and identifying a siding product associated with the first and second length measurements.

Claims

1. A computer-implemented method, comprising:

selectively aligning, via one or more processors, edges of a reference object included within an image of an object with edges of the object to provide an aligned reference object;

calculating, via one or more processors, a first dimension of the object based upon a known reference dimension of the aligned reference object and a first number of pixels spanned by the first dimension of the object within the image; and

identifying, via one or more processors, a product associated with the first dimension of the object.

2. The method of claim 1, wherein the act of selectively aligning the edges of the reference object comprises:

determining, via one or more processors, that the reference object needs to be aligned when the reference object is not parallel with a portion of the object within the image.

3. The method of claim 1, further comprising:

rectifying, via one or more processors, the image of the object so the perspective of the object faces forward if the reference object includes perspective distortion.

4. The method of claim 1, further comprising:

calculating, via one or more processors, a second dimension of the object based upon the known reference dimension of the aligned reference object and a second number of pixels spanned by the second dimension of the object within the image; and

identifying the product based upon the first and second dimensions of the object.

5. The method of claim 4, wherein:

the object is a siding sample,

the first dimension of the object corresponds to a face dimension of the siding sample,

the second dimension of the object corresponds to a profile dimension of the siding sample, and

the act of identifying the product comprises;

identifying the product as a siding product stored in a siding product reference database having a face dimension and a profile dimension matching the first dimension and the second dimension of the siding sample, respectively.

6. The method of claim 1, wherein the act of calculating the first dimension of the object comprises:

calculating, via one or more processors, a pixel-to-length ratio representative of a ratio between (i) a reference number of pixels disposed over a reference dimension of the aligned reference object within the image, and (ii) the known reference dimension of the aligned reference object; and

dividing the first number of pixels spanned by the first dimension of the object within the image by the pixel-to-length ratio to calculate the first dimension of the object.

7. The method of claim 1, further comprising:

detecting, via one or more processors, the reference object within the image using at least one of;

a Scale-Invariant Feature Transform (SIFT);

a Speeded Up Robust Features (SURF);

a Radon transform;

or

a Hough transform.

8. A product identification engine, comprising:

a memory configured to store computer-readable instructions; and

a processor configured to execute the computer-readable instructions stored in the memory to;

selectively align edges of a reference object included within an image of an object with edges of the object to provide an aligned reference object;

calculate a first dimension of the object based upon a known reference dimension of the aligned reference object and a first number of pixels spanned by the first dimension of the object within the image; and

identify a product associated with the first dimension of the object.

9. The product identification engine of claim 8, wherein the processor is further configured to determine that the reference object needs to be aligned when the reference object is not parallel with a portion of the object within the image.

10. The product identification engine of claim 8, wherein the processor is further configured to rectify the image of the object so the perspective of the object faces forward if the reference object includes perspective distortion.

11. The product identification engine of claim 8, wherein the processor is further configured to calculate a second dimension of the object based upon the known reference dimension of the aligned reference object and a second number of pixels spanned by the second dimension of the object within the image, and to identify the product based upon the first and second dimensions of the object.

12. The product identification engine of claim 11, wherein:

the first dimension of the object corresponds to a face dimension of the siding sample,

the second dimension of the object corresponds to a profile dimension of the siding sample, and

the processor is further configured to identify the product as a siding product stored in a siding product reference database having a face dimension and a profile dimension matching the first dimension and the second dimension of the siding sample, respectively.

13. The product identification engine of claim 8, wherein the processor is further configured to detect the reference object within the image using at least one of:

a Scale-Invariant Feature Transform (SIFT);

a Speeded Up Robust Features (SURF);

a Radon transform;

or

a Hough transform.

14. A non-transitory, tangible computer-readable medium storing machine readable instructions that, when executed by a processor, cause the processor to:

selectively align edges of a reference object included within an image of an object with edges of the object to provide an aligned reference object;

calculate a first dimension of the object based upon a known reference dimension of the aligned reference object and a first number of pixels spanned by the first dimension of the object within the image; and

identify a product associated with the first dimension of the object.

15. The non-transitory, tangible computer-readable medium of claim 14, wherein the instructions to selectively align the edges of the reference object further include instructions that, when executed by the processor, cause the processor to determine that the reference object needs to be aligned when the reference object is not parallel with a portion of the object within the image.

16. The non-transitory, tangible computer-readable medium of claim 14, further storing instructions that, when executed by the processor, cause the processor to rectify the image of the object so the perspective of the object faces forward if the reference object includes perspective distortion.

17. The non-transitory, tangible computer-readable medium of claim 14, further storing instructions that, when executed by the processor, cause the processor to calculate a second dimension of the object based upon the known reference dimension of the aligned reference object and a second number of pixels spanned by the second dimension of the object within the image, and to identify the product based upon the first and second dimensions of the object.

18. The non-transitory, tangible computer-readable medium of claim 17, wherein:

the first dimension of the object corresponds to a face dimension of the siding sample,

the second dimension of the object corresponds to a profile dimension of the siding sample, and

the instructions to identify the product further include instructions that, when executed by the processor, cause the processor to identify the product as a siding product stored in a siding product reference database having a face dimension and a profile dimension matching the first dimension and the second dimension of the siding sample, respectively.

19. The non-transitory, tangible computer-readable medium of claim 14, wherein the instructions to calculate the first dimension of the object further include instructions that, when executed by the processor, cause the processor to:

calculate a pixel-to-length ratio representative of a ratio between (i) a reference number of pixels disposed over the reference dimension of the aligned reference object within the image, and (ii) the known reference dimension of the aligned reference object; and

divide the first number of pixels spanned by the first dimension of the object within the image by the pixel-to-length ratio to calculate the first dimension of the object.

20. The non-transitory, tangible computer-readable medium of claim 14, further storing instructions that, when executed by the processor, cause the processor to detect the reference object within the image using at least one of:

a Scale-Invariant Feature Transform (SIFT);

a Speeded Up Robust Features (SURF);

a Radon transform;

or

a Hough transform.