Method and system for determining a volume of an object from two-dimensional images

• US 7,773,773 B2
• Filed: 10/18/2006
• Issued: 08/10/2010
• Est. Priority Date: 10/18/2006
• Status: Expired due to Fees

• Current Assignee: UT-Battelle LLC
• Original Assignee: UT-Battelle LLC
• Inventors: Schlicher, Bob G., Abercrombie, Robert K.
• Primary Examiner(s): Werner; Brian P
• Application Number:
  US11/583,473
• Publication Number:
  US 20080095404A1
• Time in Patent Office: 1,392 Days
• Field of Search: 382/104, 382/286
• US Class Current: 382/104
• CPC Class Codes: G06T 2207/20101 Interactive definition of point of interest, landmark or seed
• CPC Class Codes: G06T 7/62 of area, perimeter, diameter or volume
• CPC Class Codes: G06V 10/48 by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation

Abstract

The invention provides a method and a computer program stored in a tangible medium for automatically determining a volume of three-dimensional objects represented in two-dimensional images, by acquiring at two least two-dimensional digitized images, by analyzing the two-dimensional images to identify reference points and geometric patterns, by determining distances between the reference points and the component objects utilizing reference data provided for the three-dimensional object, and by calculating a volume for the three-dimensional object.

Claims

1. A computer program stored in a non-transitory medium from which the program can be executed by a computer, for determining a volume for a three-dimensional vehicle from at least two digitized images of the three-dimensional vehicle, the computer program further comprising:

an image acquisition group of instructions for receiving at least two two-dimensional images of the three-dimensional vehicle in a computer-readable format, wherein a first image of the two-dimensional images is acquired in a first view and a second image of the two-dimensional images is acquired for a second view that is a depth or width view relative to the first view;

a feature detector group of instructions that are executed to detect a shape and dimensions of features, including features that are smaller than the two-dimensional images of the vehicle;

a reference detector group of instructions that are executed to determine dimensions between reference points detected in, or manually entered for, the two-dimensional images of the vehicle;

a display group of instructions for displaying images of a vehicle on a visual display, including a display of symbols allowing manual entry of data representing at least one of;

a vehicle model and dimensions for selected features of the vehicle; and

a measurement calculator group of instructions that are executed to determine measurements about the three-dimensional vehicle in the image including a volume of the three-dimensional vehicle, from the dimensions of automatically detected features and from the data that has been manually entered into the computer.

2. The computer program of claim 1, wherein the vehicle is a wheeled vehicle;

wherein said features include but not limited to wheels, front axle position and a rear axle position;

wherein one reference is a distance between a front axle and a rear axle.

3. The computer program of claim 1, further comprising an adaptive learning group of instructions that determines a quality of results from executing the feature detector, reference detector and measurement calculator groups of instructions and causes these groups of instructions to be repeated one or more additional times to determine a statistical degree of certainty for identifying a particular three-dimensional object.

4. The computer program of claim 1, wherein the two-dimensional images are acquired from at least one of a digital camera, a video camera and a scanner.

5. The computer program of claim 4, wherein the two-dimensional images are acquired from image data in a non-visible spectrum.

6. The computer program of claim 1, wherein the two-dimensional images are acquired in at least one of a bit-mapped graphics format, a vector graphics format, a JPEG format or a PNG format.

7. The computer program of claim 1, wherein the vehicle is a wheeled vehicle;

and wherein the feature detector group of instructions includes instructions for receiving the two-dimensional images from said image acquisition group of instructions and detects at least one vehicle feature in the image from a group of features consisting of wheels, wheel wells, axle positions, and vehicle bottom shape.

8. The computer program of claim 1, wherein the vehicle is a wheeled vehicle;

and wherein the feature detector group of instructions includes instructions for comparing detected vehicle features with stored vehicle pattern data.

9. The computer program of claim 1, wherein said reference detector group of instructions includes instructions for receiving features that have been identified by the feature detector group of instructions, and also includes instructions for identifying spatial relationships between those features, and also includes instructions for identifying and associating reference data to with said relationships.

10. The computer program of claim 1, wherein said reference detector group of instructions includes instructions for identifying spatial relationships among features using stored relationship pattern data.

11. The computer program of claim 1, wherein said measurement calculator group of instructions identifies relationships between measurement points on the vehicle and calculates measurement distances using reference data provided by executing the reference detector group of instructions.

12. The computer program of claim 1, wherein the vehicle is a wheeled vehicle;

and wherein said feature detector group of instructions are responsive to vehicle type data either stored in memory or input into the system to utilize vehicle type pattern data to detect features in the image.

13. The computer program of claim 1, further comprising a group of image correction instructions for correcting an image received through the image acquisition system before the image is evaluated.

14. The computer program of claim 13, wherein the group of image correction instructions is executed for correcting for barrel and pincushion distortions.

15. The computer program of claim 14, wherein the barrel and pincushion distortions are corrected in relation to stored digital lens parameters.

16. The computer program of claim 15, wherein said stored digital lens parameters are derived from a digital image of a Cartesian grid captured at different focal distances to the grid.

17. The computer program of claim 15, wherein said stored digital lens parameters are derived from a digital image of a measurement scale captured at different focal distances to the scale.

18. The computer program of claim 15, wherein said stored digital lens parameters are derived from a mathematical representation of lens distortion.

19. The computer program of claim 15, wherein said stored digital lens parameters are derived from lens distortion correction algorithms.

20. The computer program of claim 13, wherein the image is corrected for pixel aspect ratio distortion and perspective distortion in the digital image.

21. The computer program of claim 20, wherein the image corrector group of instructions utilizes two tables, corresponding directly to a plane of pixel positions of the image, and wherein each value in at least one of the two tables provides an index of a pixel position in the image and a corresponding value that is a displacement from said pixel position;

and wherein a value is calculated using any function that originates in a center focus point of the image and changes with a parameterized function radially from the center focus point to edges of the image in a manner defined by the function.

22. The computer program of claim 20, wherein the image corrector group of instructions utilizes two tables, and wherein one of the two tables represents a correction of pixel position for one dimension of the image and the other of the two tables represents the pixel position correction for the other dimension.

23. The computer program of claim 20, wherein the image corrector group of instructions utilizes two tables, and wherein each table is indexed by x values for pixel position with respect to a coordinate system, and wherein y values corresponding to the x values in at least one of two tables represents an amount of correction for the pixel position.

24. A computer-implemented method for automatically determining a volume of a three-dimensional vehicle represented in at least two two-dimensional images, the method comprising:

acquiring at least two two-dimensional digitized images of the vehicle in a computer-readable format, wherein a first image of the two-dimensional images is acquired in a side view and a second image of the two-dimensional images is acquired for a second view that is a depth or width view relative to the side view;

analyzing the two-dimensional images in a computer, such that reference points on at least one of the two two-dimensional images are automatically identified;

automatically determining distances between the reference points and component objects utilizing reference data provided for the three-dimensional vehicle;

automatically identifying the vehicle by model based on a distance between the reference points on the vehicle and data stored in the computer;

analyzing the two-dimensional images in the computer to automatically identify features in addition to outer boundaries of the vehicle in each two-dimensional view for use in calculating the volume of the vehicle; and

calculating a volume for the vehicle in the computer based on the features automatically identified in each two-dimensional view.

25. The method of claim 24, further comprising comparing geometric patterns in the two-dimensional images that represent components within the images to geometric patterns stored in the computer to determine an identity of the components and an identity of the three-dimensional object.

26. The method of claim 24, wherein a digital image is received from an image acquisition subsystem and at least one vehicle feature in the image is identified from a group of vehicle features consisting of wheels, wheel wells, axle positions, and vehicle bottom shape.

27. The method of claim 24, wherein the three-dimensional object is a wheeled vehicle;

wherein said geometric patterns include but are not limited to wheels, front axle position and a rear axle position;

wherein one reference position is a distance between a front axle and a rear axle.

28. The method of claim 24, further comprising repeating the analyzing of the two-dimensional images to identify reference points, then comparing of geometric patterns in the two-dimensional images and the determining of distances between the reference points and the component objects utilizing reference data provided for the three-dimensional object;

and then calculating a volume for the three-dimensional object to determine a statistical degree of certainty for identifying a particular three-dimensional object.

29. The method of claim 24, wherein a common reference measurement is located in the two two-dimensional digitized images to relate the images to one another and to allow calculation of a depth from one of the images.

31. The method of claim 24, wherein the two-dimensional images are acquired in at least one of a bit-mapped graphics format, a vector graphics format, a JPEG format or a PNG format.

32. The method of claim 24, wherein data for features that are automatically detected from the two-dimensional images is stored in memory along with data which is manually input for other components.

30. Computer-implemented method for determining a volume of a three-dimensional vehicle represented in at least two two-dimensional images, the method comprising:

acquiring at least two two-dimensional images of the vehicle in a computer-readable format, wherein a first one of the two-dimensional images is acquired in a plane transverse to a plane for a second one of the two-dimensional images;

analyzing the two-dimensional images in a computer, such that the computer automatically identifies reference points for dimensions of the vehicle which are to be determined;

comparing geometric patterns in the two-dimensional images that represent components within the images to stored geometric patterns to automatically detect an identity and dimensions of components less than outer boundaries of the vehicle;

manually inputting data into the computer to identify selected features, and corresponding dimensions to be determined for the selected features, to be and used in a calculation of volume for the vehicle; and

determining dimensions for the three-dimensional vehicle in the computer from the features having manually input data and from data for the automatically detected components to determine a three-dimensional volume for the vehicle.

33. The method of claim 30, wherein the two-dimensional images are acquired in at least one of a bit-mapped graphics format, a vector graphics format, a JPEG format or a PNG format.

34. The method of claim 30, wherein data for components that are automatically detected from the two-dimensional images is stored in memory along with data which is manually input for other components.