Determining camera height using distributions of object heights and object image heights

Patent metadata

Field	Value
Publication number	US-9322647-B2
Application number	US-201313851991-A
Country	US
Kind code	B2
Filing date	Mar 28, 2013
Priority date	Mar 28, 2013
Publication date	Apr 26, 2016
Grant date	Apr 26, 2016

Abstract

A camera at a fixed vertical height positioned above a reference plane, with an axis of a camera lens at an acute angle with respect to a perpendicular of the reference plane. One or more processors receive images of different people. The vertical measurement values of the images of different people are determined. The one or more processors determine a first statistical measure associated with a statistical distribution of the vertical measurement values. The known heights of people from a known statistical distribution of heights of people are transformed to normalized measurements, based in part on a focal length of the camera lens, the angle of the camera, and a division operator in an objective function of differences between the normalized measurements and the vertical measurement values. The fixed vertical height of the camera is determined, based at least on minimizing the objective function.

First claim

What is claimed is:

1. A method for determining a height of a camera, the camera at an unknown fixed vertical height positioned above a reference plane, and having an axis of a lens of the camera at an acute angle with respect to a perpendicular of the reference plane, the method comprising the steps of:
receiving from the camera, by one or more processors, over a specified time period, a multiplicity of images of a multiplicity of people of unknown height within a field of view of the camera;
determining, by one or more processors, vertical measurement values of the multiplicity of images of the multiplicity of people of unknown height, wherein each vertical measurement value includes a pixel count along a vertical axis of an image of the multiplicity of images of the multiplicity of people of unknown height;
receiving, by one or more processors, heights of a multiplicity of people from a statistical distribution of known heights of people;
transforming, by one or more processors, the known heights of the multiplicity of people from the statistical distribution of known heights of people to normalized vertical measurement values that includes a pixel count, based at least on a focal length of the lens of the camera, the acute angle of the camera, and a division operator present in an objective function;
performing, by one or more processors, the objective function to sum differences between the pixel count of the vertical measurement values of the multiplicity of images and the pixel count of the normalized vertical measurement values of the heights of the multiplicity of people from the statistical distribution of the known heights of people, wherein the objective function includes an estimate of the fixed vertical height of the camera, and a transformation matrix to convert the known heights of people to normalized vertical measurement values; and
determining, by one or more processors, the fixed vertical height of the camera by adjusting within the objective function, the estimate of the fixed vertical height of the camera, until a result of the objective function is minimized.

2. The method of claim 1 , wherein transforming, by one or more processors, the heights of the multiplicity of people from the statistical distribution of known heights of people to normalized vertical measurement values that includes a pixel count, comprises:
generating, by one or more processors, a multiple row transformation matrix that includes the focal length of the lens of the camera, the acute angle of the camera, and a current estimated camera height;
transforming, by one or more processors, normalized vertical measurement values that includes a pixel count of the known heights of the multiplicity of people, based on a product of the multiple row transformation matrix and each respective known height of the multiplicity of people, including a division operator from a row of the multiple row transformation matrix; and
generating, by one or more processors, an error term as the result of the objective function that sums the differences between the pixel count of the normalized vertical measurement values of the known heights of the multiplicity of people and the pixel count of the vertical measurement values of the multiplicity of images of the multiplicity of people of unknown height.

3. The method of claim 2 , further comprising: determining, by one or more processors, a new estimated camera height by adjusting the current estimated camera height, based at least on the error term;
determining, by one or more processors, if a threshold value exceeds a value of the error term; and
in response to determining the threshold value to exceed the value of the error term, one or more processors determining the new estimated camera height to be the fixed vertical height of the camera.

4. The method of claim 1 , wherein the multiplicity of people comprises a multiplicity of people of a known gender.

5. The method of claim 1 , wherein the multiplicity of people comprises a multiplicity of people of known age groups.

6. The method of claim 1 , wherein the vertical measurement values of the images of the multiplicity of people and the normalized vertical measurement values of the known heights of people from the known statistical distribution of heights of people, are nonparametric.

7. The method of claim 1 , wherein the heights of the multiplicity of people from the statistical distribution of known heights of people is one of: a statistical median, a statistical mode, or a statistical average of the statistical distribution of the known heights of people.

8. A computer program product for determining a height of a camera, the camera at an unknown fixed vertical height positioned above a reference plane, and having an axis of a lens of the camera at an acute angle with respect to a perpendicular of the reference plane, the computer program product comprising:
one or more non-transitory computer-readable storage media and program instructions stored on the one or more non-transitory computer-readable storage media, the program instructions comprising:
program instructions to receive from the camera, over a specified time period, a multiplicity of images of a multiplicity of people of unknown height, within a field of view of the camera;
program instructions to determine vertical measurement values of the multiplicity of images from the camera of the multiplicity of people of unknown height, wherein each vertical measurement value includes a pixel count along a vertical axis of an image of the multiplicity of images of the multiplicity of people of unknown height;
program instructions to receive heights of a multiplicity of people from a statistical distribution of known heights of people;
program instructions to transform the known heights of the multiplicity of people from the statistical distribution of known heights of people to normalized vertical measurement values that includes a pixel count based at least on a focal length of the lens of the camera, the acute angle of the camera, and a division operator present in an objective function;
program instructions to perform the objective function to sum differences between the pixel count of the vertical measurement values of the multiplicity of images and the pixel count of the normalized vertical measurement values of heights of the multiplicity of people from the distribution data of the known heights of people, wherein the objective function includes an estimate of the fixed vertical height of the camera, and a transformation matrix to convert the known heights of people to normalized vertical measurement values; and
program instructions to determine the fixed vertical height of the camera by adjusting within the objective function, the estimate of the fixed vertical height of the camera, until a result of the objective function is minimized.

9. The computer program product of claim 8 , wherein program instructions to transform the heights of the multiplicity of people from the statistical distribution of known heights people to normalized vertical measurement values that includes a pixel count, further comprises:
program instructions to generate a multiple row transformation matrix that includes the focal length of the lens of the camera, the acute angle of the camera, and a current estimated camera height;
programming instructions to generate normalized vertical measurement values that includes a pixel count of the known heights of the multiplicity of people, based on a product of the multiple row transformation matrix and each respective known heights height of the multiplicity of people, including a division operator from a row of the multiple row transformation matrix; and
program instructions

Assignees

• IBM

Inventors

• Datta Ankur
• Feris Rogerio S
• Pankanti Sharathchandra U
• Zhai Yun

Classifications

• G06F18/22

  Matching criteria, e.g. proximity measures · CPC title

• G01C5/00 Primary

  Measuring height; Measuring distances transverse to line of sight; Levelling between separated points; Surveyors' levels (G01C3/20, G01C3/30 take precedence) · CPC title

• H04N7/18

  Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast · CPC title

• G06T7/60

  Analysis of geometric attributes · CPC title

• G06T2207/30196

  Human being; Person · CPC title