Method for detecting shadow of object

• US 8,189,913 B2
• Filed: 01/07/2009
• Issued: 05/29/2012
• Est. Priority Date: 11/06/2008
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Huang, Chung-Hsien, Wu, Ruei-Cheng
• Primary Examiner(s): Chen, Wenpeng
• Application Number:
  US12/350,209
• Publication Number:
  US 20100111362A1
• Time in Patent Office: 1,238 Days
• Field of Search: 382/170, 382/164, 382/103
• US Class Current: 382/170
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 10/50 by performing operations within image blocks; by using histograms, e.g. histogram of oriented gradients [HoG]; by summing image-intensity values; Projection analysis

Abstract

A method for detecting a shadow of an object in an image is provided. A moving object in a plurality of continuous images is detected. A histogram of a color variation of the moving object in each of the images is calculated. The histograms of the color variation are accumulated to obtain a cumulative histogram. A distribution of the color variation in the cumulative histogram is estimated to obtain a shadow distribution function. Whether each pixel in a received image belongs to the shadow is determined by using the shadow distribution function.

Claims

1. A shadow detection method, comprising:

detecting a moving object in a plurality of continuous images;

calculating a histogram of a color variation of the moving object in each of the images;

accumulating the histograms of the color variation in the images to obtain a cumulative histogram;

estimating a distribution of the color variation in the cumulative histogram to obtain a shadow distribution function; and

determining whether each pixel in a received image belongs to a shadow by using the shadow distribution function.

2. The shadow detection method according to claim 1, wherein the step of detecting the moving object in the continuous images comprises:

capturing a background image; and

subtracting the background image from each of the images to obtain an object image of the moving object in the image.

3. The shadow detection method according to claim 2, wherein the step of subtracting the background image from each of the images to obtain the object image of the moving object in the image further comprises:

labeling a plurality of connected components in the images after the background image is subtracted from the images.

4. The shadow detection method according to claim 2, wherein the step of calculating the histogram of the color variation of the moving object in each of the images comprises:

subtracting images in same regions of the background image from each of the object images to obtain the color variation of each pixel between the moving object and the background image; and

calculating a distribution of the color variation of the pixels in the moving object to obtain the histogram of the color variation of the moving object.

5. The shadow detection method according to claim 4, wherein the step of accumulating the histograms of the color variation in the images to obtain the cumulative histogram comprises adding, multiplying, or averaging the corresponding color variation in the histograms of the color variation in the images to obtain the cumulative histogram.

6. The shadow detection method according to claim 1, wherein the step of estimating the distribution of the color variation in the cumulative histogram to obtain the shadow distribution function comprises:

estimating a Gaussian function curve approximate to a distribution curve of the color variation in a peak region of the cumulative histogram, and serving a function of the Gaussian function curve as the shadow distribution function.

7. The shadow detection method according to claim 6, wherein the step of estimating the similar Gaussian function curve comprises estimating the similar Gaussian function curve through a robust estimation.

8. The shadow detection method according to claim 1, wherein the step of determining whether each pixel in the received image belongs to the shadow by using the shadow distribution function comprises:

subtracting a background image from the received image to obtain an object image of the moving object;

determining whether a pixel value of each pixel in the object image falls within a pixel value range defined by the shadow distribution function;

determining that the pixel belongs to the shadow of the moving object if the pixel value of the pixel falls within the pixel value range; and

determining that the pixel belongs to the moving object if the pixel value of the pixel does not fall within the pixel value range.

9. The shadow detection method according to claim 8, wherein the pixel value range is contained by a pixel value range from a pixel value corresponding to a peak of the shadow distribution function minus a multiple of a standard deviation of the shadow distribution function to the pixel value corresponding to the peak of the shadow distribution function plus the multiple of the standard deviation of the shadow distribution function.

10. The shadow detection method according to claim 1, wherein after the step of detecting the moving object in the continuous images, the shadow detection method further comprises:

transforming a color space of the moving object in the images.

11. The shadow detection method according to claim 10, wherein the step of transforming the color space of the moving object in the images comprises:

transforming the color space of the moving object in the images into an intensity (I) space, a saturation (S) space, and a hue (H) space.

12. The shadow detection method according to claim 11, wherein the step of calculating the histogram of the color variation of the moving object in each of the images comprises:

respectively calculating the histogram of the color variation of the moving object in each of the images regarding the intensity space, the saturation space, and the hue space to obtain an intensity histogram, a saturation histogram, and a hue histogram.

13. The shadow detection method according to claim 12, wherein the step of accumulating the histograms of the color variation in the images to obtain the cumulative histogram comprises:

respectively accumulating the intensity histograms, the saturation histograms, and the hue histograms of the color variation in the images to obtain a cumulative intensity histogram, a cumulative saturation histogram, and a cumulative hue histogram.

14. The shadow detection method according to claim 13, wherein the step of estimating the distribution of the color variation in the cumulative histogram to obtain the shadow distribution function comprises:

respectively estimating the distribution of the color variation in the cumulative intensity histogram, the cumulative saturation histogram, and the cumulative hue histogram to obtain an intensity shadow distribution function, a saturation shadow distribution function, and a hue shadow distribution function.

15. The shadow detection method according to claim 14, wherein the step of determining whether each pixel in the received image belongs to the shadow by using the shadow distribution function comprises:

subtracting a background image from the received image to obtain an object image of the moving object;

transforming a color space of the object image into the intensity space, the saturation space, and the hue space to obtain an intensity value, a saturation value, and a hue value of each pixel in the object image;

determining whether the intensity value of each pixel in the object image falls within an intensity value range defined by the intensity shadow distribution function;

determining whether the saturation value of the pixel falls within a saturation value range defined by the saturation shadow distribution function if the intensity value of the pixel falls within the intensity value range;

determining whether the hue value of the pixel falls within a hue value range defined by the hue shadow distribution function if the saturation value of the pixel falls within the saturation value range; and

determining that the pixel belongs to the shadow of the moving object if the hue value of the pixel falls within the hue value range.

16. The shadow detection method according to claim 15, wherein the step of determining whether each pixel in the received image belongs to the shadow by using the shadow distribution function further comprises:

determining that the pixel belongs to the moving object if the intensity value of the pixel does not fall within the intensity value range, the saturation value of the pixel does not fall within the saturation value range, or the hue value of the pixel does not fall within the hue value range.

17. The shadow detection method according to claim 10, wherein the step of transforming the color space of the moving object in the images comprises:

transforming the color space of the moving object in the images into an intensity (Y) space, a blue (Cb) space, and a red (Cr) space.