Image processing apparatus and method

• US 9,076,037 B2
• Filed: 05/22/2014
• Issued: 07/07/2015
• Est. Priority Date: 06/28/2013
• Status: Active Grant

• Current Assignee: JVC Kenwood Corporation
• Original Assignee: JVC Kenwood Corporation
• Inventors: Nakagami, Atsushi
• Primary Examiner(s): Koziol, Stephen R
• Assistant Examiner(s): Azima, Shaghayegh
• Application Number:
  US14/284,852
• Publication Number:
  US 20150003675A1
• Time in Patent Office: 411 Days
• Field of Search: 382/103
• US Class Current: 1/1
• CPC Class Codes: G06F 2218/08 Feature extraction
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/30232 Surveillance
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69

Abstract

A background image holding unit is configured to hold a background image. A threshold table holding unit is configured to hold a threshold. A previous frame image holding unit is configured to hold a previous frame image. An object extraction unit is configured to perform a background subtraction process of calculating difference values between the background image and a latest image and detecting a pixel whose calculated difference value is equal to or larger than the threshold as an extracted object, and configured to perform a process of judging the magnitude of difference between the background image and latest image. The object extraction unit is configured to update the background image and threshold in accordance with the magnitude of difference between the background image and latest image.

Claims

1. An image processing apparatus comprising:

a background image holding unit configured to hold a background image;

a threshold table holding unit configured to hold a threshold;

a previous frame image holding unit configured to hold a previous frame image which is a frame image temporally prior to the latest image; and

an object extraction unit configured to perform a background subtraction process of calculating difference values between the background image and latest image and detecting a pixel whose calculated difference value is equal to or larger than the threshold as an extracted object and a process of judging the magnitude of difference between the background image and latest image, wherein

the object extraction unit is configured to

when determining that the magnitude of difference between the background image and latest image is less than a predetermined value, update the background image by adding the latest image to the background image in a predetermined ratio,

when determining that the magnitude of difference between the background and latest image is not less than the predetermined value, update the threshold by calculating a difference average which is the average of the difference values between the previous frame image and the latest image and adding the absolute value of the calculated difference average to the threshold,

update the background image by replacing the pixels of the latest image which are not recognized as the extracted object as the result of execution of the background subtraction process using the updated threshold, for the corresponding pixels of the background image as pixels of a new background image and replacing pixels of the background image corresponding to the pixels which are recognized as the extracted object with the pixels obtained by adding the difference average to the background image as pixels of the new background image,

the background image holding unit holds the background image updated by the object extraction unit instead of the background image that the background image holding unit has held until that time, and

the threshold table holding unit holds the threshold updated by the object extraction unit instead of the threshold that the threshold table holding unit has held until that time.

2. The image processing apparatus according to claim 1, wherein, when the extracted object is located in the previous frame image held by the previous frame image holding unit, the object extraction unit removes the pixels that have the largest difference values between the previous frame image and the latest image, the number of removed pixels being equal to a number previously set based on the number of pixels of the extracted object, calculates the difference values for the residual pixels, and calculates an average of the calculated difference values as the difference average.

3. The image processing apparatus according to claim 1, wherein the object extraction unit determines that the magnitude of difference between the background image and the latest image is less than the predetermined value when the proportion of the number of pixels detected as the extracted object is less than a predetermined proportion and determines that the magnitude of difference between the background image and the latest image is not less than the predetermined value when the proportion of the number of pixels detected as the extracted object is not less than the predetermined proportion.

4. An image processing method, comprising:

calculating difference values between a background image and a latest image;

detecting a pixel whose calculated difference value is not less than a predetermined threshold as an extracted object;

judging the magnitude of difference between the background image and the latest image based on the difference values;

when determining that the magnitude of difference between the background image and the latest image is less than a predetermined value, updating the background image by adding the background image to the latest image in a predetermined ratio;

when determining that the magnitude of difference between the background and the latest image is not less than the predetermined value, updating the threshold by calculating a difference average which is the average of the difference value of each pixel between the previous frame image and the latest image and adding the absolute value of the calculated difference average to the threshold; and

updating the background image by replacing the pixels of the latest image which are not recognized as the extracted object as the result of execution of the background subtraction process using the updated threshold, for the corresponding pixels of the background image as pixels of a new background image and replacing the pixels of the background image corresponding to the pixels which are recognized as the extracted object in the latest image, with pixels obtained by adding the difference average to the background image as pixels of the new background image.