Method for identifying change of scenery and corresponding monitoring device

• US 7,130,468 B1
• Filed: 07/11/2000
• Issued: 10/31/2006
• Est. Priority Date: 07/15/1999
• Status: Expired due to Term

• Current Assignee: Robert Bosch GmbH
• Original Assignee: Robert Bosch GmbH
• Inventors: Dreves, Jens, Meyer, Michael, Hoetter, Michael
• Primary Examiner(s): Wu, Jingge
• Assistant Examiner(s): Hung, Yubin
• Application Number:
  US10/030,832
• Time in Patent Office: 2,303 Days
• Field of Search: 382/103, 382/104, 382/199, 382/209, 382218-220, 382/208, 382/305, 382/278, 340/541, 340/937, 340/938, 348/143, 348152-155
• US Class Current: 382/219
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G08B 13/19602 Image analysis to detect motion of the intruder, e.g. by frame subtraction
• CPC Class Codes: G08B 13/19606 Discriminating between target movement or movement in an area of interest and other non-signicative movements, e.g. target movements induced by camera shake or movements of pets, falling leaves, rotating fan
• CPC Class Codes: H04N 5/144 Movement detection for video coding H04N19/503; analysis of motion in general G06T7/20

Abstract

A method for detecting changes in a viewfield observed by a stationary image capturing device is proposed, in which edge images are calculated and are compared to edge images of reference recordings in order to detect static changes within the observed image range that are independent of image brightness.

Claims

1. A method of monitoring a predetermined scene to detect variations in the scene, said method comprising the steps of:

a) establishing an edge image and a reference edge image based on a reference image and a current image detected in a view field of a stationary image capturing device;

b) calculating a correlation value for at least one region of the edge image and a corresponding partial image of the reference edge image;

c) identifying a change of the at least one region when the correlation value falls below a threshold value; and

d) outputting an alarm signal when the at least one region is identified as changed for a predetermined time interval that is longer than a corresponding time interval that the edge image is detected to be changed, further comprising calculating an average quadratic deviation of said at least one region of the edge image and said corresponding partial image of the reference edge image when said correlation value exceeds said threshold value, so that possible brightness differences between said at least one region of the edge image and said corresponding partial image of the reference edge image are eliminated from the calculating;

whereby deviations of said at least one region of the edge image from said corresponding partial image of the reference edge image that do not result from image noise and said brightness differences are distinguished.

2. The method as defined in claim 1, further comprising computing current image noise and storing said current image noise for subsequent image comparisons when none of said deviations that do not result from said image noise and said brightness differences are distinguished.

3. A monitoring system with a computer (2) for processing image data from a stationary image capturing device (1) that is aimed at a predetermined viewfield, characterized in that based on a reference image of the viewfield and a current image of the viewfield, the computer (2) establishes a reference edge image and an edge image, that the computer determines a correlation of at least one region of the edge image with a corresponding partial image of the reference edge image, and that when the correlation falls below a threshold, the computer registers the region as changed, and further characterized in that when the threshold is exceeded, the computer determines the difference of the region of the edge image from the corresponding partial image of the reference edge image, where a possible brightness difference between the reference edge image and the edge image is eliminated from the calculation so that by computing the image noise, a determination can be made as to whether there is a deviation of the region of the edge image from the corresponding partial image of the reference edge image that does not result from the image noise and the brightness difference.

4. The monitoring system according to claim 3, characterized in that the computer determines the correlation of the edge image with the reference edge image, that the computer is connected to a signal indicator (5) so that an alarm is output via the signal indicator (5) when the at least one region is detected as changed for a predetermined time (T1) longer than the edge image is detected as changed.

5. The monitoring system according to claim 4, characterized in that an alarm is output via the signal indicator (5) when the at least one region of the edge image is detected as changed for longer than an additional predetermined time (T2), where the additional predetermined time (T2) is greater than the predetermined time (T1).

6. The monitoring system according to claim 3, characterized in that the image capturing device is a video camera.

7. A monitoring system for monitoring a predetermined scene to detect variations in the scene, said monitoring system comprising

means for generating a current image in a view field;

means for establishing an edge image and a reference edge image based on a reference image and the current image detected in the view field;

means for calculating a correlation value for at least one region of the edge image and a corresponding partial image of the reference edge image;

means for identifying a change of the at least one region when the correlation value falls below a threshold value; and

means for outputting an alarm signal when the at least one region is identified as changed for a predetermined time interval that is longer than a corresponding time interval that the edge image is detected to be changed;

wherein said means for generating said current image comprises a stationary image capturing device;

said means for calculating said correlation value comprises a computer for processing image data and said computer includes means for testing whether or not said correlation value is above or below said threshold value in order to identify said change, further comprising means for calculating an average quadratic deviation of said at least one region of the edge image and said corresponding partial image of the reference edge image when said correlation value exceeds said threshold value, so that possible brightness differences between said at least one region of the edge image and said corresponding partial image of the reference edge image are eliminated from the calculating.