Method for rejection of flickering lights in an imaging system

• US 6,104,831 A
• Filed: 12/10/1998
• Issued: 08/15/2000
• Est. Priority Date: 12/10/1998
• Status: Expired due to Term

• Current Assignee: ADT Security Services SA (Apollo Global Management, Inc.)
• Original Assignee: ESCO Electronics Corporation
• Inventors: Ruland, Kevin
• Primary Examiner(s): Tran, Phuoc
• Application Number:
  US09/209,097
• Time in Patent Office: 614 Days
• Field of Search: 382/103, 382/107, 382/173, 382/199, 382/219, 382/286, 348/152, 348/153, 348/154, 348/155
• US Class Current: 382/173
• CPC Class Codes: G06V 10/751 Comparing pixel values or logical combinations thereof, or feature values having positional relevance, e.g. template matching
• CPC Class Codes: G06V 20/52 Surveillance or monitoring of activities, e.g. for recognising suspicious objects recognising microscopic objects G06V20/69
• CPC Class Codes: G08B 13/19602 Image analysis to detect motion of the intruder, e.g. by frame subtraction
• CPC Class Codes: G08B 13/19604 involving reference image or background adaptation with time to compensate for changing conditions, e.g. reference image update on detection of light level change
• 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

Abstract

A video surveillance system is disclosed which readily distinguishes between changes within an observed scene, caused by the presence of an intruder within the scene, and changes resulting from lighting variations. The video surveillance system implements a retinex-based algorithm for distinguishing the presence of a new object within the scene from changes in scene lighting. Areas of change between a current image (F1) of the scene and a reference image (RF) thereof are identified to produce a difference image. Response to significant differences present in the difference image, the entire current image and a subsequent image (F2) are partitioned into uniform segments (SF1₁ -SF1_n, SF2₁ -SF2_n), and ratios of changes in light intensity along the edges of the uniform segments are calculated for each partition. If a comparison of the resultant edge ratios between corresponding segments in the images indicates a change in the scene, the process is repeated for at least one subsequent image frame. If similar results are attained, the system signals an alarm condition to a monitoring station or a system operator.

Claims

1. In a security system distinguishing between changes within a scene caused by the presence of an intruder as opposed to ambient lighting fluctuations affecting scene illumination, the improvement of a method for making such a distinction when lighting changes within the scene are produced by flickering lights, comprising:

obtaining a first image of the scene;

comparing the first image with a reference image of the scene and obtaining therefrom a difference image composed of a set of differences between the two images;

responding to said set of differences exceeding a threshold by obtaining a second image of the scene subsequent to obtaining said first image;

comparing said second image with a reference image of the scene and obtaining therefrom a second difference image composed of a second set of differences between the two images;

comparing said second image with said first image and obtaining therefrom a third difference image composed of a third set of differences between the two images;

responding to said second and third sets of differences exceeding said threshold, identically partitioning said first and second images into a plurality of uniform segments; and

comparing a portion of said plurality of uniform segments from the first image with corresponding uniform segments from the second image to determine if any differences between the respective images result from changes in the irradiation of surfaces in the scene caused by fluctuations in ambient lighting of said scene.

2. In the method of claim 1, responding to said comparison determining the differences between the respective images as not resulting from changes in the irradiation of surfaces in the scene to obtain at least one subsequent image which is partitioned and then compared with at least one previously obtained image, a predetermined number of said subsequent image comparisons not indicative of changes in surface irradiation in said scene correspondingly indicating of an intrusion.

3. In the method of claim 1, each of said difference images being created by pixel subtraction.

4. The method of claim 1 wherein said identical partitioning of said first and second images into a plurality of uniform segments yields a grid of uniform segments.

5. The method of claim 1 wherein comparing a portion of said plurality of uniform segments from the first image with corresponding uniform segments from the second image to determine if any differences between the respective images result from changes in the irradiation of surfaces in the scene caused by fluctuations in ambient lighting of said scene includes:

identifying surface edges within each uniform segment;

determining ratios of reflectance for pairs of adjacent pixels separated by said identified edges; and

comparing said determined ratios from said first image uniform segments with determined ratios from corresponding pixel pairs in said second image uniform segments, the number of differences between said compared ratios less than a predetermined number indicating of a change in said ambient lighting of said scene.

6. A method of distinguishing between changes within a scene caused by the presence of an intruder as opposed to ambient lighting fluctuations affecting scene illumination, comprising:

viewing the scene to capture video images of the scene at different points in time;

processing each video image of the scene, said processing including comparing a captured video image representing the scene at one point in time with a captured video image representing the scene at a previous point in time, and accumulating differences between the captured video images;

responsive to the number of said accumulated differences exceeding a predetermined threshold for a sequence of captured video images, evaluating at least two captured video images to determine if the differences therebetween are caused by surface differences indicative of the presence of an intrusion within the scene, or illumination changes which do not indicate the presence of an intruder; and

wherein evaluating said captured video images includes;

(a) segmenting each entire image into a plurality of identically disposed uniform image segments;

(b) selecting from each segmented image a predetermined number of identically disposed uniform image segments for further processing;

(c) identifying within said selected image segments pixel pairs disposed across significant edges of surfaces contained within said segments;

(d) calculating the reflectance ratios of said adjacent pairs of pixels;

(e) comparing the calculated reflectance ratios for pixel pairs in each image segment selected from a first image with the corresponding pixel pair ratios for each image segment selected from a second image;

(f) determining the number of image segments wherein said compared reflectance ratios differ between said first and second images; and

(g) responsive to said number of image segments having different reflectance ratios exceeding a predetermined threshold, repeating steps (a)-(f) for at least one subsequent image frame and captured image frame, and responsive to said repeated steps producing a number of image segments having different pixel ratios, identifying said differences as caused by an intrusion within said scene.