Unusual event detection in wide-angle video (based on moving object trajectories)

• US 8,866,910 B1
• Filed: 10/09/2012
• Issued: 10/21/2014
• Est. Priority Date: 09/18/2008
• Status: Expired due to Fees

• Current Assignee: Grandeye Limited (Oncam Global Group AG)
• Original Assignee: Grandeye Limited (Oncam Global Group AG)
• Inventors: Noyan, Timucin, Davey, Mark K, Cetin, A Enis
• Primary Examiner(s): BOUTAH, ALINA A
• Application Number:
  US13/647,492
• Time in Patent Office: 742 Days
• Field of Search: 348/E7.086, 348/143, 348/155, 348/161, 348/169, 348/154, 348/E5.005, 382/103, 382/181, 382/107, 382/106, 382/201, 382/202, 382/291, 382/294, 382/190, 382/77, 382/94, 345/419, 715/723, 715/202, 715/203, 715/719, 715/726, 707/E17.028
• US Class Current: 348/169
• CPC Class Codes: G06F 18/214 Generating training patterns; Bootstrap methods, e.g. bagging or boosting
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/20064 Wavelet transform [DWT]
• CPC Class Codes: G06T 7/215 Motion-based segmentation
• CPC Class Codes: G06T 7/251 involving models
• CPC Class Codes: G06T 7/262 using transform domain methods, e.g. Fourier domain methods
• CPC Class Codes: G06V 20/44 Event detection
• CPC Class Codes: G06V 20/49 Segmenting video sequences, i.e. computational techniques such as parsing or cutting the sequence, low-level clustering or determining units such as shots or scenes
• 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 13/19628 of wide angled cameras and camera groups, e.g. omni-directional cameras, fish eye, single units having multiple cameras achieving a wide angle view
• CPC Class Codes: H04N 23/60 Control of cameras or camera modules
• CPC Class Codes: H04N 7/183 for receiving images from a single remote source

Abstract

Object images captured by a wide-angle camera are distorted due to the optical effects of the wide-angle lens. The disclosed innovations allow an automatic analysis on the corrected image distinguishing normal movement from an unusual event movement. The analysis is based on Markov Modeling on moving object trajectories and motion angles.

Claims

1. A system for monitoring moving objects in a wide-angle video, comprising:

a wide-angle lens on a digital video camera capturing a sequence of video images; and

an image data processor detecting and converting moving object trajectories into vectors in a perspectively corrected image domain and interpreting the object vectors as unusual behavior according to a stochastic model, which is not a hidden Markov Model, in which the states are defined according to motion vector angles of moving objects.

2. The system of claim 1, wherein the stochastic model comprises a state transition probability model.

3. The system of claim 1, wherein the state transition probability model comprises a Markov model.

4. The system of claim 1, wherein the stochastic model indicates an unusual event based on motion vector angles associated with moving pixel blobs generating motion vectors angles defined as unusual behavior.

5. A system for monitoring moving objects in a wide-angle video, comprising:

a wide-angle lens on a digital video camera capturing a sequence of video images; and

an image data processor detecting and converting moving object trajectories into vectors in a perspectively corrected image domain and interpreting the object vectors as unusual behavior according to a stochastic model, which is not a hidden Markov Model, in which the states are defined according to motion vector angles of moving objects;

wherein the stochastic model is trained with motion vectors in Cartesian coordinates, and polar coordinates are computed for the non-rectilinear image domain associated with unusual behavior.

6. A system for classifying a movement as an unusual event in a wide-angle video, comprising:

a digital PTZ camera capturing distorted images and converting the images into a video stream of non-rectilinear video images;

image data processing circuitry operating analysis software algorithms that segment moving object pixels from stationary pixels, determine corrected motion vectors for the moving object, analyze the vectors according to Markov models, and classify the movement as normal or unusual;

wherein the movement is classified according to the conditional probability if ones of unusual event Markov models exceed the conditional probability of ones of regular event Markov models.

7. The system of claim 6, wherein the moving object pixels are derived from image data by segmenting the non-rectilinear video image data into moving foreground pixels and stationary background pixels, wherein a background image is obtained by recursively averaging the past images of the digital video.

8. The system of claim 6, wherein the moving object pixels are derived from image data by segmenting the non-rectilinear video image data into moving foreground pixels and stationary background pixels, wherein a background image is obtained by median filtering.

9. The system of claim 6, wherein the Markov models are classified as unusual event or regular event based on motion vectors associated with unusual movement and regular movement.

10. The system of claim 6, wherein Markov models trained with motion vectors in Cartesian coordinates and polar coordinates defined as regular behavior indicate regular events.

11. The system of claim 6, wherein Markov models trained with motion vectors in Cartesian coordinates and polar coordinates defined as unusual behavior indicate unusual events.