HUMAN PRESENCE DETECTION IN A HOME SURVEILLANCE SYSTEM

• US 20170109613A1
• Filed: 10/19/2015
• Published: 04/20/2017
• Est. Priority Date: 10/19/2015
• Status: Active Grant

• Current Assignee: Ademco Inc. (Resideo Technologies, Inc.)
• Original Assignee: Honeywell International Inc.
• Inventors: Kolavennu, Soumitri N., Venkatesha, Sharath
• Granted Patent: US 10,083,376 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 18/217 Validation; Performance evaluation; Active pattern learning techniques
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/24 Classification techniques
• CPC Class Codes: G06F 18/40 Software arrangements specially adapted for pattern recognition, e.g. user interfaces or toolboxes therefor
• CPC Class Codes: G06T 2207/20081 Training; Learning
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06T 7/60 Analysis of geometric attributes
• CPC Class Codes: G06T 7/73 using feature-based methods
• CPC Class Codes: G06V 10/776 Validation; Performance evaluation
• 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: G06V 40/10 Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
• 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: H04N 7/18 Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast

Abstract

A surveillance system includes an image sensing device for capturing images of a plurality of moving objects at a plurality of different locations in a field of view of the image sensing device. The system also includes a user interactive processor for performing a human-distance calibration for a given location and position of the image sensing device. The system further includes a processor for detecting a human presence among the plurality of moving objects based on the human-distance calibration.

Claims

1. A surveillance system comprising:

an image sensing device for capturing images of a plurality of moving objects at a plurality of different locations in a field of view of the image sensing device;

a user interactive processor for performing a human-distance calibration for a given location and position of the image sensing device; and

a processor for detecting a human presence among the plurality of moving objects based on the human-distance calibration.

2. The surveillance system of claim 1, wherein the image sensing device is installable by an end user;

and wherein extrinsic and intrinsic parameters of the image sensing device and distance of the moving objects from the image sensing device are unknown.

3. The surveillance system of claim 1, wherein the system receives an initial configuration calibration via the user interactive processor;

and wherein the initial configuration calibration comprises a calibration of the surveillance system.

4. The surveillance system of claim 3, wherein the user interactive processor receives a human voice and the initial configuration calibration uses voice commands for generating a query and providing feedback.

5. The surveillance system of claim 3, wherein an output of the user interactive processor comprises a look up table (LUT) for detection of a human in a field of view of the image sensing device at different distances from the image sensing device, at different positions in the field of view, and under different lighting conditions.

6. The surveillance system of claim 5, wherein the initial configuration calibration comprises:

capturing a plurality of images of a person as image blobs at the plurality of different locations in the field of view of the image sensing device;

determining image blob features including one or more of a size, a position relative to the field of view, a shape, a centroid of the blob, and a position of a top of the blob at each of the plurality of different locations; and

constructing the LUT such that the LUT comprises multiple features extracted from the image blobs and associated locations of the image blobs in the field of view.

7. The surveillance system of claim 6, wherein the system:

detects a motion blob in the field of view of the image sensing device;

determines a size of the motion blob;

determines a position of the motion blob in the field of view of the image sensing device relative to one or more of the plurality of different locations in the field of view;

uses the LUT to compare the size of the motion blob at the relative position of the motion blob in the field of view to the sizes of the image blobs of the person at one or more locations that are in proximity to the location of the motion blob; and

determines whether the motion blob is a person based on the comparison.

8. The surveillance system of claim 6, wherein the estimating multiple features of the image blobs comprises:

determining a minimum vertical position of a bottom of the image blob, a maximum vertical position of a top of the image blob, a shape of the image blob, and a center position of the image blob; and

generating a histogram of oriented gradients calculated at a plurality of orientations for grayscale image pixels corresponding to the motion blob.

9. The surveillance system of claim 8, wherein the system

detects a motion blob in the field of view of the image sensing device;

determines characteristics of the motion blob;

uses the LUT to compare the size and characteristics of the motion blob at the relative position of the motion blob in the field of view to the sizes and characteristics of the image blobs of the person at one or more locations that are in proximity to the relative location of the motion blob; and

determines whether the motion blob is a person based on the comparison.

10. The surveillance system of claim 5, wherein the initial configuration calibration comprises:

establishing a ground plane in the field of view;

associating each of a plurality of points on the ground plane with one of the plurality of different locations in the field of view;

capturing image blobs of a person at each of the plurality of points on the ground plane;

determining a size of the image blobs at each of the plurality of points on the ground plane; and

storing in the LUT the size of the image blobs at each of the plurality of points on the ground plane.

11. The surveillance system of claim 10, wherein

the system detects a motion blob in the field of view of the image sensing device;

determines characteristics of the motion blob;

uses the LUT to compare the size and characteristics of the motion blob at the relative position of the motion blob in the field of view to the sizes and characteristics of the image blobs of the person at the points on the ground plan that are in proximity to the relative location of the motion blob; and

determines whether the motion blob is a person based on the comparison.

12. The surveillance system of claim 6, wherein the initial configuration comprises a pre-computed LUT based on an approximate size of an area under surveillance;

and wherein the system updates the LUT based on a self-learning algorithm, which uses features computed on the motion blobs in the field of view.

13. The surveillance system of claim 1, comprising a processor for accounting for, during the human-distance calibration, different sizes of persons including adults, adolescents, and children.

14. The surveillance system of claim 5, comprising self-learning and updating the LUT based on the plurality of image blobs of the person in a plurality of poses including sitting, squatting, bending down, and stretching upwards.

15. The surveillance system of claim 3, wherein the initial configuration calibration comprises:

establishing approximate boundary regions of a living space in the field of view based on extreme positions where the moving objects can be present in the field of view; and

establishing, by self-learning over a period of time or by involving user assistance, positions of doors, windows, stairs or other areas of interest in the field of view that can cause false motion alarms.

16. The surveillance system of claim 1, comprising a computer processor for removing an effect of changes in lighting.

17. The surveillance system of claim 1, comprising a computer processor for using temporal analysis to remove false alarms.

18. The surveillance system of claim 3, comprising a computer processor for receiving, during the initial configuration calibration, an indication of a window, a door, and a region of partial occlusion in the field of view of the image sensing device;

and for estimating correction parameters when a person is located in proximity to the window, the door, or the region of partial occlusion.

19. The surveillance system of claim 1, comprising a server coupled to the system via a network, the server operable for notifying a user of the human presence via an application on a mobile device.

20. A surveillance system comprising:

an image sensing device for capturing images of a plurality of moving objects at a plurality of different locations in a field of view of the image sensing device;

a user interactive processor for performing a human-distance calibration for a given location and position of the image sensing device; and

a processor for detecting a human presence among the plurality of moving objects based on the human-distance calibration;

wherein the system receives an initial configuration calibration via the user interactive processor; and

wherein the initial configuration calibration comprises a calibration of the surveillance system;

.wherein the initial configuration calibration comprises;

establishing approximate boundary regions of a living space in the field of view based on extreme positions where the moving objects can be present in the field of view; and

establishing, by self-learning over a period of time or by involving user assistance, positions of doors, windows, stairs or other areas of interest in the field of view that can cause false motion alarms.