Abnormality detection apparatus and method

• US 6,504,482 B1
• Filed: 01/04/2001
• Issued: 01/07/2003
• Est. Priority Date: 01/13/2000
• Status: Expired due to Fees

• Current Assignee: Sanyo Electric Company Limited (Panasonic Holdings Corporation)
• Original Assignee: Sanyo Electric Company Limited (Panasonic Holdings Corporation)
• Inventors: Oyamada, Kenji, Iinuma, Toshiya, Okada, Seiji, Mori, Yukio
• Primary Examiner(s): MULLEN, THOMAS J
• Application Number:
  US09/753,672
• Time in Patent Office: 733 Days
• Field of Search: 340/573.1, 340/573.6, 340/573.7, 340/541, 340/529
• US Class Current: 340/573.1
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G08B 21/043 detecting an emergency event, e.g. a fall
• CPC Class Codes: G08B 21/0476 Cameras to detect unsafe condition, e.g. video cameras

Abstract

An abnormality detection apparatus for detecting an event where a monitored object in a room has lapsed into an abnormal state. The apparatus includes an image pickup system for picking up an image of a scene in the room, and feature-quantity extraction devices for extracting an image feature quantity from the image picked up by the image pickup devices. The apparatus also includes judgment devices for determining whether the monitored object in the room has lapsed into the abnormal state or not based on the time-variations of the image feature quantity extracted by the feature-quantity extraction devices, and outputting an alarm that responds to the determination of the occurrence of the abnormal state.

Claims

1. An abnormality detection apparatus for detecting an event where a monitored object in a room has lapsed into an abnormal state, the apparatus comprising:

image pickup means for picking up an image of a scene in the room;

feature-quantity extraction means for extracting an image feature quantity from the image picked up by the image pickup means; and

judgment means for determining whether the monitored object in the room has lapsed into the abnormal state or not based on the time-variations of the image feature quantity extracted by the feature-quantity extraction means.

2. The abnormality detection apparatus as claimed in claim 1,

wherein the feature-quantity extraction means calculates the image feature quantity of each of plural feature-quantity calculation regions defined in one screen, the calculation performed on the feature-quantity calculation regions at predetermined time intervals, and

wherein the judgment means determines whether the monitored object in the room has lapsed into the abnormal state or not based on the time-variations of the respective image feature quantities of the feature-quantity calculation regions.

3. The abnormality detection apparatus as claimed in claim 2, wherein the judgment means comprises:

means for judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

means for determining whether the feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the image feature quantities is made; and

means for determining that the monitored object has lapsed into the abnormal state when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the image feature quantities has continued for a predetermined period of time.

4. The abnormality detection apparatus as claimed in claim 2, further comprising a first alarm device installed in the room and a second alarm device installed outside the room,

wherein the judgment means comprises:

means for judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

means for determining whether the feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the image feature quantities is made;

means for triggering the first alarm device when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the image feature quantities has continued for a first predetermined period of time; and

means for triggering the second alarm device when the state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the image feature quantities has continued for a second predetermined period of time which is longer than the first predetermined period of time.

5. The abnormality detection apparatus as claimed in claim 2, wherein the judgment means comprises:

first judging means for judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

second judging means for determining whether feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variation of the image feature quantities is made; and

third judging means for determining whether the monitored object has lapsed into the abnormal state or not based on a predetermined number of preceding judgment results given by the second judging means.

6. The abnormality detection apparatus as claimed in claim 2, further comprising a first alarm device installed in the room and a second alarm device installed outside the room,

wherein the judgment means comprises:

first judging means for judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

second judging means for determining whether the monitored object is in motion or not each time the judgment as to the time-variations of the image feature quantities is made, the judgment made based on whether the feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number;

means for triggering the first alarm device when a first predetermined number of preceding judgment results given by the second judging means include not more than a first predetermined number of determinations that the monitored object is in motion; and

means for triggering the second alarm device when a second predetermined number of preceding judgment results given by the second judging means include not more than a second predetermined number of determinations that the monitored object is in motion.

7. The abnormality detection apparatus as claimed in claim 1,

wherein the feature-quantity extraction means calculates an average of image feature quantities for a predetermined number of fields with respect to each of plural feature-quantity calculation regions defined in one screen, the calculation performed on the calculation regions at predetermined time intervals;

and

the judgment means determines whether the monitored object in the room has lapsed into the abnormal state or not based on the time-variations of the respective average image feature quantities of the feature-quantity calculation regions.

8. The abnormality detection apparatus as claimed in claim 7, wherein the judgment means comprises:

means for judging the respective feature-quantity calculation regions as to whether the average image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

means for determining whether the feature-quantity calculation regions determined to be time-varied in the average image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the average image feature quantities is made; and

means for determining that the monitored object has lapsed into the abnormal state when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the average image feature quantities has continued for a predetermined period of time.

9. The abnormality detection apparatus as claimed in claim 7, further comprising a first alarm device installed in the room and a second alarm device installed outside the room,

wherein the judgment means comprises:

means for judging the respective feature-quantity calculation regions as to whether the average image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

means for determining whether the feature-quantity calculation regions determined to be time-varied in the average image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the average image feature quantities is made;

means for triggering the first alarm device when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the average image feature quantities has continued for a first predetermined period of time; and

means for triggering the second alarm device when the state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the average image feature quantities has continued for a second predetermined period of time which is longer than the first predetermined period of time.

10. An abnormality detection method for detecting an event where a monitored object in a room has lapsed into an abnormal state, the method comprising the steps of:

a first step of picking up an image of a scene in the room via an image pickup device;

a second step of extracting an image feature quantity from the image captured at the first step; and

a third step of determining whether the monitored object in the room has lapsed into the abnormal state or not based on the time-variations of the image feature quantity extracted at the second step.

11. The abnormality detection method as claimed in claim 10,

wherein the second step calculates an image feature quantity of each of plural feature-quantity calculation regions defined in one screen, the calculation performed on the calculation regions at predetermined time intervals, and

wherein the third step determines whether the monitored object in the room has lapsed into the abnormal state or not based on the time-variations of the respective image feature quantities of the feature-quantity calculation regions.

12. The abnormality detection method as claimed in claim 11, wherein the third step comprises the steps of:

judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

determining whether the feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the image feature quantities is made; and

determining that the monitored object has lapsed into the abnormal state when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the image feature quantities has continued for a predetermined period of time.

13. The abnormality detection method as claimed in claim 11, wherein a first alarm device is installed in the room while a second alarm device is installed outside the room, and

wherein the third step comprises the steps of:

judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

determining whether the feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the image feature quantities is made;

triggering the first alarm device when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the image feature quantities has continued for a first predetermined period of time; and

triggering the second alarm device when the state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the image feature quantities has continued for a second predetermined period of time which is longer than the first predetermined period of time.

14. The abnormality detection method as claimed in claim 11, wherein the third step comprises:

Step ‘

a’

of judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

Step ‘

b’

of determining whether the feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the image feature quantities is made; and

Step ‘

c’

of determining whether the monitored object has lapsed into the abnormal state or not based on a predetermined number of preceding judgment results given by Step ‘

b’

.

15. The abnormality detection method as claimed in claim 11, wherein a first alarm device is installed in the room while a second alarm device is installed outside the room, and

wherein the third step comprises:

Step ‘

a’

of judging the respective feature-quantity calculation regions as to whether the image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

Step ‘

b’

of determining whether the feature-quantity calculation regions determined to be time-varied in the image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the image feature quantities is made, thereby determining whether the monitored object is in motion or not;

Step ‘

c’

of triggering the first alarm device when a first predetermined number of preceding judgment results given by Step ‘

b’

include not more than a first predetermined number of determinations that the monitored object is in motion; and

Step ‘

d’

of triggering the second alarm device when a second predetermined number of preceding judgment results given by Step ‘

b’

include not more than a second predetermined number of determinations that the monitored object is in motion.

16. The abnormality detection method as claimed in claim 10,

wherein the second step calculates an average of image feature quantities for a predetermined number of fields with respect to each of plural feature-quantity calculation regions defined in one screen, the calculation performed on the calculation regions at predetermined time intervals;

and

wherein the third step determines whether the monitored object in the room has lapsed into the abnormal state or not based on the time-variations of the respective average image feature quantities of the feature-quantity calculation regions.

17. The abnormality detection method as claimed in claim 16, wherein the third step comprises the steps of:

judging the respective feature-quantity calculation regions as to whether the average image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

determining whether the feature-quantity calculation regions determined to be time-varied in the average image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the average image feature quantities is made; and

determining that the monitored object has lapsed into the abnormal state when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the average image feature quantities has continued for a predetermined period of time.

18. The abnormality detection method as claimed in claim 16, wherein a first alarm device is installed in the room while a second alarm device is installed outside the room, and wherein the third step comprises the steps of:

judging the respective feature-quantity calculation regions as to whether the average image feature quantity is time-varied or not, the judgment made on the calculation regions at predetermined time intervals;

determining whether the feature-quantity calculation regions determined to be time-varied in the average image feature quantities are equal to or greater than a predetermined number or less than the predetermined number each time the judgment as to the time-variations of the average image feature quantities is made; and

triggering the first alarm device when a state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the average image feature quantities has continued for a first predetermined period of time; and

triggering the second alarm device when the state with less than the predetermined number of feature-quantity calculation regions determined to be time-varied in the average image feature quantities has continued for a second predetermined period of time which is longer than the first predetermined period of time.