Method and apparatus for automatically detecting intrusion object into view of image pickup device

• US 6,754,367 B1
• Filed: 09/28/2000
• Issued: 06/22/2004
• Est. Priority Date: 09/30/1999
• Status: Expired due to Fees

• Current Assignee: Hitachi Denshi Kabushiki Kaisha
• Original Assignee: Hitachi Denshi Kabushiki Kaisha
• Inventors: Ito, Wataru, Ueda, Hirotada
• Primary Examiner(s): Johnson, Timothy M.
• Assistant Examiner(s): Kassa, Yosef
• Application Number:
  US09/671,178
• Time in Patent Office: 1,363 Days
• Field of Search: 382/103, 382/104, 382/171, 382/190, 382/209, 382/219, 382/278, 382/282, 382/170, 348/152-155, 348/352, 348/699, 348/700
• US Class Current: 382/103
• CPC Class Codes: G06T 7/277 involving stochastic approaches, e.g. using Kalman filters
• CPC Class Codes: G06V 10/28 Quantising the image, e.g. histogram thresholding for discrimination between background and foreground patterns
• 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 13/19606 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

An intruding object detecting method and an intruding object detecting apparatus for judging whether an detected object is an intruding object or not even in the condition that there are shadows of moving objects other than target objects to be detected, such as shaking trees, leaves thereof, etc. An image signal is supplied from an image pickup device; a moving object is extracted from the image signal by a subtraction method; division is made between the image signal and a reference background image signal for every pixel so as to produce a ratio image; a histogram of luminance values (pixel values) is calculated on a part of the ratio image corresponding to the moving object extracted in the extracting step; a pixel value on the part of the ratio image corresponding to the extracted moving object is calculated in ratio representation from the calculated histogram; the calculated pixel value in ratio representation is estimated; and the extracted moving object is judged to be an intruding object or not, on the basis of the estimation of the pixel value in ratio representation. Thus, intruding objects are detected in the field of view to be monitored.

Claims

1. An intruding object detecting method comprising the steps of:

inputting an image signal from an image pickup apparatus;

extracting a moving object from said image signal and a reference background image signal stored in a memory device by a subtraction method;

making division between said image signal and said reference background image signal for every pixel so as to produce a ratio image;

calculating a histogram of pixel values on said ratio image of said moving object extracted from said image signal;

detecting symmetry of said histogram of the extracted moving object; and

judging whether said extracted moving object is an intruding object or not, on the basis of detection result of said symmetry of said histogram.

2. An intruding object detecting method according to claim 1, wherein said step of detecting symmetry of said histogram including the steps of:

calculating mean pixel values in ratio representation on a part of said ratio image corresponding to said extracted image object from said histogram (109); and

comparing an absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation with a predetermined threshold value (T1);

wherein said judging step judges said extracted image object to be an intruding object when comparison result in said comparing step concludes that said absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation is equal to or larger than said predetermined threshold value (T1), and judges said extracted image object to be not an intruding object when said absolute value is smaller than said predetermined threshold value (T1).

3. An intruding object detecting method according to claim 1, wherein said step of detecting symmetry of said histogram including the steps of:

calculating sharpnesses of pixel value distribution on a part of said ratio image corresponding to said extracted image object from said histogram; and

comparing each of said calculated sharpnesses of pixel value distribution with a predetermined threshold value (T2);

wherein said judging step judges said extracted image object to be an intruding object or not on the basis of comparison result in said comparing step.

4. An intruding object detecting method according to claim 3, wherein said judging step judges said extracted image object to be an intruding object when at least one of said sharpnesses of pixel value distribution is not smaller than said predetermined threshold value (T2).

5. An intruding object detecting method according to claim 3, wherein said judging step judges said extracted image object to be not an intruding object when both of said sharpnesses of pixel value distribution are smaller than said predetermined threshold value (T2).

6. An intruding object detecting method according to claim 3, wherein said predetermined threshold value (T2) is variable.

7. An intruding object detecting method according to claim 1, wherein said step of detecting symmetry of said histogram including:

a first comparing step of calculating mean pixel values in ratio representation on a part of said ratio image corresponding to said extracted image object from said histogram, and comparing an absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation with a first predetermined threshold value (T1); and

a second comparing step of calculating sharpnesses of pixel value distribution on a part of said ratio image corresponding to said extracted image object from said histogram, and comparing each of said calculated sharpnesses of pixel value distribution with a second predetermined threshold value (T2);

wherein said first comparing step executes said second comparing step when said difference between one mean pixel value in ratio representation and a reciprocal of the other mean pixel value in ratio representation is smaller than said first predetermined threshold value (T1).

8. An intruding object detecting method according to claim 7, wherein said second predetermined threshold value (T2) is variable.

9. An intruding object detecting method according to claim 7, wherein said second comparing step judges said extracted image object to be an intruding object when at least one of said sharpnesses of pixel value distribution is not smaller than said predetermined threshold value (T2).

10. An intruding object detecting method according to claim 7, wherein said second comparing step judges said extracted image object to be not an intruding object when both of said sharpnesses of pixel value distribution are smaller than said predetermined threshold value (T2).

11. An intruding object detecting method according to claim 1, wherein in said step of producing a ratio image, said ratio image is produced with respect to a part of said image signal corresponding to said extracted image object.

12. An intruding object detecting method according to claim 1, further comprising a step of storing judgement result of said step of judging whether said extracted image object is an intruding object or not, and a step of comparing a plurality of past judgement results.

13. An intruding object detecting method according to claim 1, further comprising a step of judging whether a mean pixel value of said part of said image signal corresponding to said extracted moving object is smaller than a predetermined value (T3) or not, and said histogram calculating step is executed about said extracted image object which said mean pixel value is not smaller than said predetermined value (T3).

14. An intruding object detecting method according to claim 1, wherein said step of detecting symmetry of said histogram including the step of:

detecting, from said histogram, symmetry between a first area where a luminance value of said input image is larger than a luminance value of said reference background image and a second area where a luminance value of said input image is larger than a luminance value of said reference background image, with respect to a part of said ratio image corresponding to said extracted moving object.

15. An intruding object detecting method according to claim 14, wherein said step of detecting symmetry of said histogram including the step of:

calculating mean pixel values in ratio representation in said first and second areas on a part of said ratio image corresponding to said extracted moving object from said histogram; and

comparing an absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation with a predetermined threshold value (T1);

wherein said judging step judges said extracted moving object to be an intruding object when said comparing step shows comparison result that said absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation is equal to or larger than said predetermined threshold value (T1), and judges said extracted moving object to be not an intruding object when said absolute value is smaller than said predetermined threshold value (T1).

16. An intruding object detecting method according to claim 14, wherein said step of detecting symmetry of said histogram including the step of:

calculating sharpnesses of pixel value distribution in said first and second areas on a part of said ratio image corresponding to said extracted moving object from said histogram; and

comparing each of said pair of calculated sharpnesses of pixel value distribution with a predetermined threshold value (T2);

wherein said judging step judges whether said extracted moving object is an intruding object or not on the basis of comparison result in said comparing step.

17. An intruding object detecting method according to claim 14, wherein said step of detecting symmetry of said histogram including:

a first comparing step of calculating mean pixel values in ratio representation in said first and second areas on a part of said ratio image corresponding to said extracted moving object from said histogram, and comparing an absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation with a first predetermined threshold value (T1);

a second comparing step of calculating sharpnesses of pixel value distribution in said first and second areas on a part of said ratio image corresponding to said extracted moving object from said histogram, and comparing each of said two calculated sharpnesses of pixel value distribution with a second predetermined threshold value (T2);

wherein said first comparing step executes said second comparing step when said absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation is smaller than said first predetermined threshold value (T1).

18. An intruding object detecting method comprising the steps of:

inputting an image signal from an image pickup apparatus;

extracting a moving object from said image signal and a reference background image signal stored in a memory device by a subtraction method;

making division between said image signal and said reference background image signal for every pixel so as to produce a ratio image;

calculating a histogram of pixel values on said ratio image of said moving object extracted from said image signal;

calculating sharpnesses of pixel value distribution on a part of said ratio image corresponding to said extracted moving object from said calculated histogram;

comparing each of said calculated sharpnesses with a predetermined threshold value (T2); and

judging whether said extracted moving object is an intruding object or not on the basis of detection result of said sharpnesses.

19. An intruding object detecting method according to claim 18, wherein said judging step judges said extracted moving object to be an intruding object when at least one of said sharpnesses of pixel value distribution is not smaller than said predetermined threshold value (T2).

20. An intruding object detecting method according to claim 18, wherein said judging step judges said extracted moving object to be not an intruding object when both of said sharpnesses of pixel value distribution are smaller than said predetermined threshold value (T2).

21. An intruding object detecting method according to claim 18, wherein in said step of producing a ratio image, said ratio image is produced with respect to a part of said image signal corresponding to said extracted moving object.

22. An intruding object detecting method according to claim 18, further comprising a step of judging whether a mean pixel value of said part of said image signal corresponding to said extracted moving object is smaller than a predetermined value (T3) or not, and said histogram calculating step is executed about said extracted image object which said mean pixel value is not smaller than said predetermined value (T3).

23. An intruding object detecting method according to claim 18, further comprising a step of storing judgement result of said judging step, and a step of comparing a plurality of past judgement results, and stored;

wherein said extracted moving object is judged to be an intruding object or not on the basis of comparison result of said past judgement results.

24. An intruding object detecting apparatus comprising:

an image pickup device for generating an image signal to detect an intruding object in a field to be monitored; and

an image processor for receiving said image signal;

wherein said image processor comprising;

a moving object extracting unit for extracting a moving object from said image signal and a reference background image signal stored in a memory device by a subtraction method;

a ratio image producing unit for generating a ratio image by making division between said image signal and said reference background image signal;

a histogram producing unit for producing a histogram of pixel values on said ratio image of said extracted moving object;

a detecting unit for detecting symmetry of said histogram of the extracted moving object; and

a decision unit for deciding whether said extracted moving object is an intruding object or not on the basis of detection result of said symmetry of said histogram.

25. A computer program product comprising:

a computer usable medium having computer readable program code means embodied in said medium for detecting an intruding object appearing in an image signal supplied from an image pickup device, said computer readable program code means comprising;

means for applying a subtraction method to said image signal and a reference background image signal stored in a memory device to extract an moving object;

means for making division between said image signal and said reference background image signal so as to produce a ratio image;

means for producing a histogram of pixel values on said ratio image of said extracted moving object from said ratio image;

means for calculating at least one of mean pixel values in ratio representation and sharpnesses of pixel value distribution about said extracted moving object from said histogram;

means for comparing said calculated mean pixel values in ratio representation and/or said calculated sharpnesses with a predetermined threshold value; and

means for judging whether said extracted moving object is an intruding object or not, on the basis of comparison result obtained by said comparing means.

26. An intruding object detecting method comprising the steps of:

inputting an image signal from an image pickup device;

extracting a moving object from said image signal and a reference background image signal stored in a memory device by a subtraction method;

making division between said image signal and said reference background image signal for every pixel so as to produce a ratio image;

calculating a histogram of pixel values on a part of said ratio image corresponding to said moving object extracted from said image signal;

calculating mean pixel values in ratio representation on a part of said ratio image corresponding to said extracted moving object from said calculated histogram;

comparing said mean pixel values in ratio representation; and

judging whether said extracted moving object is an intruding object on the basis of comparison result of said mean pixel values in ratio representation.

27. An intruding object detecting method according to claim 26, wherein said comparing step including the steps of:

calculating mean pixel values in ratio representation on a part of said ratio image corresponding to said extracted moving object from said histogram; and

comparing an absolute value of a difference between one of said mean pixel value in ratio representation and a reciprocal of the other mean pixel value in ratio representation with a predetermined threshold value (T1);

wherein said judging step judges that said extracted object is an intruding object when said comparing step shows comparison result that said absolute value of a difference between one of said mean pixel values in ratio representation and a reciprocal of the other mean pixel value in ratio representation is equal to or larger than said predetermined threshold value (T1), and judges that said extracted object is not an intruding object when said comparing step indicates that said absolute value is smaller than said predetermined threshold value (T1).

28. An intruding object detecting method according to claim 26, wherein in said step of producing a ratio image, said ratio image is produced with respect to a part of said image signal corresponding to said extracted moving object.

29. An intruding object detecting method according to claim 26, further comprising a step of judging whether a mean pixel value of a part of said image signal corresponding to said extracted moving object is smaller than a predetermined value (T3) or not, and said histogram calculating step is executed on said moving object in which said mean pixel value shows a value not smaller than said predetermined value.

30. An intruding object detecting method according to claim 26, further comprising a step of storing judgement result in said judging step, and a step of comparing a plurality of past judgement results, wherein said extracted moving object is judged to be an intruding object or not on the basis of comparison result of said past judgement results.