Apparatus for identifying individual animals and image processing method

• US 6,285,780 B1
• Filed: 03/27/1998
• Issued: 09/04/2001
• Est. Priority Date: 03/28/1997
• Status: Expired due to Fees

• Current Assignee: OKI Electric Industry Company Limited
• Original Assignee: OKI Electric Industry Company Limited
• Inventors: Yamakita, Osamu, Kuno, Yuji
• Primary Examiner(s): MEHTA, BHAVESH M
• Application Number:
  US09/049,073
• Time in Patent Office: 1,257 Days
• Field of Search: 382/110, 382/117, 382/115, 382/100, 351/205, 351/206, 351/211
• US Class Current: 382/110
• CPC Class Codes: A01K 11/006 Automatic identification systems for animals, e.g. electronic devices, transponders for animals identification systems for access to feeding pens A01K1/0209; ear tags with electronic identification means A01K11/004
• CPC Class Codes: A01K 29/00 Other apparatus for animal husbandry

Abstract

An individual animal identification apparatus has an eye region image pick-up unit for obtaining an input image by taking a picture of an eye, including an iridial granule, of an animal to be registered and identified, a geometric normalizer for performing a coordinate transformation process by using a specified parameter to move the iridial granule region of the input image onto predetermined coordinates, a feature extractor for extracting the feature of the normalized iridial granule, a dictionary memory for registering data on the features of the iridial granule regions corresponding to animals to be registered, the data having been extracted by the feature extractor, and a collator for checking whether or not an animal to be identified has been registered by comparing data on the feature of the iridial granule region corresponding to the animal under identification, the feature having been extracted by the feature extractor, with the feature data registered in the dictionary memory.

Claims

1. An individual animal identification apparatus comprising:

an eye region image pick-up unit for obtaining an input image by taking a picture of an eye, including an iridial granule in an iridial granule region, of an animal to be registered and identified;

a geometric normalizer for performing a coordinate transformation process to normalize said iridial granule by using a specified parameter to move said iridial granule region of the input image onto predetermined coordinates;

a feature extractor for extracting a feature of said normalized iridial granule;

a dictionary memory for registering feature data on features of iridial granule regions corresponding to animals to be registered, said data having been extracted by said feature extractor; and

a collator for checking whether or not an animal to be identified has been registered by comparing data on said feature of said iridial granule region corresponding to the animal under identification, said feature having been extracted by said feature extractor, with said feature data registered in said dictionary memory.

2. An individual animal identification apparatus according to claim 1, wherein said geometric normalizer decides a center-of-gravity position of said iridial granule region as said parameter and performs said coordinate transformation process using said parameter.

3. An individual animal identification apparatus according to claim 2, wherein said geometric normalizer performs the coordinate transformation process by converting said center-of-gravity position to a predetermined coordinate position.

4. An individual animal identification apparatus according to claim 1, wherein said geometric normalizer decides an inclination of a principal component of said iridial granule region as said parameter and performs said coordinate transformation process by using said parameter.

5. An individual animal identification apparatus according to claim 4, wherein said geometric normalizer performs the coordinate transformation process by converting said inclination to a predetermined value.

6. An individual animal identification apparatus according to claim 1, wherein said geometric normalizer decides as a longitudinal dimension of said iridial granule region as said parameter and performs said coordinate transformation process by using said parameter.

7. An individual animal identification apparatus according to claim 6, wherein said geometric normalizer performs the coordinate transformation process by converting said longitudinal dimension to a predetermined value.

8. An individual animal identification apparatus according to claim 1, further comprising a density normalizer for calculating a mean value of the density of said iridial granule region in connection with the contrast, and normalizing the density of said iridial granule region on the basis of said mean value.

9. An individual animal identification apparatus according to claim 8, wherein said density normalizer calculates a standard deviation of the density of said iridial granule region by regarding said standard deviation as the contrast, and normalizes the density by using said standard deviation and said mean value.

10. An individual animal identification apparatus according to claim 1, wherein said feature extractor divides said iridial granule region into a plurality of subdivisions, converts said subdivisions into compressed regions each having a mean value of pixel densities, and uses density values of said compressed regions as data on said feature.

11. An individual animal identification apparatus according to claim 10, said collator performs an arithmetic operation on differences between the density values of the compressed regions extracted by said feature extractor and the density values of the compressed regions registered in said dictionary memory, and uses a obtained difference as a difference for identification of said animal.

12. An individual animal identification apparatus according to claim 11, wherein said collator calculates a mean density difference of said compressed regions by dividing the total sum of said density values by the number of said compressed regions, and uses a mean density difference as said difference.

13. An individual animal identification apparatus according to claim 1, further comprising at least one additional geometric normalizer, wherein said geometric normalizers respectively perform a coordinate transformation process on said iridial granule region by using a parameter which is different from the parameters used by the other normalizers.

14. An individual animal identification apparatus comprising:

an eye region image pick-up unit for capturing images of an eye, including an iridial granule in an iridial granule region, of an animal at a plurality of camera angles to obtain input images, said animal to be registered in a registration mode of the apparatus and to be identified in an identification mode of the apparatus;

a camera angle determining section for deriving data on said camera angles;

a feature extracting section for extracting a feature of said iridial granule region from said input images to derive corresponding feature data;

a dictionary memory for, in said registration mode, registering said feature data of the animal to be registered along with the corresponding camera angle data; and

a collating section for, in said identification mode, collating the feature data of the animal to be identified, derived at said feature extracting section, with the feature data registered in said dictionary memory during said registration mode and having the camera angle data corresponding to the camera angle data of the feature data of the animal to be identified, so as to determine whether the animal to be identified is registered in said dictionary memory.

15. An individual animal identification apparatus comprising:

an eye region image pick-up unit for capturing images of an eye, including an iridial granule in an iridial granule region, of an animal to obtain input images, said animal to be registered in a registration mode of the apparatus and to be identified in an identification mode of the apparatus;

a feature extracting section for extracting a feature of said iridial granule region from said input images to derive corresponding feature data;

a data classifying means for, in said registration mode, classifying a plurality of the feature data, received from said feature extracting section, of the animal to be registered into a plurality of groups based on image pick-up conditions, respectively;

a representative dictionary setting means for, in said registration mode, determining representative feature data in each of said groups and registering in a dictionary memory said representative feature data determined in each of said groups; and

a collating section for, in said identification mode, collating the feature data of the animal to be identified, derived at said feature extracting section, with each of said representative feature data registered in said dictionary memory during said registration mode, so as to determine whether the animal to be identified is registered in said dictionary memory.

16. An individual animal identification apparatus according to claim 15, wherein said data classifying means determines a plurality of temporary representative feature data among said plurality of the feature data received from said feature extracting section and classifies said plurality of the feature data into said groups each having one of said temporary representative feature data and those feature data having small difference degrees relative to said one of said temporary representative feature data.

17. An individual animal identification apparatus according to claim 15, wherein said representative dictionary setting means derives mean pixel values of the feature data in each of said groups and determines, as said representative feature data, the feature data having pixel values which are the closest to said mean pixel values.

18. An individual animal identification apparatus comprising:

an eye region image pick-up unit for capturing an image of an eye, including an iridial granule, in an iridial granule region of an animal to obtain an input image, said animal to be registered in a registration mode of the apparatus an to be identified in an identification mode of the apparatus;

a feature extracting section for extracting a feature of a region of said iridial granule region from said input image to derive corresponding feature data;

a region selecting means for, in said registration mode, selecting a stable identification region of said iridial granule based on a plurality of the feature data, received from said feature extracting section, of the animal to be registered;

a dictionary preparing means for, in said registration mode, producing regional feature data of said stable identification region and registering said regional feature data in a dictionary memory; and

a collating section for, in said identification mode, collating the feature data on an identification region of the iridial granule of the animal to be identified, derived at said feature extracting section, with the regional feature data registered in said dictionary memory during said registration mode, so as to determine whether the animal to be identified is registered in said dictionary memory.

19. An individual animal identification apparatus comprising:

an eye region image pick-up unit for capturing an image of an eye, including an iridial granule, of an animal to obtain an input image, said animal to be registered in a registration mode of the apparatus and to be identified in an identification mode of the apparatus;

a position data output means for outputting position data representing a position of said iridial granule relative to a particular position in a region of said eye in said input image, said position data output means including

a region extracting section for extracting a region of said iridial granule from said input image, deriving a position of the center of gravity of said iridial granule region as said position of said iridial granule and further deriving a position of the center of gravity of a pupil in said eye region as said particular position,

a shadow removing section for accumulating pixel values along a contour of said iridial granule region extracted at said region extracting section, and determining a contour represented by an accumulated darkest pixel value to be a new contour of said iridial granule region, and

a center of gravity coordinate calculating section for deriving a distance from the position of the center of gravity of said pupil to the position of the center of gravity of aid iridial granule region as said position data, said center of gravity coordinate calculating section correcting the position of the center of gravity of said iridial granule region corresponding to said new contour so as to derive said position data;

a memory section for, in said registration mode, registering said position data, received from said position data output means of the animal to be registered;

a distance calculating section for, in said identification mode, deriving a position difference between said position data, received form said position data output means, of the animal to be identified and each of a plurality of the position data registered in said memory section; and

a collating section for, in said identification mode, selecting the registered position data whose position difference is the smallest and comparing the smallest position difference with a predetermined value, so as to determine whether the animal to be identified is registered in said memory section.