Method for object recognition

• US 8,639,448 B2
• Filed: 05/18/2010
• Issued: 01/28/2014
• Est. Priority Date: 05/18/2009
• Status: Active Grant

• Current Assignee: Airbus Defence and Space GmbH (Airbus SE)
• Original Assignee: EADS Deutschland GmbH (Airbus SE)
• Inventors: Hiebl, Manfred
• Primary Examiner(s): Bui, Bryan
• Application Number:
  US12/782,414
• Publication Number:
  US 20100318300A1
• Time in Patent Office: 1,351 Days
• Field of Search: 702/19
• US Class Current: 702/19
• CPC Class Codes: G01S 13/726 Multiple target tracking
• CPC Class Codes: G01S 7/412 based on a comparison between measured values and known or stored values
• CPC Class Codes: G06F 18/2111 by using evolutionary computational techniques, e.g. genetic algorithms
• CPC Class Codes: G06F 18/24765 Rule-based classification

Abstract

A method for recognizing an object that has a plurality of expressions of abstract object characteristics, and is associated with an object characteristic class of a hierarchical system of object characteristic classes stored in a first memory. The method includes i) observing at least one location at which the object is presumed to be present, using a plurality of sensors in a sensor population, each of said sensors responding to at least one object characteristic and accordingly emitting a sensor signal; ii) checking whether each of the emitted sensor signals exceeds a specified threshold value for the sensor signals, and accepting sensor signals which exceed the threshold value; iii) pairing combinations of the sensor characteristics, for the accepted sensor signals obtained in ii) to form identification characteristic pairs; iv) comparing the population of identification characteristic pairs obtained in iii) to the object characteristic classes stored in the first memory; and v) identifying the object, based on the object characteristic class, whose object characteristic pairs are identical to the identification characteristic pairs obtained in iii).

Claims

1. A method for recognizing an object that has a plurality of expressions of abstract object characteristics, and is associated with an object characteristic class of a hierarchical system of object characteristic classes stored in a first memory;

said method comprising;

1.1) observing at least one location at which the object is presumed to be present, using a plurality of sensors in a sensor suite, each of said sensors responding to at least one object characteristic, and accordingly emitting a sensor signal;

1.2) checking whether each of the emitted sensor signals exceeds a specified threshold value for the sensor signals, accepting sensor signals which exceed the threshold value, and storing the accepted sensor signals in a second memory;

1.3) pairing combinations of the sensor characteristics for the accepted sensor signals obtained in 1.2) to form identification characteristic pairs;

1.4) comparing the identification characteristic pairs obtained in 1.3) to the object characteristic classes stored in the first memory; and

1.5) identifying the object, based on the object characteristic class, whose object characteristic pairs are identical to the identification characteristic pairs obtained in 1.3).

2. The method according to claim 1, wherein:

1.3) is carried out according to the rules of Mendelian genetics, whereby the number of identification characteristic pairs in each collection corresponds to a value of 2ⁿ, where n satisfies the condition nμ

1; and

to achieve this condition, doublets of identification characteristic pairs determined multiple times are initially omitted, and if the condition still cannot be satisfied, identification characteristic pairs which are already present are then multiplied until the condition is satisfied.

3. The method according to claim 1, wherein the at least one location in 1.1) is observed by scanning a region in which the object is presumed to be present.

4. The method according to claim 3, wherein the following are carried out after 1.2):

1.2.1) detecting location coordinates for each accepted sensor signal;

1.2.2) associating the location coordinates with the accepted sensor signal; and

1.2.3) storing the associated location coordinates together with the accepted sensor signals in the second memory.

5. The method according to claim 1, wherein:

a time signal representing a point in time at which the sensor signal was detected is associated with each accepted sensor signal; and

the respective associated time signals are stored together with the accepted sensor signals in the second memory.

6. The method according to claim 1, wherein:

in addition to a sensor that is already present, the plurality of sensors includes at least one additional sensor, which responds to at least one object characteristic; and

said additional sensor emits a sensor signal which is also accepted for further processing only when it exceeds a specified threshold value.

7. The method according to claim 1, wherein:

the observation of the at least one location in 1.1) by different sensors is carried out by at least a portion of the sensors at different points in time; and

1.3) and 1.4) are carried out when all sensors have observed the location.

8. The method according to claim 1, wherein:

observation of the location in 1.1) by different sensors is carried out multiple times in succession by at least a portion of the sensors;

1.3) is run through for the population of observations carried out;

identification characteristic pairs determined from the population of observations carried out are combined according to the rules of Mendelian genetics to form a collection of identification characteristic pairs;

the number of the identification characteristic pairs in each collection corresponds to a value of 2ⁿ, where n satisfies the condition nμ

1; and

to achieve said condition, in combining the determined identification characteristic pairs, multiply determined identification characteristic pairs are initially accepted into the collection only once, and if said condition still cannot be satisfied, identification characteristic pairs which are already present are multiplied until this condition is satisfied; and

in 1.4) the identification characteristic pairs combined in the collection are compared to the stored object characteristic pairs.

9. The method according to claim 1, wherein when the sensor characteristics are combined in 1.3), doublets of multiply present identification characteristic pairs are not considered, and the sequence of identification characteristics in an identification characteristic pair remains unconsidered.

10. The method according to claim 1, wherein the hierarchical object characteristic classes are formed by classifying the object characteristic pairs formed from the object characteristics.

11. The method according to claim 10, wherein:

the number of object characteristic pairs in an object characteristic class corresponds to a value of 2ⁿ, where n satisfies the condition nμ

1;

to achieve said condition, in the classification of the object characteristic pairs, doublets of multiply present identification characteristic pairs are initially not considered, and for this purpose the sequence of identification characteristics in an identification characteristic pair remains unconsidered; and

if the condition still cannot be satisfied, identification characteristic pairs which are already present are multiplied until the condition is satisfied.

12. A system for recognizing an object that has a plurality of expressions of abstract object characteristics, and is associated with an object characteristic class of a hierarchical system of object characteristics classes;

said system comprising;

a first memory having said hierarchical system of object classes stored therein;

a plurality of sensors, each of which responds to at least one object characteristic and emits a corresponding sensor signal;

a data processor; and

a non-transitory computer readable medium encoded with a series of commands which cause the system to perform following steps;

1.1) said plurality of sensors observing at least one location at which the object is presumed to be present;

1.2) checking whether each of the emitted sensor signals exceeds a specified threshold value for the sensor signals, accepting sensor signals which exceed the threshold value, and storing the accepted sensor signals in a second memory;

1.3) pairing combinations of the sensor characteristics for the accepted sensor signals obtained in 1.2) to form identification characteristic pairs;

1.4) comparing the identification characteristic pairs obtained in 1.3) to the object characteristic classes stored in the first memory; and

1.5) identifying the object, based on the object characteristic class, whose object characteristic pairs are identical to the identification characteristic pairs obtained in 1.3).

13. The system according to claim 12, wherein:

1.3) is carried out according to the rules of Mendelian genetics, whereby the number of identification characteristic pairs in each collection corresponds to a value of 2ⁿ, where n satisfies the condition nμ

1; and

to achieve said condition, doublets of identification characteristic pairs determined multiple times are initially omitted, and if the condition still cannot be satisfied, identification characteristic pairs which are already present are then multiplied until the condition is satisfied.

14. The system according to claim 13, wherein the following are carried out after 1.2):

1.2.1) detecting location coordinates for each accepted sensor signal;

1.2.2) associating the location coordinates with the accepted sensor signal; and

1.2.3) storing the associated location coordinates together with the accepted sensor signals in the second memory.

15. The system according to claim 14, wherein:

a time signal representing a point in time at which the sensor signal was detected is associated with each accepted sensor signal; and

the respective associated time signals are stored together with the accepted sensor signals in the second memory.

16. The system according to claim 12, wherein:

observation of the location in 1.1) by different sensors is carried out multiple times in succession by at least a portion of the sensors;

1.3) is run through for the population of observations carried out;

identification characteristic pairs determined from the population of observations carried out are combined according to the rules of Mendelian genetics to form a collection of identification characteristic pairs;

the number of the identification characteristic pairs in each collection corresponding to a value of 2ⁿ where n satisfies the condition nμ

1; and

to achieve said condition, in combining the determined identification characteristic pairs, multiply determined identification characteristic pairs are initially accepted into the collection only once, and if said condition still cannot be satisfied, identification characteristic pairs which are already present are multiplied until this condition is satisfied; and

in 1.4) the identification characteristic pairs combined in the collection are compared to the stored object characteristic pairs.

17. A non-transitory computer readable medium encoded in a non-transient manner with a program for causing an object recognition system to perform the following steps:

1.1) observing at least one location at which the object is presumed to be present, using a plurality of sensors in a sensor suite, each of said sensors responding to at least one object characteristic, and accordingly emitting a sensor signal;

1.2) checking whether each of the emitted sensor signals exceeds a specified threshold value for the sensor signals, accepting sensor signals which exceed the threshold value, and storing the accepted sensor signals in a second memory;

1.3) pairing combinations of the sensor characteristics for the accepted sensor signals obtained in 1.2) to form identification characteristic pairs;

1.4) comparing the identification characteristic pairs obtained in 1.3) to the object characteristic classes stored in the first memory; and

1.5) identifying the object, based on the object characteristic class, whose object characteristic pairs are identical to the identification characteristic pairs obtained in 1.3).

18. The computer readable medium according to claim 17, wherein:

1.3) is carried out according to the rules of Mendelian genetics, whereby the number of identification characteristic pairs in each collection corresponds to a value of 2ⁿ, where n satisfies the condition nμ

1; and

to achieve this condition, doublets of identification characteristic pairs determined multiple times are initially omitted, and if the condition still cannot be satisfied, identification characteristic pairs which are already present are then multiplied until the condition is satisfied.

19. The computer readable medium according to claim 18, wherein:

1.2.1) detecting location coordinates for each accepted sensor signal;

1.2.2) associating the location coordinates with the accepted sensor signal; and

1.2.3) storing the associated location coordinates together with the accepted sensor signals in the second memory.

20. The computer readable medium according to claim 17, wherein:

observation of the location in 1.1) by different sensors is carried out multiple times in succession by at least a portion of the sensors;

1.3) is run through for the population of observations carried out;

identification characteristic pairs determined from the population of observations carried out are combined according to the rules of Mendelian genetics to form a collection of identification characteristic pairs;

the number of the identification characteristic pairs in each collection corresponds to a value of 2ⁿ, where n satisfies the condition nμ

1; and

to achieve said condition, in combining the determined identification characteristic pairs, multiply determined identification characteristic pairs are initially accepted into the collection only once, and if said condition still cannot be satisfied, identification characteristic pairs which are already present are multiplied until this condition is satisfied; and

in 1.4) the identification characteristic pairs combined in the collection are compared to the stored object characteristic pairs.