Image recognition apparatus

1. An image recognition apparatus, comprising:

a storage unit, using a non-transitory computer-readable storage medium, which stores a plurality of pieces of identical recognition reference data which are used as references for recognizing a specific object from an image;

an image recognition unit, using a processor, which recognizes the specific object in images acquired by an image acquisition unit by switching between the plurality of pieces of identical recognition reference data stored in the storage unit, on a frame basis for consecutive frames of the images acquired by the image acquisition unit;

a recognition determining unit, using the processor, which, on condition that recognition results obtained by the image recognition unit by using the plurality of pieces of identical recognition reference data meet a predetermined matching condition, outputs the recognition results as effective recognition results;

an examining unit, using the processor, which examines a presence or an absence of a fault in a plurality of pieces of recognition reference data stored in the storage unit; and

a recovery unit, using the processor, which recovers the plurality of pieces of identical recognition reference data if the examining unit determines that there is a fault.

2. The image recognition apparatus according to claim 1 , wherein

the examining unit periodically examines the presence or the absence of a fault.

3. An image recognition apparatus, comprising:

a storage unit, using a non-transitory computer-readable storage medium, which stores a plurality of pieces of identical recognition reference data which are used as references for recognizing a specific object from an image;

an image recognition unit, using a processor, which recognizes the specific object in images acquired by an image acquisition unit by switching between the plurality of pieces of identical recognition reference data stored in the storage unit, on a frame basis for consecutive frames of the images acquired by the image acquisition unit;

a recognition determining unit, using the processor, which, on condition that recognition results obtained by the image recognition unit by using the plurality of pieces of identical recognition reference data meet a predetermined matching condition, outputs the recognition results as effective recognition results;

a fault determining unit, using the processor, which determines a fault in the recognition reference data of an odd number of recognition results based on majority rule logic, wherein recognition results comprising a minority are determined as being faulty, and

a recovery unit, using the processor, which recovers the recognition reference data determined as being faulty by the fault determining unit.

4. A method for recognizing a specific object from acquired images, the method comprising:

storing a plurality of pieces of identical recognition reference data within a storage unit, the plurality of pieces of identical recognition reference data including references for recognizing the specific object;

recognizing the specific object within the acquired images by switching between the plurality of pieces of identical recognition reference data within the storage unit on a frame basis for consecutive frames of the acquired images;

determining whether the plurality of pieces of identical recognition reference data meet a predetermined matching condition;

outputting the recognition results as effective recognition results based on whether the plurality of pieces of identical recognition reference data meet the predetermined matching condition;

examining for a presence or an absence of a fault in a plurality of pieces of recognition reference data; and

recovering the plurality of pieces of identical recognition reference data if a fault is present in the plurality of pieces of recognition reference data.

5. The method according to claim 4 , wherein

the plurality of pieces of recognition reference data are periodically examined for the presence or the absence of a fault.

6. A method for recognizing a specific object from acquired images, the method comprising:

storing a plurality of pieces of identical recognition reference data within a storage unit, the plurality of pieces of identical recognition reference data including references for recognizing the specific object;

recognizing the specific object within the acquired images by switching between the plurality of pieces of identical recognition reference data within the storage unit on a frame basis for consecutive frames of the acquired images;

determining whether the plurality of pieces of identical recognition reference data meet a predetermined matching condition;

outputting the recognition results as effective recognition results based on whether the plurality of pieces of identical recognition reference data meet the predetermined matching condition;

determining a fault in the recognition reference data of an odd number of recognition results based on majority rule logic, wherein recognition results comprising a minority are determined as being faulty, and

recovering the recognition reference data determined as being faulty.

7. A system for recognizing a specific object from acquired images, the system comprising:

a central processor;

a computer-readable storage medium which stores a plurality of pieces of identical recognition reference data which are used as references for recognizing a specific object from an image; and

a set of computer-executable instructions stored on the computer-readable storage medium that cause the central processor to implement:

recognizing the specific object within the acquired images by switching between the plurality of pieces of identical recognition reference data within the computer-readable storage medium on a frame basis for consecutive frames of the acquired images;

determining whether the plurality of pieces of identical recognition reference data meet a predetermined matching condition;

outputting the recognition results as effective recognition results based on whether the plurality of pieces of identical recognition reference data meet the predetermined matching condition;

examining for a presence or an absence of a fault in a plurality of pieces of recognition reference data; and

recovering the plurality of pieces of identical recognition reference data if a fault is present in the plurality of pieces of recognition reference data.

8. The system according to claim 7 , wherein

the plurality of pieces of recognition reference data are periodically examined for the presence or the absence of a fault.

9. A system for recognizing a specific object from acquired images, the system comprising:

a central processor;

a computer-readable storage medium which stores a plurality of pieces of identical recognition reference data which are used as references for recognizing a specific object from an image; and

a set of computer-executable instructions stored on the computer-readable storage medium that cause the central processor to implement:

recognizing the specific object within the acquired images by switching between the plurality of pieces of identical recognition reference data within the computer-readable storage medium on a frame basis for consecutive frames of the acquired images;

determining whether the plurality of pieces of identical recognition

reference data meet a predetermined matching condition; and

outputting the recognition results as effective recognition results based on whether the plurality of pieces of identical recognition reference data meet the predetermined matching condition;

determining a fault in the recognition reference data of an odd number of recognition results based on majority rule logic, wherein recognition results comprising a minority are determined as being faulty; and

recovering the recognition reference data determined as being faulty.