Apparatus and method for recognizing object

1. An object recognition apparatus comprising:

an object information collector configured to collect object information based on driving information of a vehicle and surrounding detection information detected around the vehicle;

an object selector configured to select an object corresponding to a predetermined first tolerance range based on the object information;

a tolerance range setter configured to set a second tolerance range based on the driving information and the object information when a moving object exists among the objects corresponding to the first tolerance range; and

a guardrail recognizer configured to determine whether the moving object falls within the second tolerance range based on the object information, and recognize a guardrail based on object information of the moving object corresponding to the second tolerance range,

wherein the object selector clusters the objects located within a specific range using a clustering with respect to the objects included in the object information, and selects the objects corresponding to a cluster distributed within a predetermined reference longitudinal interval by checking a longitudinal distribution of the clustered clusters.

2. The object recognition apparatus of claim 1 , wherein the object information collector calculates a relative speed of the object with respect to the vehicle based on the driving information and the surrounding detection information, and collects object information on at least one object from among the moving object and a stationary object classified according to the relative speed.

3. The object recognition apparatus of claim 1 , wherein the tolerance range setter obtains a vehicle speed of the vehicle from the driving information, extracts an incidence angle of the moving object from the object information, and sets the second tolerance range based on the incidence angle and the vehicle speed.

4. The object recognition apparatus of claim 3 , wherein the tolerance range setter calculates a reference vector which has a direction opposite to a direction of the vehicle speed and is equal to a magnitude of the vehicle speed, calculates a unit vector having an angle equal to the incidence angle from the reference vector, performs a projection operation of the reference vector onto the unit vector, and sets the second tolerance range based on a reference speed calculated by the projection operation.

5. The object recognition apparatus of claim 4 , wherein the tolerance range setter calculates the unit vector having an angle equal to the sum of the pre-stored mounting angle of a radar sensor and the incidence angle from the reference vector.

6. The object recognition apparatus of claim 1 , wherein the guardrail recognizer acquires the relative speed of the moving object from the object information, compares the relative speed of the moving object with a reference speed included in the second tolerance range, and determines whether the moving object falls within the second tolerance range based on the comparison result.

7. The object recognition apparatus of claim 6 , wherein the guardrail recognizer calculates a difference value between the relative speed of the moving object and the reference speed, and if the difference value is within a predetermined error range, determines that the moving object falls within the second tolerance range.

8. The object recognition apparatus of claim 6 , wherein the guardrail recognizer, when the moving object falls within the second tolerance range, generates correction object information by inversely compensating the relative speed of the moving object, and recognizes the correction object included in the correction object information as the guardrail.

9. The object recognition apparatus of claim 1 , wherein the guardrail recognizer, when the object included in the first tolerance range is a stationary object, recognizes the guardrail based on object information of the object included in the first tolerance range.

10. An object recognition method comprising:

collecting object information based on driving information of a vehicle and surrounding detection information detected around the vehicle;

selecting an object corresponding to a predetermined first tolerance range based on the object information;

setting a second tolerance range based on the driving information and the object information when a moving object exists among the objects corresponding to the first tolerance range; and

determining whether the moving object falls within the second tolerance range based on the object information, and recognize a guardrail based on object information of the moving object corresponding to the second tolerance range,

wherein collecting object information comprises clustering the objects located within a specific range using a clustering with respect to the objects included in the object information, and selecting the objects corresponding to a cluster distributed within a predetermined reference longitudinal interval by checking a longitudinal distribution of the clustered clusters.

11. The object recognition method of claim 10 , wherein setting the second tolerance range comprises obtaining a vehicle speed of the vehicle from the driving information, extracting an incidence angle of the moving object from the object information, and setting the second tolerance range based on the incidence angle and the vehicle speed.

12. The object recognition method of claim 11 , wherein setting the second tolerance range comprises calculating a reference vector which has a direction opposite to a direction of the vehicle speed and is equal to a magnitude of the vehicle speed, calculating a unit vector having an angle equal to the incidence angle from the reference vector, performing a projection operation of the reference vector onto the unit vector, and setting the second tolerance range based on a reference speed calculated by the projection operation.

13. The object recognition method of claim 10 , wherein recognizing the guardrail comprises acquiring the relative speed of the moving object from the object information, comparing the relative speed of the moving object with a reference speed included in the second tolerance range, and determining whether the moving object falls within the second tolerance range based on the comparison result.

14. The object recognition method of claim 10 , wherein recognizing the guardrail comprises, when the object included in the first tolerance range is a stationary object, recognizing the guardrail based on object information of the object included in the first tolerance range.