Granted Patent B2

US 12,307,644 · App. 17/548,927 · Granted May 20, 2025

Inventors: Minhoe Hur (Seoul, KR); Jaesik Min (Gyeonggi-do, KR)

Assignees: Hyundai Motor Company; Kia Corporation

G06T7/0004G06T7/70G06V10/87G06V10/98G07C5/0866G06T2207/20081G06V2201/08

Abstract

Claims (46)

1. A device for inspecting a vehicle, comprising:

a first camera configured to obtain a bottom image of the vehicle;

a second camera configured to obtain an exterior appearance image of the vehicle;

a memory; and

a processor configured to:

generate a plurality of trained models based on a plurality of reference bottom images of the vehicle, the plurality of trained models being varied according to a vehicle model,

store the plurality of trained models in the memory,

determine the vehicle model of the vehicle based on the exterior appearance image of the vehicle,

determine a trained model corresponding to the vehicle model of the vehicle among the plurality of trained models,

recognize a first object and a second object in the bottom image of the vehicle using the determined trained model,

calculate a relative distance between the first object and the second object based on a distance between the first object and the second object in the bottom image of the vehicle and sizes of the first object and the second object in the bottom image of the vehicle according to the determined trained model, and

determine that the vehicle is defective when the relative distance between the first object and the second object is not within a threshold distance.

2. The device of claim 1 , wherein the processor is configured to:

recognize a predetermined object as the first object in the bottom image of the vehicle, and

recognize an object to be inspected as the second object in the bottom image of the vehicle.

3. The device of claim 1 , wherein when the processor fails to recognize one of a plurality of predetermined objects, the processor is configured to recognize a remaining of the plurality of predetermined objects as first objects except for the object failed to be recognized.

4. The device of claim 1 , wherein the processor is configured to:

determine a third object,

calculate a relative distance between the third object and the second object based on the bottom image of the vehicle, and

determine whether the vehicle is defective based on the relative distance between the third object and the second object.

5. The device of claim 1 , wherein the processor is configured to determine that the vehicle is normal, in response to determining that a position of the second object in the bottom image of the vehicle is within the threshold distance from the second object of one of the reference bottom images of the vehicle.

6. The device of claim 1 , further comprising a display,

wherein the processor is configured to operate the display to display whether the vehicle is defective.

7. The device of claim 1 ,

wherein the processor is configured to store information on whether the vehicle is defective in the memory.

8. A method of inspecting a vehicle, the method comprising:

obtaining, by a processor, a bottom image of the vehicle from a first camera;

obtaining, by the processor, an exterior appearance image of the vehicle from a second camera;

generating, by the processor, a plurality of trained models based on a plurality of reference bottom images of the vehicle, the plurality of trained models being varied according to a vehicle model;

storing, by the processor, the plurality of trained models in a memory;

determining, by the processor, the vehicle model of the vehicle based on the exterior appearance image of the vehicle;

determining, by the processor, a trained model corresponding to the vehicle model of the vehicle among the plurality of trained models;

recognizing, by the processor, a first object and a second object in the bottom image of the vehicle using the determined trained model;

calculating, by the processor, a relative distance between the first object and the second object based on a distance between the first object and the second object in the bottom image of the vehicle and sizes of the first object and the second object in the bottom image of the vehicle according to the determined trained model; and

determining, by the processor, that the vehicle is defective when the relative distance between the first object and the second object is not within a threshold distance.

9. The method of claim 8 , wherein the recognizing of at least one first object and at least one second object includes:

recognizing, by the processor, a predetermined object as the first object in the bottom image of the vehicle; and

recognizing, by the processor, an object to be inspected as the second object in the bottom image of the vehicle.

10. The method of claim 9 , wherein the recognizing of a predetermined object as the first object in the bottom image of the vehicle includes, in response to the processor failing to recognize one of a plurality of predetermined objects, recognizing the other predetermined objects as first objects except for the object failed to be recognized.

11. The method of claim 8 , further comprising:

determining, by the processor, a third object;

calculating, by the processor, a relative distance between the third object and the second object based on the bottom image of the vehicle; and

determining, by the processor, whether the vehicle is defective based on the relative distance between the third object and the second object.

12. The method of claim 8 , further comprising determining that the vehicle is normal, in response to determining that a position of the second object in the bottom image of the vehicle is within the threshold distance from the second object of one of the reference bottom images of the vehicle.

13. The method of claim 8 , further comprising operating a display to display whether the vehicle is defective.

14. The method of claim 8 , further comprising storing information on whether the vehicle is defective in the memory.

Assignments (1)

Priority Claims (1)

Continuity (1)

References Cited (9)