Method for autonomously parking a motor vehicle

1. A system comprising a computer including a processor and a memory, the memory including instructions such that the processor is programmed to:

receive an image depicting a parking spot;

determine a length of the depicted parking spot based on a classified endpoint of the parking spot;

compare the determined length to an average length;

determine an endpoint of the depicted parking spot when the determined length is less than the average length, wherein the determined endpoint is distal to the classified endpoint; and

realign the determined endpoint with an endpoint of an adjacent parking spot such that the depicted parking spot is oriented with the adjacent parking spot regardless of whether parking spot markings of the depicted parking spot are aligned properly with the parking spot markings of the adjacent parking spot.

2. The system as recited in claim 1 , wherein the processor is further programmed to actuate a vehicle system of a vehicle using the determined endpoint and the classified endpoint.

3. The system as recited in claim 1 , wherein the processor is further programmed to classify the classified endpoint via a convolutional neural network.

4. The system as recited in claim 3 , wherein the convolutional neural network comprises at least one of a single shot detector, a You-Only-Look-Once network, or a Fast-RCNN network.

5. The system as recited in claim 1 , further comprising a sensor configured to capture images representative of a vehicle environment.

6. The system as recited in claim 5 , wherein the sensor is configured to capture Red-Green-Blue (RGB) images.

7. The system as recited in claim 1 , wherein the classified endpoint and the determined endpoint comprise an x-coordinate and a y-coordinate within the image.

8. The system as recited in claim 1 , wherein the image comprises a stitched image.

9. The system as recited in claim 1 , wherein the processor is further programmed to stitch the stitched image based on a plurality of Red-Green-Blue images received from a plurality of sensors.

10. A method comprising:

receiving an image depicting a parking spot;

determining a length of the depicted parking spot based on a classified endpoint of the parking spot;

comparing the determined length to an average length;

determining an endpoint of the parking spot when the determined length is less than the average length, wherein the determined endpoint is distal to the classified endpoint; and

realigning the determined endpoint with an endpoint of an adjacent parking spot such that the depicted parking spot is oriented with the adjacent parking spot regardless of whether parking spot markings of the depicted parking spot are aligned properly with the parking spot markings of the adjacent parking spot.

11. The method as recited in claim 10 , further comprising: actuating a vehicle system of a vehicle using the determined endpoint and the classified endpoint.

12. The method as recited in claim 10 , wherein the classified endpoint is classified via a convolutional neural network.

13. The method as recited in claim 12 , wherein the convolutional neural network comprises at least one of a single shot detector, a You-Only-Look-Once network, or a Fast-RCNN network.

14. The method as recited in claim 10 , wherein the image is captured by a sensor configured to capture images representative of a vehicle environment.

15. The method as recited in claim 14 , wherein the sensor is configured to capture Red-Green-Blue (RGB) images.

16. The method as recited in claim 10 , wherein the classified endpoint and the determined endpoint comprise an x-coordinate and a y-coordinate within the image.

17. The method as recited in claim 10 , wherein the image comprises a stitched image.

18. The method as recited in claim 17 , further comprising stitching the stitched image based on a plurality of Red-Green-Blue images captured by a plurality of sensors.

19. A system comprising a computer including a processor and a memory, the memory including instructions such that the processor is programmed to:

receive a stitched image depicting a parking spot, the stitched image based on a plurality of images captured by a plurality of sensors;

determine a length of the depicted parking spot based on a classified endpoint of the depicted parking spot;

compare the determined length to an average length;

determine an endpoint of the depicted parking spot when the determined length is less than the average length, wherein the determined endpoint is distal to the classified endpoint;

realign the determined endpoint with an endpoint of an adjacent parking spot such that the depicted parking spot is oriented with the adjacent parking spot regardless of whether parking spot markings of the depicted parking spot are aligned properly with the parking spot markings of the adjacent parking spot; and

actuate a vehicle system of a vehicle using the determined endpoint and the classified endpoint.

20. The system of claim 19 , wherein the classified endpoint is classified via a convolutional neural network.