US 12,254,706 B2
Parking slot detection method and system
Lai Tuan Ho, Ha Noi (VN); Tran Hai Anh Vo, Ha Noi (VN); Viet Toan Do, Ha Noi (VN); and Hai Hung Bui, Ha Noi (VN)
Assigned to VINAI ARTIFICIAL INTELLIGENCE APPLICATION AND RESEARCH JOINT STOCK COMPANY, Ha Noi (VN)
Filed by VINAI ARTIFICIAL INTELLIGENCE APPLICATION AND RESEARCH JOINT STOCK COMPANY, Ha Noi (VN)
Filed on Aug. 24, 2022, as Appl. No. 17/894,983.
Claims priority of application No. 1-2021-07087 (VN), filed on Nov. 5, 2021.
Prior Publication US 2023/0146185 A1, May 11, 2023
Int. Cl. G06V 20/56 (2022.01); G06V 10/26 (2022.01); G06V 10/764 (2022.01); G06V 10/766 (2022.01); G06V 10/77 (2022.01); G06V 10/774 (2022.01); G06V 10/94 (2022.01)
CPC G06V 20/588 (2022.01) [G06V 10/26 (2022.01); G06V 10/764 (2022.01); G06V 10/766 (2022.01); G06V 10/7715 (2022.01); G06V 10/774 (2022.01); G06V 10/94 (2022.01)] 8 Claims
OG exemplary drawing
 
1. A parking slot detection method comprising:
receiving a plurality of images taken from a plurality of cameras mounted on a vehicle in a parking environment;
generating a top view image comprising a surrounding view of the vehicle based on the plurality of images;
processing the top view image using a parking line detection model that has been trained using an annotated dataset to detect parking lines for a parking slot in the parking environment, estimate a bounding box for the parking slot and identify an occupancy state of the parking slot; and
converting pixel coordinate information of the bounding box to vehicle information,
wherein the plurality of images include at least four images comprising a left view image, a right view image, a front view image, and a rear view image,
wherein the plurality of cameras include at least four cameras comprising a left camera mounted on a left side of the vehicle, a right camera mounted on a right side of the vehicle, a front camera mounted on a front side of the vehicle, and a rear camera mounted on a rear side of the vehicle,
wherein the left view image, the right view image, the front view image, and the rear view image are taken from the left camera, the right camera, the front camera, and the rear camera, respectively, and
wherein the parking line detection model comprises:
a backbone configured to extract features from the top view image;
a neck being a feature pyramid network configured to enrich the extracted features;
an anchor sampling mechanism configured to generate an anchor on the top of the enriched features;
a segmentation head configured to perform a segmentation task that detects the parking lines for the parking slot using the generated features;
a regression head configured to perform a regression task that estimates the bounding box for the parking slot and estimates a center of the parking slot using the generated features; and
a classification head configured to perform a classification task that identifies the occupancy state of the parking slot using the generated features.