US 12,260,653 B2
Safety belt detection method, apparatus, computer device, and storage medium
Yuqian Zeng, Shanghai (CN); and Tao Zhang, Shanghai (CN)
Assigned to Black Sesame Technologies Inc., San Jose, CA (US)
Filed by Black Sesame Technologies Inc., San Jose, CA (US)
Filed on Apr. 13, 2022, as Appl. No. 17/720,042.
Claims priority of application No. 202110814893.3 (CN), filed on Jul. 19, 2021.
Prior Publication US 2023/0020385 A1, Jan. 19, 2023
Int. Cl. G06V 20/59 (2022.01); G06N 5/01 (2023.01); G06T 7/11 (2017.01); G06T 7/70 (2017.01); G06V 10/25 (2022.01); G06V 10/82 (2022.01); G06V 40/10 (2022.01); G06V 40/16 (2022.01)
CPC G06V 20/593 (2022.01) [G06N 5/01 (2023.01); G06T 7/11 (2017.01); G06T 7/70 (2017.01); G06V 10/25 (2022.01); G06V 10/82 (2022.01); G06V 40/11 (2022.01); G06V 40/162 (2022.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/20132 (2013.01); G06T 2207/30201 (2013.01); G06T 2207/30268 (2013.01)] 15 Claims
OG exemplary drawing
 
1. A safety belt detection method, comprising:
obtaining an image to be detected, wherein the step of obtaining an image to be detected further comprises:
capturing an original image by a camera;
detecting a face in the original image based on a face detection algorithm to obtain a face bounding box and a head posture information;
determining whether a part below the face in the original image meets a predetermined size;
determining a detection result as insufficient detection area if the predetermined size is not met;
dynamically cropping the part below the face in the original image based on the face bounding box and the head posture information to obtain a cropped image if the predetermined size is met; and
pre-processing the cropped image to obtain the image to be detected;
inputting the image to be detected into a detection network, wherein the detection network comprises a global dichotomous branch network and a grid classification branch network;
obtaining a dichotomous result output from the global dichotomous branch network and obtaining a grid classification diagram output from the grid classification branch network based on image classification, wherein the dichotomous result indicates whether a driver is wearing a safety belt, and the grid classification diagram indicates a position information of the safety belt; and
obtaining a detection result of the safety belt based on the dichotomous result and the grid classification diagram, wherein the detection result indicates whether the driver is wearing the safety belt normatively.