US 11,756,306 B2
Anti-drowning safety alarm method and device for swimming pool
Xiao Feng Ren, Hangzhou (CN); Xin Xie, Hangzhou (CN); Yu Guo, Hangzhou (CN); Dong Yan Guo, Hangzhou (CN); Zhen Hua Wang, Hangzhou (CN); Jian Hua Zhang, Hangzhou (CN); and Du Si Zhang, Hangzhou (CN)
Assigned to Hangzhou Juyan Xincheng Technology Co. Ltd., Hangzhou (CN)
Filed by Hangzhou Juyan Xincheng Technology Co. Ltd., Hangzhou (CN)
Filed on Jan. 8, 2023, as Appl. No. 18/94,388.
Claims priority of application No. 202210022875.6 (CN), filed on Jan. 10, 2022.
Prior Publication US 2023/0222804 A1, Jul. 13, 2023
Int. Cl. G06V 20/52 (2022.01); G06T 7/292 (2017.01); G06T 7/246 (2017.01); G06V 10/82 (2022.01); G06V 10/774 (2022.01); G06T 7/73 (2017.01); H04N 7/18 (2006.01); G08B 21/02 (2006.01); G06V 40/10 (2022.01); H04N 23/90 (2023.01)
CPC G06V 20/52 (2022.01) [G06T 7/248 (2017.01); G06T 7/292 (2017.01); G06T 7/74 (2017.01); G06V 10/774 (2022.01); G06V 10/82 (2022.01); G06V 40/10 (2022.01); G08B 21/02 (2013.01); H04N 7/181 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/20221 (2013.01); G06T 2207/30196 (2013.01); G06T 2207/30241 (2013.01); G06T 2207/30244 (2013.01); G06V 2201/07 (2022.01); H04N 23/90 (2023.01)] 7 Claims
OG exemplary drawing
 
1. An anti-drowning safety alarm method for swimming pools, which includes:
Acquiring images taken by multiple cameras;
Inputting the images into a target detection model to detect human bodies and human heads, so as to obtain human body target detection boxes and human head target detection boxes;
Calculating a relationship between the human heads and the human bodies according to the human body target detection boxes and the human head target detection boxes, and integrating the human body target detection boxes and the human head target detection boxes to obtain a bounding box sequence;
Performing multi-camera fusion on the human body target detection boxes and the human head target detection boxes to obtain human body boxes and human head boxes;
Inputting the human body boxes and the human head boxes into a relative position discrimination model to determine a relative position relationship between the human bodies or the human heads and the water surface, so as to obtain a relative position relationship sequence;
Calculating a correlation between the bounding box sequence at a current time and that at a previous moment by using the human body boxes and the human head boxes and the relative position relationship sequence, so as to obtain a human tracking trajectory;
Updating a state database based on the bounding box sequence;
Generating alarm information according to the state database and sending the alarm information to a terminal;
Wherein, the multi-camera fusion on the human body target detection boxes and the human head target detection boxes to obtain the human body boxes and the human head boxes includes the following steps:
Determining a reference camera image and a registration camera image;
Using Zhang Zhengyou calibration method to calibrate internal and external parameters of a reference camera and a registration camera;
Preprocessing the reference camera image and the registration camera image to obtain a local world coordinate system;
Calculating a relative pose between the reference camera and the registration camera;
Associating multiple cameras through the relative pose to establish a global unified world coordinate system;
Determining a position of each camera in the global unified world coordinate system to obtain a spatial pose of each camera;
Mapping the bounding box sequence into a three-dimensional world coordinate system, and calculating a matching degree of two bounding boxes in different images according to a center point position of the human body target detection boxes and the human head target detection boxes, and a size of the bounding boxes; and
Screening out the human body boxes and the human head boxes whose matching degree exceeds a threshold to obtain the human body boxes and the human head boxes.