| Life is above all else.In the context of the increasingly prosperous national economy,people’s requirements for material and spiritual life have also increased.In their spare time,people can always be seen by lakes such as campuses and parks.Unfortunately,due to various reasons Accidents of people accidentally falling into the water often occur.At present,the rescue facilities for people falling into the water are passive rescue facilities,which can only be found passively and rescued manually.The discovery time is late and the rescue is not timely,which leads to the problem of drowning and death of people who fall into the water.To this end,this paper proposes an intelligent recognition and positioning rescue system for falling water on the lake.Using this active rescue facility,under the circumstance of focusing on prevention,people can be found in the water in time,and an alarm will be issued at the same time.Rescue and save precious lives by popping ropes at those who fall into the water.Aiming at lakeside application scenarios,this paper is based on deep learning-based personnel image recognition technology,through a practical drowning person locating algorithm,using an automatic rescue device installed on the lakeside,to intelligently rescue people who accidentally fall into the water by the lakeside.First,the pedestrian detection is carried out in the dangerous area near the lake by combining motion detection and PSPNet deep learning model.Image processing technology is used to solve the centroid of the pedestrian,and the distance from the centroid to the outline of the lake is calculated to determine whether the pedestrian falls into the water,and the pixel position of the centroid of the drowning person is obtained.;Then,according to the lake surface environment,a plane-based positioning algorithm is derived,and the actual position of the drowning person relative to the camera can be obtained through the pixel position of the drowning person.Finally,an automatic rescue device is designed to control the direction with a stepping motor,and the spring The elastic force is used as the driving force,and the rope is automatically and quickly ejected to the direction of the drowning person according to the information of the falling position sent by the monitoring center to rescue.The system is tested by building an experimental platform.Using a laptop as the monitoring center,the Honor V20 mobile phone records the video of falling into the water.The camera has a focal length of 26 mm and a resolution of 1280*720,which can accurately identify the range of 20 m;locate the person falling into the water.The average absolute error of the algorithm within 20 m is less than 0.5 m,and the average error of the declination angle is less than 0.5.;When someone is found falling into the water in the monitoring area by the lake,an alarm can be issued in time and the rescue device can be triggered,and the rescue rope can be ejected to the direction of the falling point.The experimental results show that the system has the characteristics of initiative,rapidity and accuracy,which is of great significance for reducing lakeside drowning incidents. |