CN-122009438-A - Machine vision-based non-blind area active drowning prevention monitoring and rescuing system and method

CN122009438ACN 122009438 ACN122009438 ACN 122009438ACN-122009438-A

Abstract

The invention discloses a machine vision-based non-blind area active drowning prevention monitoring and rescue system and method, and belongs to the technical field of drowning prevention monitoring and rescue. The intelligent rescue ship comprises a visual monitoring module and an intelligent rescue ship module, wherein each module consists of specific hardware and forms a hardware structure for cooperative work. The visual monitoring module is erected above a water area, wireless communication connection is established between the visual monitoring module and an intelligent rescue ship module on the water surface through the HC-05 Bluetooth module, the visual monitoring module is an instruction transmitting end, the ship module is an instruction receiving end, and meanwhile, the ship module transmits the real-time position information of the ship module back to the visual monitoring module to form a hardware communication closed loop of perception-decision-execution-feedback, so that all-weather drowning monitoring, state identification and automatic rescue of an open field water area are realized, and the drowning prevention rescue effect is improved.

Inventors

YANG JINRONG
YANG LIQIAO

Assignees

榕力信息科技(上海)有限公司

Dates

Publication Date: 20260512
Application Date: 20260327

Claims (7)

1. The machine vision-based non-blind area active drowning prevention monitoring and rescue system is characterized by comprising a vision monitoring module and an intelligent rescue ship module, wherein the vision monitoring module is an instruction transmitting end, the intelligent rescue ship module is an instruction receiving end, and a bidirectional real-time communication connection is established between the vision monitoring module and the intelligent rescue ship module through a Bluetooth module; the visual monitoring module is erected at a non-blind area position above a field water area and comprises a camera, a single board computer, a first Bluetooth module and an acrylic mounting bracket; the camera is fixed at the global covering position of the water area through an acrylic bracket and is directly connected with a camera interface of the single board computer; the single board computer is arranged on the acrylic bracket base, and is respectively connected with the camera and the first Bluetooth module through the USB/GPIO interface to serve as a data processing unit; The first Bluetooth module is connected with a communication interface of the raspberry group single board computer; the intelligent rescue ship module comprises an Arduino open source development board, a second Bluetooth module, a motor driving module, a steering engine mounting platform, RGB state indication lamp beads and a double-motor power system; the Arduino open source development board is respectively connected with the second Bluetooth module, the motor driving module, the steering engine and the RGB lamp beads through digital/analog interfaces; The second Bluetooth module is connected with the Arduino development board communication interface; the motor driving module is connected in series between the Arduino development board and the double-motor power system; The steering engines and the steering engine installation platforms are 2 in number, are arranged on a CAD customized acrylic installation platform, a steering engine rotating shaft is connected with a T-shaped bearing structure, the bearing structure is used for placing a life buoy and an underwater self-rescue bracelet, the rotating angle of the steering engines is 10-130 degrees, and the stable throwing of life-saving equipment is realized; the RGB state indicator lamp beads are connected with the I/O interface of the Arduino development board and are exposed at the obvious positions of the ship body; The double-motor power system is connected with the motor driving module.
2. The machine vision non-blind area based active drowning prevention monitoring and rescuing system according to claim 1, wherein the camera is 720p 60-frame high-definition non-distortion camera.
3. The machine vision non-blind area based active drowning prevention monitoring and rescuing system according to claim 1, wherein the single board computer is a raspberry-based single board computer.
4. The machine vision non-blind area based active drowning prevention monitoring and rescuing system according to claim 1, wherein the first Bluetooth module and the second Bluetooth module are HC-05 Bluetooth modules.
5. The machine vision non-blind area based active drowning prevention monitoring and rescue system of claim 4, wherein the baud rate of the HC-05 bluetooth module is set to 9600bps.
6. The machine vision non-blind area based active drowning prevention monitoring and rescue system of claim 1, wherein the steering engine is an MG90D steering engine.
7. The machine vision-based non-blind area active drowning prevention monitoring and rescue method is characterized by comprising the following steps of: S1, a visual monitoring module starts real-time image acquisition; s2, performing human body key point identification by a raspberry group single board computer, and extracting characteristic parameters; s3, the raspberry group single board computer executes swimming/drowning state judgment to generate a disposal instruction; S4, the raspberry-pie single board computer performs dynamic path planning based on a PID algorithm; S5, the visual monitoring module sends an instruction to the rescue ship module; S6, the intelligent rescue ship module receives the instruction and realizes accurate navigation; s7, performing grading rescue treatment by the Arduino development board according to the state code; and S8, after the rescue action is executed, the visual monitoring module restores the real-time monitoring state of the water surface, waits for the next target to appear, and forms an all-weather circular monitoring and rescue closed loop.

Description

Machine vision-based non-blind area active drowning prevention monitoring and rescuing system and method Technical Field The invention relates to the technical field of drowning prevention monitoring and rescue, in particular to a machine vision-based non-blind area active drowning prevention monitoring and rescue system and method. Background In recent years, teenagers frequently suffer from drowning accidents, and the drowning accidents are continuously and highly caused by the lack of effective supervision means and intelligent early warning mechanisms for the first time of accidental death of children in China in long-term residence, especially in open field water areas (ponds, river ditches, reservoirs and the like) in rural areas and urban and rural joints. The existing drowning rescue equipment is mainly divided into three types, ①, namely, a remote control unmanned aerial vehicle which throws a life buoy, is high in flying speed but is greatly influenced by weather, can not cruise for 24 hours, needs to be controlled by a police to arrive at the scene, is limited in application scene, ②, namely, a water remote control life boat, needs manual field operation, waits for a time difference of arrival of rescue workers to cause rescue lag, ③, namely, a wearable underwater self-rescue device, needs a user to purchase and wear in advance, is impractical for key groups such as students and the like, and can not actively find a drowning target. Meanwhile, the existing drowning prevention means mainly comprise propaganda education, warning mark establishment and manual inspection, have the limitations of high manpower inspection cost, limited coverage range and incapability of all-weather monitoring, and have the defects of perception-execution integration design of part of intelligent drowning prevention equipment, are easily influenced by water shielding and visual angle blind areas, have low positioning precision and response lag, and are difficult to adapt to sudden drowning scenes of open wild water areas. The above prior art mainly has the following core problems: The monitoring layer is lack of non-blind area and all-weather automatic monitoring equipment aiming at open water areas in the field, the existing equipment is either required to be manually attended, or the sensing range is limited, so that a drowning target cannot be found in time; The state discrimination method of the existing intelligent equipment is complex, or depends on a wearable sensor, a non-contact rapid state discrimination means matched by a user is not needed, and swimming and drowning states are difficult to distinguish; The existing rescue equipment is mostly manually operated, has the problems of lag response and low positioning precision, has no grading treatment mechanism, and can not realize differential rescue according to swimming/drowning states; The adaptation level is that the existing high-end intelligent rescue equipment is complex to deploy, expensive in price, and lack of a low-cost and easy-to-install anti-drowning system for adapting rural and urban and rural junction field waters. Disclosure of Invention The invention provides an active drowning prevention monitoring and rescuing system and method based on machine vision, which effectively solve the problems of monitoring blind areas, complex recognition, rescue lag, low positioning precision, poor suitability and the like in the prior art, so as to realize all-weather drowning monitoring, state recognition and automatic rescue of open field water areas and improve the drowning prevention rescuing effect. In order to achieve the above purpose, the invention adopts the following technical scheme: The invention provides a machine vision-based non-blind area active drowning prevention monitoring and rescue system, which comprises a vision monitoring module and an intelligent rescue ship module, wherein the vision monitoring module is an instruction transmitting end, the intelligent rescue ship module is an instruction receiving end, and a bidirectional real-time communication connection is established between the vision monitoring module and the intelligent rescue ship module through a Bluetooth module; the visual monitoring module is erected at a non-blind area position above a field water area and comprises a camera, a single board computer, a first Bluetooth module and an acrylic mounting bracket; the camera is fixed at the global covering position of the water area through an acrylic bracket and is directly connected with a camera interface of the single board computer; the single board computer is arranged on the acrylic bracket base, and is respectively connected with the camera and the first Bluetooth module through the USB/GPIO interface to serve as a data processing unit; The first Bluetooth module is connected with a communication interface of the raspberry group single board computer; the intelligent rescue ship module comprises an Arduino open source