CN117842825A - Real-time three-dimensional protection method and system for escalator - Google Patents

Abstract

The invention provides a real-time three-dimensional protection method and a real-time three-dimensional protection system for an escalator, which are based on passenger identification and behavior detection of image identification; counting the number of people riding on the escalator, calculating the current people flow density, and judging whether the current people flow density exceeds the maximum transportation capacity; analyzing the pressure of each step on the elevator supporting frame, judging the standing distribution condition of personnel on the escalator, and analyzing whether each step exceeds the maximum bearing capacity; analyzing whether passengers fall down or not by monitoring the change condition of the step pressure; analyzing whether gaps between the steps are normal or not and whether the steps are damaged or not, and judging whether foreign matters are involved between the steps or not; monitoring the running power, temperature and traction power system state of the escalator, and judging whether the running state of the escalator is normal or not; and comprehensively judging by adopting a safety protection method based on rules.

Description

Real-time three-dimensional protection method and system for escalator

Technical Field

The invention relates to the field of escalator safety protection, in particular to a real-time three-dimensional protection method and system for an escalator.

Background

Escalator is a common facility in life, and provides convenience for people to go upstairs and downstairs. But safety accidents also occur in the running process of the escalator, so that personal injury and property loss of passengers are caused. The potential safety hazards of the escalator include the following aspects:

(1) Safety problem of escalator itself

a. Mechanical failure: mechanical components such as a driving system, a transmission device, a control device and the like of the escalator can be out of order, so that the escalator cannot normally operate, and even stops operating.

b. Electrical failure: the electrical control system, the motor, the power supply and other electrical components of the escalator possibly fail, so that the escalator cannot normally operate, and even safety accidents such as electric shock and the like occur.

c. Protection against device shortages or lack of standardization: protection devices for escalators such as scram switches, safety belts, guard rails, etc. may be deficient or irregular, resulting in injury to passengers or other safety hazards.

d. Overload operation: the escalator is easy to damage mechanical parts of the escalator when running under overload conditions, thereby causing safety accidents.

(2) Safety problem of non-normative behavior of passengers

a. Incorrect use of the escalator: incorrect use of the escalator by passengers is prone to fall or injury. For example, passengers walk or run on an escalator or stand on unsafe locations such as the sides of an escalator.

b. Passengers do not notice the safety of children or elderly people: children or elderly people need more attention and protection while riding on the escalator, but passengers often do not notice their safety, resulting in injuries or other safety accidents for the children or elderly people.

c. Passengers do not comply with regulations: passengers do not comply with regulations, such as carrying large luggage, pushing a baby carriage, etc., which easily affect the normal operation of the escalator and even cause safety accidents.

The existing safety protection device of the escalator mainly comprises: anti-pinch device, anticreeper, blocking device, antiskid device, protective baffle, emergency stop device, handrail area entry protection device, fishback protection device, access cover and floor cover protection device, overspeed protection device, non-manipulation reverse protection device, drive element protection device of step, the distance between drive device and the turning device shortens protection device, the subsidence protection device of step, the loss protection device of step, handrail area velocity deviation protection device etc. these devices better have protected passenger's health, but still have the emergence of some incident. The main reasons are that the protection devices do not form a three-dimensional real-time protection system, for example, emergency stop opening needs manual operation, timeliness is low, and the emergency situation can not respond timely. These devices do not actively identify passenger behavior and do not predict and identify risk. The safety and bearing condition of the escalator steps are not sensed finely. The early warning prompt and the warning information are insufficient. The information of various sensors does not form fusion decision, and the safety protection benefit is not exerted to the maximum.

Disclosure of Invention

The invention aims to solve the problems in the prior art and provides a real-time three-dimensional protection method of an escalator, which comprises the following steps of S1: occupant identification and behavior detection based on image recognition; the method comprises the steps of identifying video streams of IP cameras at the top of the escalator, and judging abnormal behaviors of the entrance by an intelligent terminal with a binocular camera at the entrance of the escalator;

step S2: counting the number of people riding the escalator through video analysis, calculating the current people flow density, judging whether the current people flow density exceeds the maximum transportation capacity or not, and scheduling the people;

step S3: judging falling, clamping feet and detecting steps through a plurality of cameras and sensors; analyzing the pressure of each step on the elevator supporting frame, judging the standing distribution condition of personnel on the escalator, and analyzing whether each step exceeds the maximum bearing capacity; analyzing whether passengers fall down or not by monitoring the change condition of the step pressure;

step S4: analyzing whether gaps between the steps are normal or not and whether the steps are damaged or not, and judging whether foreign matters are involved between the steps or not; the elevator comb plate is provided with cameras with two opposite sides, and the damage, foreign matters and holes in the steps and shoes are judged through video analysis, and any one or more of the long skirts easily cause dangerous situations caused by the step winding;

step S5: reading the fault state of the escalator from the control cabinet; detecting the handrail, namely detecting crack abnormality on the handrail through a handrail photoelectric sensor;

step S6: and combining the step S1 to the step S5, and judging by adopting a safety protection method based on rules.

Preferably, the step S1 further includes:

the image recognition includes risk factor recognition, risk behavior recognition, and abnormal device recognition.

Preferably, the number of people taking the escalator is calculated according to the following method, the length of the escalator is L, the running speed is V, the count of the entering end is I (t) at the moment t, and the count of the exiting end is O (t); at time t, the number of people Np (t) on the escalator is

And analyzing the bearing condition of each step by combining the step pressure sensor.

Preferably, when analyzing whether the passengers fall down, judging by combining the behavior recognition result of the passengers of the cameras; determining the position of a step pressure sensor at the moment according to the abnormal change of the step pressure of the escalator;

determining the position of a camera for monitoring the area according to the position, analyzing the information of the camera, and judging whether jumping and walking behaviors exist or not;

if yes, abnormality is eliminated, no alarm is given, otherwise, personnel fall alarm information is sent;

if the pressure of a certain step is large and the step is uniformly changed according to the running direction of the elevator, the situation that the number of people or articles standing on the step is large is indicated, passengers need to be reminded of safety, and the running speed of the elevator can be reduced if necessary.

Preferably, when the monitoring camera recognizes that the child sits on/takes a large piece of luggage, fusion is not needed, and the decision result is that an audible and visual alarm is adopted to send out a prompt sound to take care of the child/take the large piece of luggage;

when the monitoring camera recognizes that a person walks on the elevator, fusion is not needed, and the decision result is that an acousto-optic alarm gives out a prompt sound and a stable stop;

when the monitoring camera recognizes that a person falls down, the monitoring camera needs to fuse, the pressure sensor value is fused, normal squatting actions are eliminated, the result is perceived as an audible and visual alarm prompt tone, the ordered evacuation and the alarm lamp flash, the emergency execution is that the elevator is stopped, and the personnel operation is that the person on duty is notified;

when the passenger gets in and out of the photoelectric sensor recognition result is that the passenger flow of the elevator is large, the passenger needs to be fused, elevator operation parameters are fused, the problem of elevator failure is solved, an audible and visual alarm gives out a prompt tone, people are crowded, safety is noted, emergency execution is not needed, and personnel operation is to inform an operator on duty to limit the passenger flow;

the step pressure sensing recognition result is that the pressure distribution is uneven, when the individual step pressure is large, fusion is needed, the step fault is eliminated by fusing the information of the escalator step photoelectric sensor, and the acousto-optic warning result is that the escalator step is scattered standing;

when the handrail belt photoelectric sensor identification result shows that the handrail belt has safety risk, fusion is not needed, emergency execution is carried out to reduce the running speed, and personnel operation is carried out to inform the operator on duty to check and process.

Preferably, step S6 further includes decision making for multi-sensor data fusion, and a fusion method based on DS evidence theory is used to fusion identify the pedestrian fall situation by using the monitoring camera and the step pressure sensor, and the specific method is as follows:

determining a fused sensor type and a hypothesis identification frame, wherein the fused sensor is a monitoring camera and a step pressure sensor, and the identification frame is omega= { omega 1, omega 2, & gt, omega N }, such as { fall, bend over, squat down, shelter };

a basic overrule distribution table is established, two kinds of sensors are fused for example, the identification results given by each sensor are given in a probability form, the sensor identification results are fused based on DS evidence theory,

wherein->

Wherein A represents elevator behaviors needing fusion recognition, B and C are different sensor recognition results, and B, C is one of { falling, bending, squatting and shielding }, taking a monitoring camera and a step pressure sensor as examples; m1 and m2 are respectively recognition rate functions of two sensors and are related to the characteristics of the sensors.

The invention further provides a real-time three-dimensional protection system of the escalator, which mainly comprises an information acquisition sensor, a control host, an alarm device and an execution device;

the information acquisition sensor is an information acquisition sensor aiming at each part of the escalator and a sensor for acquiring information aiming at the load of the escalator;

the control host is mainly used for receiving information acquired by each sensor, processing the information, judging whether abnormal operation conditions of the escalator exist or not, giving an alarm aiming at the special conditions, and outputting a control instruction to act on the execution device;

the alarm device comprises an audible and visual alarm device on the escalator site and alarm information sent to an escalator maintenance attendant;

the executing device is mainly an elevator safety control device.

Preferably, the information acquisition sensor comprises a plurality of paths of image sensors, and the plurality of paths of image sensors input information to the control host for information processing.

Preferably, the information acquisition sensor comprises a photoelectric sensor and a step pressure sensor, wherein the photoelectric sensor counts the number of passengers riding on the escalator based on the passengers entering and exiting the escalator, and the step pressure sensor analyzes the bearing condition of each step.

Preferably, the information acquisition sensor comprises an escalator step photoelectric sensor and is used for analyzing whether gaps between steps are normal or not and whether the steps are damaged or not and judging whether foreign matters are involved between the steps or not; the photoelectric sensors of the escalator steps are distributed at the bottoms of the escalator steps in a scattered mode, and the escalator is monitored in a full time mode.

Preferably, the information acquisition sensor comprises an escalator self sensor, monitors the running power, the temperature and the traction power system state of the escalator, and judges whether the running state of the escalator is normal or not.

Compared with the prior art, the invention has the beneficial effects that:

according to the invention, by arranging the sensors of multiple types, the elevator is monitored and the passenger behaviors are monitored, and the control host is utilized to process the sensor information, so that the running safety state of the escalator is analyzed, early warning is performed in advance, the safety operation is rapidly executed, and the safety protection effect of the escalator is improved.

Drawings

Fig. 1 is a schematic diagram of an escalator real-time three-dimensional protection system according to a first embodiment of the present invention;

fig. 2 is a schematic diagram of a real-time three-dimensional protection method of an escalator according to a second embodiment of the present invention;

fig. 3 is a schematic diagram of a correspondence relationship between pressure and time in an escalator operation according to a second embodiment of the present invention.

Detailed Description

The invention is described in further detail below with reference to the attached drawing figures:

for a better understanding of the present invention, embodiments of the present invention are explained in detail below with reference to the drawings.

Real-time three-dimensional protection system of escalator mainly includes: the system comprises an information acquisition sensor, a control host, an alarm device and an execution device.

(1) The information acquisition sensor is an additional sensor based on the escalator basic sensor. To ensure the reliability of escalator monitoring and protection, two main categories are added. The information acquisition sensor for each part of the escalator is provided with a step pressure sensor and various photoelectric sensors; the other is a sensor for information acquisition of escalator load (riding escalator personnel), mainly a video image sensor.

(2) The control host is mainly used for receiving information acquired by each sensor, processing the information, judging whether the abnormal operation condition of the escalator exists, giving out an alarm aiming at the special condition, outputting a control instruction to act on the execution device, and playing a role in safety protection. For special cases of multi-path video image information processing, a control host generally needs to configure higher-performance computing and storage resources, especially multi-path parallel video image recognition hardware and software, such as GPU clusters and image recognition programs based on deep neural networks.

(3) The alarm device comprises an audible and visual alarm device on the escalator site and alarm information sent to maintenance operators of the escalator; passengers can evacuate the dangerous areas of the elevator through alarm audible and visual information, and on-duty personnel can also timely process faults and problems.

(4) The executing device is mainly an elevator safety control device, such as emergency pause, slow speed, reverse running and the like, and controls the entrance and exit of an escalator gate, an electronic enclosure and the like.

On the other hand, the invention provides a real-time three-dimensional protection method of an escalator, which is based on the fact that a plurality of types of sensors, a control host and an audible and visual alarm device are additionally arranged on the traditional escalator, so that the real-time three-dimensional protection of the escalator is realized.

The main method comprises the following steps:

(1) Occupant identification and behavior detection based on image recognition. And inputting the information of the multipath image sensors to a control host for information processing. Based on the current deep neural network open source image recognition software, the method can recognize:

(1) whether risk factors such as children, big pieces of luggage and the like exist;

(2) and identifying the actions of passengers, and judging whether dangerous behaviors such as walking, playing mobile phones, falling and the like exist on the elevator.

(3) And identifying devices such as an escalator skirt guard board, a tooth comb, a handrail and the like, and analyzing whether the devices are abnormal or not.

Step S1: occupant identification and behavior detection based on image recognition; the intelligent terminal with the binocular camera at the entrance of the escalator judges low-head mobile phones at the entrance and the exit, and the luggage and the like judge early warning abnormal behaviors in advance.

In order to be able to accurately identify the behavior of passengers in crowds, it is necessary to install a plurality of cameras facing the passengers of the escalator.

Camera number n=escalator length L/base spacing r+1.

In areas with dense escalators and large passenger flow, such as shopping malls, stations and the like, the density of cameras can be increased.

Wherein the recognition algorithm comprises: fall, retrograde, access & exit is crowded, big object, wheelchair, perambulator, child take advantage of ladder, play handrail area, squat, low head play cell-phone, hole shoes detects, long skirt detects, step crosses the limit and detects, presss from both sides foot detection, comb plate foreign matter detects, maliciously shelters from the detection, automatic judgement staircase direction of travel detects, empty ladder detects, the statistics of the number of people detects, automatic capture tread detects.

(2) Step S2: counting the number of people riding the escalator through video analysis, calculating the current people flow density, judging whether the current people flow density exceeds the maximum transportation capacity or not, and scheduling the people;

and counting the number of passengers taking the escalator based on the photoelectric sensors of passengers entering and exiting the escalator, calculating the current people flow density, and judging whether the maximum transportation capacity is exceeded.

The number of people riding the escalator is calculated as follows. Let the elevator length be L, the running speed be V, the entry end count be I (t) and the exit end count be O (t) at time t. At time t, the number of people Np (t) on the escalator is:

because photoelectric sensors can cause inaccurate counting due to factors such as shielding of large articles carried, the number of people calculated here is only used as a reference, and the bearing condition of each step needs to be analyzed by combining a step pressure sensor.

(3) Step S3: judging falling, clamping feet and detecting steps through a plurality of cameras and sensors; analyzing the pressure of each step on the elevator supporting frame, judging the standing distribution condition of personnel on the escalator, and analyzing whether each step exceeds the maximum bearing capacity; analyzing whether passengers fall down or not by monitoring the change condition of the step pressure;

based on the pressure of each step to the elevator supporting frame, the step pressure sensor is used for analyzing the distribution situation of personnel standing on the escalator and analyzing whether each step exceeds the maximum bearing capacity. In addition, by monitoring the change of the step pressure, whether the passenger falls down or not can be analyzed.

When analyzing whether the passenger falls down, the judgment is also needed to be carried out by combining the behavior recognition result of the passenger with the camera. And determining the position of the step pressure sensor at the moment according to the abnormal change of the step pressure of the escalator. And determining the position of the camera for monitoring the area according to the position, analyzing the camera information, and judging whether the actions such as jumping, walking and the like exist. If so, the abnormality is eliminated, no alarm is given, otherwise, the personnel fall alarm information is sent out.

In addition, if the pressure of a certain step is large and the pressure is uniformly changed according to the running direction of the elevator, the number of people or articles standing on the step is large, passengers are required to be reminded of safety, and the running speed of the elevator can be reduced if necessary.

Optionally, a plurality of cameras and sensors can be used for judging falling, clamping feet and detecting steps.

(4) Step S4: analyzing whether gaps between the steps are normal or not and whether the steps are damaged or not, and judging whether foreign matters are involved between the steps or not; the elevator comb plate is provided with cameras with two opposite sides, and the damage, foreign matters and holes in the steps and shoes are judged through video analysis, and any one or more of the long skirts easily cause dangerous situations caused by the step winding; currently, some escalators are equipped with such sensors, and the number of such sensors is increased, and the sensors are distributed and deployed at the bottom of the whole escalator step, so that the escalator is monitored at all times, and safety is provided.

(5) Step S5: directly reading the fault state of the escalator from an RS485 interface of the control cabinet, detecting the handrail, and detecting crack abnormality on the handrail through a handrail photoelectric sensor; based on the sensor of the escalator, the running power, the temperature, the traction power system state and the like of the escalator are monitored, and whether the running state of the escalator is normal is judged. People flow density can be identified based on people counting and camera image analysis, but the escalator running state monitoring also needs to be realized by means of sensors of the escalator, particularly running power, system temperature and the like, and alarm information needs to be sent out in time once abnormality occurs.

(6) Based on the sensor information, a security decision is made. A rule-based security approach is employed. And merging analysis results of the sensor information to comprehensively judge.

When the monitoring camera recognizes that the child takes/holds the large piece of luggage, fusion is not needed, and the decision result is that an acousto-optic alarm is adopted to send out prompt sound to take care of the child/hold the large piece of luggage;

when the monitoring camera recognizes that a person walks on the elevator, fusion is not needed, and the decision result is that an acousto-optic alarm gives out a prompt sound and a stable stop;

when the monitoring camera recognizes that a person falls down, the monitoring camera needs to fuse, the pressure sensor value is fused, normal squatting actions are eliminated, the result is perceived as an audible and visual alarm prompt tone, the ordered evacuation and the alarm lamp flash, the emergency execution is that the elevator is stopped, and the personnel operation is that the person on duty is notified;

when the passenger gets in and out of the photoelectric sensor recognition result is that the passenger flow of the elevator is large, the passenger needs to be fused, elevator operation parameters are fused, the problem of elevator failure is solved, an audible and visual alarm gives out a prompt tone, people are crowded, safety is noted, emergency execution is not needed, and personnel operation is to inform an operator on duty to limit the passenger flow;

the step pressure sensing recognition result is that the pressure distribution is uneven, when the individual step pressure is large, fusion is needed, the information of the escalator step photoelectric sensor is fused, and if the step fault is eliminated, the sent acousto-optic warning result is scattered standing;

when the identification result of the handrail belt photoelectric sensor is that the handrail belt has safety risks such as cracks, fusion is not needed, emergency operation is performed to reduce the running speed, and personnel operation is to inform operators on duty to check and process.

The invention provides a fusion method based on DS evidence theory, which aims at decision making of multi-sensor data fusion. The method has universality for fusion of various sensors, and for easy understanding, the pedestrian falling situation is identified by fusion of the monitoring camera and the step pressure sensor, and the method is specifically described as follows:

1. a fusion sensor type and hypothesis recognition framework is determined. The sensors fused here are monitoring cameras and step pressure sensors. The recognition frame Ω= { Ω 1, Ω 2,..q.n }, e.g., { fall, bend over, squat, shelter }.

2. A basic profile allocation table is built, taking two kinds of sensor fusion as an example. The recognition results given by the individual sensors are usually given in the form of probabilities, so that the recognition probabilities can be said to be filled in the following table.

Identifying frames/hypotheses	Elevator monitoring sensor 1	Elevator monitoring sensor 2
			Ω1	m1(Ω1)	m2(Ω1)
Ω2	m1(Ω2)	m2(Ω2)
			...	...	...
ΩN	m1(ΩN)	m2(ΩN)

3. And fusing the sensor identification results based on DS evidence theory.

Wherein the method comprises the steps ofIn the above formula, a represents the elevator behavior to be identified in a fusion way, B and C are the results of identification by different sensors, and B, C is one of { falling, bending down, squatting down, shielding }, taking a monitoring camera and a step pressure sensor as examples. m1 and m2 are respectively recognition rate functions of the two sensors, and parameters of the sensors can be obtained according to historical data or other experiments in the characteristic correlation of the sensors.

Through the formula, the two sensor identification results can be fused, and the identification error rate is reduced. Meanwhile, the method can be popularized to fusion of a plurality of sensor identification results.

In the description of the present invention, it should be noted that, unless explicitly specified and limited otherwise, the terms "connected," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present invention can be understood by those of ordinary skill in the art according to the specific circumstances.

In the description of the present invention, unless otherwise indicated, the terms "upper," "lower," "left," "right," "inner," "outer," and the like are used for convenience in describing the present invention and simplifying the description based on the orientation or positional relationship shown in the drawings, and do not denote or imply that the devices or elements in question must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention.

Finally, it should be noted that the above-mentioned technical solution is only one embodiment of the present invention, and various modifications and variations can be easily made by those skilled in the art based on the application methods and principles disclosed in the present invention, and are not limited to the methods described in the above-mentioned specific embodiments of the present invention, therefore, the foregoing description is only preferred, and not meant to be limiting.

Claims (11)

1. The real-time three-dimensional protection method for the escalator is characterized by comprising the following steps of:

step S1: occupant identification and behavior detection based on image recognition; the method comprises the steps of identifying video streams of IP cameras at the top of the escalator, and judging abnormal behaviors of the entrance by an intelligent terminal with a binocular camera at the entrance of the escalator;

step S2: counting the number of people riding the escalator through video analysis, calculating the current people flow density, judging whether the current people flow density exceeds the maximum transportation capacity or not, and scheduling the people;

step S3: judging falling, clamping feet and detecting steps through a plurality of cameras and sensors; analyzing the pressure of each step on the elevator supporting frame, judging the standing distribution condition of personnel on the escalator, and analyzing whether each step exceeds the maximum bearing capacity; analyzing whether passengers fall down or not by monitoring the change condition of the step pressure;

step S4: analyzing whether gaps between the steps are normal or not and whether the steps are damaged or not, and judging whether foreign matters are involved between the steps or not; the elevator comb plate is provided with cameras with two opposite sides, and the damage, foreign matters and holes in the steps and shoes are judged through video analysis, and any one or more of the long skirts easily cause dangerous situations caused by the step winding;

step S5: reading the fault state of the escalator from the control cabinet; detecting the handrail, namely detecting crack abnormality on the handrail through a handrail photoelectric sensor;

step S6: and combining the step S1 to the step S5, and judging by adopting a safety protection method based on rules.

2. The method according to claim 1, characterized in that said step S1 further comprises:

the image recognition includes risk factor recognition, risk behavior recognition, and abnormal device recognition.

3. The method according to claim 1, characterized in that said step S2 further comprises:

the number of people taking the escalator is calculated according to the following method, the length of the escalator is L, the running speed is V, at the moment t, the count of the entering end is I (t), and the count of the exiting end is O (t); at time t, the number of people Np (t) on the escalator is

And analyzing the bearing condition of each step by combining the step pressure sensor.

4. The method according to claim 1, characterized in that said step S3 further comprises:

when analyzing whether passengers fall down, judging by combining the behavior recognition result of the passengers of the cameras; determining the position of a step pressure sensor at the moment according to the abnormal change of the step pressure of the escalator;

determining the position of a camera for monitoring the area according to the position, analyzing the information of the camera, and judging whether jumping and walking behaviors exist or not;

if yes, abnormality is eliminated, no alarm is given, otherwise, personnel fall alarm information is sent;

if the pressure of a certain step is large and the step is uniformly changed according to the running direction of the elevator, the situation that the number of people or articles standing on the step is large is indicated, passengers need to be reminded of safety, and the running speed of the elevator can be reduced if necessary.

5. The method according to claim 1, characterized in that said step S6 further comprises:

when the monitoring camera recognizes that the child takes/holds the large piece of luggage, fusion is not needed, and the decision result is that an acousto-optic alarm is adopted to send out prompt sound to take care of the child/hold the large piece of luggage;

when the monitoring camera recognizes that a person walks on the elevator, fusion is not needed, and the decision result is that an acousto-optic alarm gives out a prompt sound and a stable stop;

when the monitoring camera recognizes that a person falls down, the monitoring camera needs to fuse, the pressure sensor value is fused, normal squatting actions are eliminated, the result is perceived as an audible and visual alarm prompt tone, the ordered evacuation and the alarm lamp flash, the emergency execution is that the elevator is stopped, and the personnel operation is that the person on duty is notified;

when the passenger gets in and out of the photoelectric sensor recognition result is that the passenger flow of the elevator is large, the passenger needs to be fused, elevator operation parameters are fused, the problem of elevator failure is solved, an audible and visual alarm gives out a prompt tone, people are crowded, safety is noted, emergency execution is not needed, and personnel operation is to inform an operator on duty to limit the passenger flow;

the step pressure sensing recognition result is that the pressure distribution is uneven, when the individual step pressure is large, fusion is needed, the step fault is eliminated by fusing the information of the escalator step photoelectric sensor, and the acousto-optic warning result is that the escalator step is scattered standing;

when the handrail belt photoelectric sensor identification result shows that the handrail belt has safety risk, fusion is not needed, emergency execution is carried out to reduce the running speed, and personnel operation is carried out to inform the operator on duty to check and process.

6. The method according to claim 1, wherein the step S6 further comprises making a decision for multi-sensor data fusion, and the method for fusing the monitoring camera and the step pressure sensor to identify the pedestrian falling based on the DS evidence theory fusion is as follows:

determining a fused sensor type and a hypothesis identification frame, wherein the fused sensor is a monitoring camera and a step pressure sensor, and the identification frame is omega= { omega 1, omega 2, & gt, omega N }, such as { fall, bend over, squat down, shelter };

a basic overrule distribution table is established, two kinds of sensors are fused for example, the identification results given by each sensor are given in a probability form, the sensor identification results are fused based on DS evidence theory,

wherein->

Wherein A represents elevator behaviors needing fusion recognition, B and C are different sensor recognition results, and B, C is one of { falling, bending, squatting and shielding }, taking a monitoring camera and a step pressure sensor as examples; m1 and m2 are respectively recognition rate functions of two sensors and are related to the characteristics of the sensors.

7. An escalator real-time three-dimensional protection system for executing the method of claims 1-6, characterized in that the system mainly comprises an information acquisition sensor, a control host, an alarm device and an execution device;

the information acquisition sensor is an information acquisition sensor aiming at each part of the escalator and a sensor for acquiring information aiming at the load of the escalator;

the control host is mainly used for receiving information acquired by each sensor, processing the information, judging whether abnormal operation conditions of the escalator exist or not, giving an alarm aiming at the special conditions, and outputting a control instruction to act on the execution device;

the alarm device comprises an audible and visual alarm device on the escalator site and alarm information sent to an escalator maintenance attendant;

the executing device is mainly an elevator safety control device.

8. The system according to claim 7, wherein: the information acquisition sensor comprises a plurality of paths of image sensors, and the plurality of paths of image sensors input information to the control host for information processing.

9. The system according to claim 7, wherein: the information acquisition sensor comprises a photoelectric sensor and a step pressure sensor, wherein the photoelectric sensor counts the number of passengers taking the escalator based on the passengers entering and exiting the escalator, and the step pressure sensor analyzes the bearing condition of each step.

10. The system according to claim 7, wherein: the information acquisition sensor comprises an escalator step photoelectric sensor and is used for analyzing whether gaps between steps are normal or not and whether the steps are damaged or not and judging whether foreign matters are involved between the steps or not; the photoelectric sensors of the escalator steps are distributed at the bottoms of the escalator steps in a scattered mode, and the escalator is monitored in a full time mode.

11. The system according to claim 7, wherein: the information acquisition sensor comprises an escalator sensor, monitors the running power, the temperature and the traction power system state of the escalator, and judges whether the running state of the escalator is normal or not.