CN112347862A - Elevator door fault real-time detection method based on machine vision - Google Patents
Elevator door fault real-time detection method based on machine vision Download PDFInfo
- Publication number
- CN112347862A CN112347862A CN202011123569.9A CN202011123569A CN112347862A CN 112347862 A CN112347862 A CN 112347862A CN 202011123569 A CN202011123569 A CN 202011123569A CN 112347862 A CN112347862 A CN 112347862A
- Authority
- CN
- China
- Prior art keywords
- door
- elevator door
- elevator
- time
- open
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000011897 real-time detection Methods 0.000 title claims abstract description 10
- 238000012544 monitoring process Methods 0.000 claims abstract description 26
- 238000001514 detection method Methods 0.000 claims abstract description 23
- 238000012549 training Methods 0.000 claims description 30
- 238000012360 testing method Methods 0.000 claims description 12
- 230000002159 abnormal effect Effects 0.000 claims description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 description 7
- 230000005856 abnormality Effects 0.000 description 4
- 230000003068 static effect Effects 0.000 description 4
- 230000001133 acceleration Effects 0.000 description 3
- 238000001914 filtration Methods 0.000 description 3
- 238000012423 maintenance Methods 0.000 description 3
- 238000002372 labelling Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000005034 decoration Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000000491 multivariate analysis Methods 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/0006—Monitoring devices or performance analysers
- B66B5/0018—Devices monitoring the operating condition of the elevator system
- B66B5/0031—Devices monitoring the operating condition of the elevator system for safety reasons
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
Abstract
The invention relates to a machine vision-based elevator door fault real-time detection method, which comprises the following steps: s1, collecting a video stream containing an elevator door image, and intercepting the elevator door image in the video stream according to a time sequence according to a time interval of a fixed threshold value; s2, obtaining the elevator door image, judging the state of an elevator door in the elevator door image based on a door state monitoring model, and obtaining the elevator door time sequence of the operation of the elevator door; and S3, detecting the fault of the elevator door according to the elevator door time sequence. The invention has good detection real-time property and high accuracy.
Description
Technical Field
The invention relates to the technical field of elevator safe operation monitoring, in particular to a real-time elevator door fault detection method based on machine vision.
Background
In recent years, with the coming of a series of policies such as urbanization, old building transformation, elevator installation and the like, the number of urban elevators is increasing day by day, so that the urban elevators become indispensable transportation means in modern urban buildings and bring great convenience for people to go out. But at the same time the problem of operational safety of elevators is getting more and more attention. Door trouble is one of the common trouble of elevator, to the real-time accurate detection of door trouble, in time maintenance, can avoid more serious incident to take place on the one hand, and on the other hand also can reduce the elevator and stop long when the ladder, promotes resident's the comfort of taking, and among the prior art, the real-time detection to the lift-cabin door is realized based on the sign that is showing on the door usually, and it receives the influence of elevator internal environment, light etc. very easily and leads to the rate of accuracy step-down of whole testing process. For example, chinese patent CN201610637687 discloses an elevator operation and maintenance monitoring method based on an intelligent visual light curtain, which analyzes elevator door closing abnormality, door opening operation and elevator peak-shifting and stopping abnormality based on an intelligent visual algorithm, and specifically, the identification of the elevator door closing abnormality is to collect images of multiple door switches in a state without foreign matter intrusion or when the doors can be normally closed, then perform feature extraction and feature dimension reduction on the images, then model the closed door images by applying kmeans and Rbf model methods, and when a new image is collected, compare the new image with the modeled template to know whether the door is closed, on one hand, the invention needs to pre-collect some sample images for modeling each newly accessed elevator, which is complicated, on the other hand, even if the door is closed in the elevator, there may be multiple scenes, light change, elevator decoration, elevator advertisement, etc., only through the pre-collected model, all situations are difficult to cover, so that the accuracy of the model is lowered when a specific situation is met; in addition, the elevator door closing abnormity of the elevator door can not distinguish more detailed subcategories, such as the door full-open abnormity or the door half-open abnormity of the elevator door closing abnormity for a long time. The elevator door opening operation detection is based on the identified door opening and closing information, the elevator operation state is detected by combining a Kalman filtering method, and the identification precision is possibly reduced due to the diversity of scenes of pictures when the elevator operation is judged based on the images.
Disclosure of Invention
The invention aims to provide a real-time elevator door fault detection method and a real-time elevator door fault detection system based on multivariate analysis, which solve the defect of low accuracy of real-time abnormal elevator detection.
In order to achieve the above object, the present invention provides a real-time elevator door fault detection method based on machine vision, which comprises:
s1, collecting a video stream containing an elevator door image, and intercepting the elevator door image in the video stream according to a time sequence according to a time interval of a fixed threshold value;
s2, obtaining the elevator door image, judging the state of an elevator door in the elevator door image based on a door state monitoring model, and obtaining the elevator door time sequence of the operation of the elevator door;
and S3, detecting the fault of the elevator door according to the elevator door time sequence.
According to one aspect of the invention, the status of the elevator door comprises: full open, half open and close;
the elevator door failure comprises: the door is repeatedly opened and closed, the door is difficult to open and close, the door opening and closing speed is abnormal, the door is fully opened for a long time, the door is half opened for a long time, and the ladder is driven when the door is opened.
According to an aspect of the present invention, in step S3, in the step of detecting the elevator door fault according to the elevator door timing sequence, if the elevator door has multiple cycles of "closed door, half open, and full open" in the elevator door timing sequence obtained according to the video stream, it is determined that the elevator door fault is a repeated door opening and closing.
According to an aspect of the present invention, in step S3, in the step of detecting the failure of the elevator door according to the elevator door timing sequence, if the elevator door in the elevator door timing sequence obtained according to the video stream has multiple cycles of "closed door-half open" or multiple cycles of "full open-half open", it is determined that the failure of the elevator door is difficult to open or close the door.
According to an aspect of the present invention, in step S3, in the step of detecting the elevator door fault according to the elevator door timing sequence, if the time consumed for the "closing/fully opening" of the elevator door in the elevator door timing sequence acquired according to the video stream exceeds a preset threshold or the time consumed for the "fully opening/closing" of the elevator door is lower than a preset threshold, it is determined that the elevator door fault is an abnormal door opening/closing speed.
According to an aspect of the present invention, in step S3, in the step of detecting the elevator door fault according to the elevator door time sequence, if the duration of the fully opened state of the elevator door in the elevator door time sequence obtained according to the video stream is greater than a preset threshold, it is determined that the elevator door fault is that the door is fully opened for a long time;
and if the half-open state duration time of the elevator door in the elevator door time sequence acquired according to the video stream is greater than a preset threshold value, judging that the elevator door fault is long-time half-open of the elevator door.
According to an aspect of the present invention, in step S3, in the step of detecting the elevator door fault according to the elevator door timing sequence, if the elevator door is fully opened or half opened in the elevator door timing sequence obtained according to the video stream and the elevator is in a motion state, it is determined that the elevator door fault is open door and elevator is going.
According to one aspect of the invention, the door condition monitoring model is generated based on the steps comprising:
s21, acquiring an image containing the opening and closing state of the elevator door as a training sample;
and S22, after the training samples are labeled, inputting the training samples into a convolutional neural network for model parameter training, and outputting the door state monitoring model.
According to an aspect of the invention, further comprising:
s23, obtaining a test sample and inputting the test sample into the door state monitoring model, judging the output precision of the door state monitoring model according to the output of the door state monitoring model, outputting the door state monitoring model if the output precision meets the requirement, and otherwise, re-executing the steps S21 to S23.
According to an aspect of the present invention, in step S22, the labeled training samples are compressed and gray-scale converted, and then input to the convolutional neural network for parameter training of the model.
According to one aspect of the invention, the elevator door open and close states in the training sample comprise closed door, half open and full open;
the distribution proportion of the opening and closing states of the elevator door in the test sample is consistent with that of the training sample.
According to the scheme of the invention, the running time sequence of the elevator door can be conveniently obtained by obtaining the video stream of the running of the elevator, the whole running process of the elevator door is simply and effectively simplified, meanwhile, the accuracy of judging the running state of the whole elevator door is also ensured, and the effect of quick and accurate detection is achieved.
According to the scheme of the invention, the running state of the elevator door is judged through the elevator door running time sequence obtained from the obtained video stream, and the requirement of judging the time node of starting running of the elevator door is directly and effectively eliminated, so that the detection efficiency of the method is higher.
According to one scheme of the invention, the method can be realized without other obvious marks on the elevator door, and particularly, the method can still keep good detection precision under the condition of dark light, and is beneficial to realizing the real-time detection of the state of the elevator door.
According to one scheme of the invention, the opening and closing states of the door are identified based on the deep convolutional neural network, after the model is trained by collecting samples, the model is independently modeled without pre-collecting samples for each elevator, the unified model can be applied to all elevators, the modeling process is simple and convenient, and the precision is high. In addition, when the tags are set, the tags of the door closing type, the half opening type and the door opening type are established, and the newly acquired pictures can be easily distinguished to which category the newly acquired pictures belong. Therefore, the situation that the elevator is opened or is half opened in a traditional mode is effectively avoided, and the detection efficiency is greatly improved.
According to one scheme of the invention, the door fault types which can be identified by the invention are more, and more specifically, when the elevator is abnormally closed, the more detailed abnormity, such as the abnormity of full opening of the door for a long time or the abnormity of half opening of the door for a long time or the abnormity of difficulty in closing the door, can be given in more detail, and the more detailed abnormity can help maintenance personnel to more clearly position the elevator door fault problem.
Drawings
Fig. 1 is a block diagram schematically illustrating the steps of a method for real-time detection of elevator door faults according to one embodiment of the present invention.
Fig. 2 is a flow chart schematically illustrating a method for real-time detection of elevator door faults according to an embodiment of the present invention;
fig. 3 is a fully open view schematically illustrating an elevator door according to an embodiment of the present invention;
fig. 4 is a half-open view schematically showing an elevator door according to an embodiment of the present invention.
Fig. 5 is a view schematically showing the closing of an elevator door according to an embodiment of the present invention;
fig. 6 is a flow chart schematically showing the generation of a door state monitoring model according to an embodiment of the present invention.
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
In describing embodiments of the present invention, the terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like are used in an orientation or positional relationship that is based on the orientation or positional relationship shown in the associated drawings, which is for convenience and simplicity of description only, and does not indicate or imply that the referenced device or element must have a particular orientation, be constructed and operated in a particular orientation, and thus, the above-described terms should not be construed as limiting the present invention.
The present invention is described in detail below with reference to the drawings and the specific embodiments, which are not repeated herein, but the embodiments of the present invention are not limited to the following embodiments.
Referring to fig. 1 and 2, according to an embodiment of the present invention, a method for detecting elevator door faults in real time based on machine vision includes:
s1, collecting a video stream containing an elevator door image, and intercepting the elevator door image in the video stream according to a time sequence according to a time interval of a fixed threshold value;
s2, obtaining an elevator door image, judging the state of an elevator door in the elevator door image based on a door state monitoring model, and obtaining an elevator door time sequence of elevator door operation;
and S3, detecting the fault of the elevator door according to the time sequence of the elevator door.
According to an embodiment of the present invention, in step S1, a corresponding elevator door image is obtained by capturing a screenshot according to a fixed threshold time interval according to a video stream collected by a camera.
Referring to fig. 3, 4 and 5, according to an embodiment of the present invention, the state of the elevator door includes: full open (see fig. 3), half open (see fig. 4) and closed (see fig. 5). In the present embodiment, it is known from the analysis of the opening and closing process of the elevator door that the elevator door always repeats the cycle from closing to opening or from opening to closing, and there is an intermediate process that does not open or close, which is a type of process that is difficult to avoid in the detection of the door state. Therefore, this intermediate state is named half-open for the following door fault detection categories to be clearer. Therefore, the types of the states of the doors are determined to be three, i.e., closed door, half open, and full open.
According to one embodiment of the invention, the elevator door timing refers to the door state of the elevator at each time, i.e. the elevator is in a closed, half-open or full door state. In this embodiment, the elevator door timing sequence detection includes two main components, namely, offline training of the door state detection model and online real-time detection of the door state timing sequence.
According to one embodiment of the invention, an elevator door fault comprises: the door is repeatedly opened and closed, the door is difficult to open and close, the door opening and closing speed is abnormal, the door is fully opened for a long time, the door is half opened for a long time, and the ladder is driven when the door is opened.
In the present embodiment, the repeated opening and closing of the door means that the elevator door repeats a cycle of "closing the door-half opening the door" a plurality of times when the elevator is stationary. Therefore, in step S3, in the step of detecting the elevator door fault according to the elevator door timing sequence, if the elevator door has multiple cycles of "closed door, half open, full open" in the elevator door timing sequence obtained according to the video stream, it is determined that the elevator door fault is a repeatedly opened and closed door. The number of cycles for determining whether the elevator door is opened or closed repeatedly may be set, for example, 2 times, 3 times, etc.
In the present embodiment, the door opening and closing difficulty includes two scenarios, i.e., multiple attempts to open the door and multiple attempts to close the door. The repeated door opening attempts mean that when the elevator is static, the elevator door repeats the cycle of door closing and half opening for a plurality of times, but the elevator is difficult to open the door completely; the multiple-time door closing attempt means that when the elevator is static, the door repeats a cycle of opening the door and half opening the door for multiple times, but the elevator cannot completely close the door. Therefore, in step S3, in the step of detecting the failure of the elevator door according to the elevator door timing sequence, if the elevator door has multiple cycles of "closed door-half open" or multiple cycles of "fully open-half open" in the elevator door timing sequence obtained according to the video stream, it is determined that the failure of the elevator door is difficult to open or close the door. The number of cycles for determining whether the elevator door is difficult to open or close may be set, for example, 2 times, 3 times, or the like.
In this embodiment, the abnormal door opening and closing speed includes two scenarios, i.e., the door is opened and closed slowly and the door is opened and closed quickly. The slow door opening and closing means that the time consumed by the elevator from closing to full opening is long and is far longer than the normal door opening and closing time of the elevator; the fast door opening and closing means that the time for the elevator to close the door from the full door is short, and even the elevator is accompanied with a collision phenomenon. Therefore, in step S3, in the step of detecting the failure of the elevator door according to the elevator door timing sequence, if the consumed time of "closing door-full opening" of the elevator door exceeds the preset threshold or the consumed time of "full opening-closing" of the elevator door is less than the preset threshold in the elevator door timing sequence obtained according to the video stream, it is determined that the failure of the elevator door is abnormal in door opening and closing speed. The preset threshold value for determining whether the elevator door is abnormally opened or closed may be set, for example, the preset threshold value for opening or closing the door very slowly may be set to 7 seconds, and the preset threshold value for opening or closing the door very quickly may be set to 1 second.
In the present embodiment, the long-time non-closing of the door means that the door remains in a non-closed state for a long time when the elevator is stationary, and includes two scenes of a long-time full-open door and a long-time half-open door. When the elevator is static, the door is normally closed, even if people go in and out of the elevator or block the door, the door opening duration cannot be long, and therefore once the elevator is kept in the door opening state for a long time, the elevator door may have certain abnormity. The long-time full-open door is that the detection door keeps the state of full-open for a long time, and it needs to be noted that the elevator (welcome mode and the like) with the door fully-open when the elevator is static does not need to be considered through special setting, and the situation can be eliminated by obtaining the setting of the elevator at the moment in the detection process (for example, when the elevator is recorded, the information can be recorded, such as the welcome mode and the like, so that when the door is identified to be fully-open for a long time, a blacklist can be used for filtering the situation). In addition, the time required for the elevator door to be opened from the closed state to the fully opened state or from the fully opened state to the closed state is about 2s, and therefore, once the half-opened state is maintained for a long time, the elevator door may have some abnormality. The long-time half-open state of the door is detected to be kept for a long time.
Therefore, in step S3, in the step of detecting the failure of the elevator door according to the elevator door time sequence, if the duration of the fully opened state of the elevator door in the elevator door time sequence obtained according to the video stream is greater than the preset threshold, it is determined that the failure of the elevator door is fully opened for a long time; among them, a preset threshold value for determining whether the elevator door is fully opened for a long time may be set, and for example, the preset threshold value may be set to 60 seconds.
And if the half-open state duration time of the elevator door in the elevator door time sequence acquired according to the video stream is greater than a preset threshold value, judging that the elevator door fault is long-time half-open of the elevator door. Among them, a preset threshold for determining whether the elevator door is half-opened for a long time may be set, and for example, the preset threshold may be set to 60 seconds.
In the present embodiment, the door-open elevator-walking refers to the situation that the elevator door is fully opened or half opened and the elevator is moving, and the foregoing has described in detail an algorithm for detecting that the elevator door is opened for a long time and half opened for a long time, so the key point of the door-open elevator-walking is to detect whether the elevator is moving. The motion state data of the elevator is combined, so that the motion stillness of the elevator can be judged, and the phenomenon that the elevator is opened and walks can be detected. Therefore, in step S3, in the step of detecting the elevator door fault according to the elevator door time sequence, if the elevator door is fully opened or half opened in the elevator door time sequence obtained according to the video stream and the elevator is in a motion state, it is determined that the elevator door fault is open door and elevator is going. It should be pointed out that, whether the in-process that detects the elevator and move, acceleration data that can be based on the acceleration sensor collection of thing networking, the acceleration obtains speed data after the filtering integral, if the speed absolute value is greater than 0, then can think that the elevator has taken place the operation, and when the elevator door opened or half opened, the elevator had taken place the operation, can think that the ladder is walked to opening the door.
As shown in fig. 6, according to an embodiment of the present invention, a door state monitoring model is generated based on the steps including:
s21, acquiring an image containing the opening and closing state of the elevator door as a training sample; in the present embodiment, the open/close state of the elevator door in the training sample (i.e., the learning target) includes closed door, half open, and full open, see fig. 3 to 5;
and S22, after the training samples are labeled, inputting the training samples into a convolutional neural network for parameter training of the model, and outputting a gate state monitoring model. In this embodiment, the training samples are labeled by manual labeling, and then the labeled training samples are processed, such as picture compression and gray level conversion. And inputting the processed training sample into a convolutional neural network for parameter training of the model.
According to an embodiment of the present invention, further comprising:
s23, obtaining a test sample and inputting the test sample into the door state monitoring model, judging the output precision of the door state monitoring model according to the output of the door state monitoring model, outputting the door state monitoring model if the output precision meets the requirement, and otherwise, re-executing the steps S21 to S23. In the embodiment, when the model training is completed, the precision of the model is evaluated, in the embodiment, the adopted test sample set is not repeated with the training sample set, and the distribution proportion of the opening and closing states of the elevator door in the test sample is consistent with that of the training sample. In the embodiment, the model can be applied to the online detection environment of the elevator door fault only when the detection precision of the test sample set meets the requirement, and if the detection precision does not meet the requirement, the labeling of the sample and the model training need to be repeated.
According to one embodiment of the invention, the door condition monitoring model is a deep learning machine vision model.
The foregoing is merely exemplary of particular aspects of the present invention and devices and structures not specifically described herein are understood to be those of ordinary skill in the art and are intended to be implemented in such conventional ways.
The above description is only one embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (11)
1. A real-time elevator door fault detection method based on machine vision comprises the following steps:
s1, collecting a video stream containing an elevator door image, and intercepting the elevator door image in the video stream according to a time sequence according to a time interval of a fixed threshold value;
s2, obtaining the elevator door image, judging the state of an elevator door in the elevator door image based on a door state monitoring model, and obtaining the elevator door time sequence of the operation of the elevator door;
and S3, detecting the fault of the elevator door according to the elevator door time sequence.
2. The method of real-time elevator door fault detection according to claim 1, wherein the status of the elevator door includes: full open, half open and close;
the elevator door failure comprises: the door is repeatedly opened and closed, the door is difficult to open and close, the door opening and closing speed is abnormal, the door is fully opened for a long time, the door is half opened for a long time, and the ladder is driven when the door is opened.
3. The method of claim 2, wherein in step S3, in the step of detecting the failure of the elevator door according to the elevator door timing sequence, if the elevator door has multiple cycles of "closed door, half open, full open" in the elevator door timing sequence obtained according to the video stream, the failure of the elevator door is determined to be the repeated opening and closing of the door.
4. The method of claim 3, wherein in step S3, in the step of detecting the failure of the elevator door according to the elevator door timing sequence, if the elevator door has multiple cycles of "closed door-half open" or multiple cycles of "fully open-half open" in the elevator door timing sequence obtained according to the video stream, the failure of the elevator door is determined to be difficult to open or close the door.
5. The method of claim 4, wherein in step S3, in the step of detecting the failure of the elevator door according to the elevator door timing sequence, if the time consumed for the door to be closed-fully opened exceeds a preset threshold or the time consumed for the door to be fully opened-closed is lower than the preset threshold in the elevator door timing sequence obtained according to the video stream, the failure of the elevator door is determined as abnormal door opening and closing speed.
6. The method for detecting the fault of the elevator door in real time as claimed in claim 5, wherein in the step of detecting the fault of the elevator door according to the elevator door timing sequence in step S3, if the duration of the fully opened state of the elevator door in the elevator door timing sequence obtained according to the video stream is greater than a preset threshold, the fault of the elevator door is determined to be fully opened for a long time;
and if the half-open state duration time of the elevator door in the elevator door time sequence acquired according to the video stream is greater than a preset threshold value, judging that the elevator door fault is long-time half-open of the elevator door.
7. The method of claim 6, wherein in step S3, in the step of detecting the elevator door fault according to the elevator door timing sequence, if the elevator door is fully opened or half opened in the elevator door timing sequence obtained according to the video stream and the elevator is in motion, the elevator door fault is determined as open door and elevator moving.
8. The method for real-time detection of elevator door faults according to any one of claims 1 to 7, characterized in that the door state monitoring model is generated based on the following steps, including:
s21, acquiring an image containing the opening and closing state of the elevator door as a training sample;
and S22, after the training samples are labeled, inputting the training samples into a convolutional neural network for model parameter training, and outputting the door state monitoring model.
9. The method of real-time detection of elevator door faults according to claim 8, further comprising:
s23, obtaining a test sample and inputting the test sample into the door state monitoring model, judging the output precision of the door state monitoring model according to the output of the door state monitoring model, outputting the door state monitoring model if the output precision meets the requirement, and otherwise, re-executing the steps S21 to S23.
10. The method of claim 9, wherein in step S22, the labeled training samples are compressed and gray-scale converted and then input to the convolutional neural network for parameter training of the model.
11. The method of claim 10, wherein the elevator door open/close states in the training samples include closed door, half open, and full open;
the distribution proportion of the opening and closing states of the elevator door in the test sample is consistent with that of the training sample.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011123569.9A CN112347862A (en) | 2020-10-20 | 2020-10-20 | Elevator door fault real-time detection method based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011123569.9A CN112347862A (en) | 2020-10-20 | 2020-10-20 | Elevator door fault real-time detection method based on machine vision |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112347862A true CN112347862A (en) | 2021-02-09 |
Family
ID=74361014
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011123569.9A Pending CN112347862A (en) | 2020-10-20 | 2020-10-20 | Elevator door fault real-time detection method based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112347862A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112801072A (en) * | 2021-04-14 | 2021-05-14 | 浙江大学 | Elevator non-flat-layer door opening fault recognition device and method based on computer vision |
| CN113516179A (en) * | 2021-06-24 | 2021-10-19 | 北京航空航天大学 | Method and system for identifying water leakage performance of underground infrastructure |
| CN113581956A (en) * | 2021-07-19 | 2021-11-02 | 浙江新再灵科技股份有限公司 | Elevator noise level monitoring method and system based on audio signal |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106204659A (en) * | 2016-07-26 | 2016-12-07 | 浙江捷尚视觉科技股份有限公司 | Elevator switch door detection method based on light stream |
| CN106219367A (en) * | 2016-08-05 | 2016-12-14 | 沈阳聚德视频技术有限公司 | A kind of elevator O&M based on intelligent vision light curtain monitoring method |
| CN108584588A (en) * | 2017-12-31 | 2018-09-28 | 浙江工业大学 | A kind of tor door faults detection method based on extensive flow data |
| CN109775486A (en) * | 2019-03-13 | 2019-05-21 | 上海臻颖智能科技有限公司 | A kind of elevator operation monitoring system of view-based access control model intelligence |
| CN109896386A (en) * | 2019-03-08 | 2019-06-18 | 浙江新再灵科技股份有限公司 | A kind of detection method and system that the elevator door based on computer vision technique switchs repeatedly |
| CN110002302A (en) * | 2018-08-09 | 2019-07-12 | 浙江新再灵科技股份有限公司 | A kind of elevator switch door detection system and method based on deep learning |
| CN110127479A (en) * | 2019-04-17 | 2019-08-16 | 浙江工业大学 | A kind of elevator door switch method for detecting abnormality based on video analysis |
| CN110363144A (en) * | 2019-07-16 | 2019-10-22 | 中国民航科学技术研究院 | A kind of aircraft door switch state detecting system and method based on image processing techniques |
| CN110589647A (en) * | 2019-08-13 | 2019-12-20 | 福建工程学院 | Method for real-time fault detection and prediction of elevator door through monitoring |
| CN111453578A (en) * | 2019-01-18 | 2020-07-28 | 株式会社日立大厦系统 | Door abnormality detection system and elevator system |
| CN111731962A (en) * | 2020-06-29 | 2020-10-02 | 浙江新再灵科技股份有限公司 | Opening and closing fault detection method and detection early warning system for door |
| CN111731960A (en) * | 2020-06-22 | 2020-10-02 | 浙江新再灵科技股份有限公司 | Elevator door opening and closing state detection method |
-
2020
- 2020-10-20 CN CN202011123569.9A patent/CN112347862A/en active Pending
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106204659A (en) * | 2016-07-26 | 2016-12-07 | 浙江捷尚视觉科技股份有限公司 | Elevator switch door detection method based on light stream |
| CN106219367A (en) * | 2016-08-05 | 2016-12-14 | 沈阳聚德视频技术有限公司 | A kind of elevator O&M based on intelligent vision light curtain monitoring method |
| CN108584588A (en) * | 2017-12-31 | 2018-09-28 | 浙江工业大学 | A kind of tor door faults detection method based on extensive flow data |
| CN110002302A (en) * | 2018-08-09 | 2019-07-12 | 浙江新再灵科技股份有限公司 | A kind of elevator switch door detection system and method based on deep learning |
| CN111453578A (en) * | 2019-01-18 | 2020-07-28 | 株式会社日立大厦系统 | Door abnormality detection system and elevator system |
| CN109896386A (en) * | 2019-03-08 | 2019-06-18 | 浙江新再灵科技股份有限公司 | A kind of detection method and system that the elevator door based on computer vision technique switchs repeatedly |
| CN109775486A (en) * | 2019-03-13 | 2019-05-21 | 上海臻颖智能科技有限公司 | A kind of elevator operation monitoring system of view-based access control model intelligence |
| CN110127479A (en) * | 2019-04-17 | 2019-08-16 | 浙江工业大学 | A kind of elevator door switch method for detecting abnormality based on video analysis |
| CN110363144A (en) * | 2019-07-16 | 2019-10-22 | 中国民航科学技术研究院 | A kind of aircraft door switch state detecting system and method based on image processing techniques |
| CN110589647A (en) * | 2019-08-13 | 2019-12-20 | 福建工程学院 | Method for real-time fault detection and prediction of elevator door through monitoring |
| CN111731960A (en) * | 2020-06-22 | 2020-10-02 | 浙江新再灵科技股份有限公司 | Elevator door opening and closing state detection method |
| CN111731962A (en) * | 2020-06-29 | 2020-10-02 | 浙江新再灵科技股份有限公司 | Opening and closing fault detection method and detection early warning system for door |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112801072A (en) * | 2021-04-14 | 2021-05-14 | 浙江大学 | Elevator non-flat-layer door opening fault recognition device and method based on computer vision |
| CN113516179A (en) * | 2021-06-24 | 2021-10-19 | 北京航空航天大学 | Method and system for identifying water leakage performance of underground infrastructure |
| CN113581956A (en) * | 2021-07-19 | 2021-11-02 | 浙江新再灵科技股份有限公司 | Elevator noise level monitoring method and system based on audio signal |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112347862A (en) | Elevator door fault real-time detection method based on machine vision | |
| CN106219367B (en) | A kind of elevator O＆M monitoring method based on intelligent vision light curtain | |
| CN107679471B (en) | Indoor personnel air post detection method based on video monitoring platform | |
| US11524876B2 (en) | Method and monitoring device for monitoring an operation of an elevator door arrangement | |
| US11760605B2 (en) | Elevator door monitoring system, elevator system and elevator door monitoring method | |
| CN109867186B (en) | Elevator trapping detection method and system based on intelligent video analysis technology | |
| CN105452139A (en) | Monitoring system of a lift installation | |
| US20010045327A1 (en) | Elevator door control device | |
| CN110002302A (en) | A kind of elevator switch door detection system and method based on deep learning | |
| CN109896386B (en) | Method and system for detecting repeated opening and closing of elevator door based on computer vision technology | |
| CN109409315B (en) | Method and system for detecting remnants in panel area of ATM (automatic Teller machine) | |
| CN108178035A (en) | A kind of elevator cage door state monitoring apparatus and monitoring method | |
| CN112357707B (en) | Elevator detection method and device, robot and storage medium | |
| CN110790101A (en) | Elevator trapping false alarm identification method based on big data analysis | |
| CN110589647A (en) | Method for real-time fault detection and prediction of elevator door through monitoring | |
| CN111731960B (en) | Elevator door opening and closing state detection method | |
| CN101996307A (en) | Intelligent video human body identification method | |
| CN108584586A (en) | A kind of elevator door state line detection method based on video image | |
| US20220083782A1 (en) | Item monitoring for doorbell cameras | |
| CN108584588B (en) | Elevator door fault detection method based on large-scale flow data | |
| CN108776452B (en) | Special equipment field maintenance monitoring method and system | |
| CN110723621A (en) | Device and method for detecting smoking in elevator car based on deep neural network | |
| CN114436087B (en) | Deep learning-based elevator passenger door-pulling detection method and system | |
| CN112127718B (en) | Method for detecting running state of sliding door lock of platform door system | |
| CN112560650A (en) | Elevator door label extraction and monitoring picture abnormity detection method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |