CN116863409B - Intelligent elevator safety management method and system based on cloud platform - Google Patents
Intelligent elevator safety management method and system based on cloud platform Download PDFInfo
- Publication number
- CN116863409B CN116863409B CN202311135363.1A CN202311135363A CN116863409B CN 116863409 B CN116863409 B CN 116863409B CN 202311135363 A CN202311135363 A CN 202311135363A CN 116863409 B CN116863409 B CN 116863409B
- Authority
- CN
- China
- Prior art keywords
- identification object
- identification
- hidden danger
- probability
- dynamic
- 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.)
- Active
Links
- 238000007726 management method Methods 0.000 title claims abstract description 36
- 238000012544 monitoring process Methods 0.000 claims abstract description 52
- 230000011218 segmentation Effects 0.000 claims abstract description 37
- 231100001267 hazard identification Toxicity 0.000 claims abstract description 22
- 238000000034 method Methods 0.000 claims abstract description 19
- 238000004364 calculation method Methods 0.000 claims description 17
- 238000012549 training Methods 0.000 claims description 10
- 230000003213 activating effect Effects 0.000 claims description 5
- 238000013507 mapping Methods 0.000 claims description 5
- 230000003068 static effect Effects 0.000 claims description 5
- 230000004913 activation Effects 0.000 claims description 4
- 230000003542 behavioural effect Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 5
- 230000008859 change Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 241000282472 Canis lupus familiaris Species 0.000 description 3
- 238000012935 Averaging Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/26—Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/94—Hardware or software architectures specially adapted for image or video understanding
- G06V10/95—Hardware or software architectures specially adapted for image or video understanding structured as a network, e.g. client-server architectures
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02B—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO BUILDINGS, e.g. HOUSING, HOUSE APPLIANCES OR RELATED END-USER APPLICATIONS
- Y02B50/00—Energy efficient technologies in elevators, escalators and moving walkways, e.g. energy saving or recuperation technologies
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Artificial Intelligence (AREA)
- Image Analysis (AREA)
Abstract
The application provides a cloud platform-based intelligent elevator safety management method and a cloud platform-based intelligent elevator safety management system, which relate to the technical field of intelligent identification, wherein the intelligent elevator safety management method comprises the following steps: according to the method, object segmentation is carried out on multi-frame images in a ladder according to a target segmentation model embedded in a cloud platform, a first identification object and a second identification object are determined to be input into a potential safety hazard identification model, a first potential hazard probability is obtained, when the first potential hazard probability is larger than a first preset potential hazard probability, a tracking monitoring instruction is activated to track and monitor the first identification object and the second identification object, feature identification is carried out on the obtained tracking multi-frame images, a first dynamic potential hazard probability is output, and when the first dynamic potential hazard probability is larger than a second preset potential hazard probability, first safety early warning information is generated.
Description
Technical Field
The application relates to the technical field of intelligent identification, in particular to a cloud platform-based intelligent elevator safety management method and system.
Background
With the rapid development of urban construction, the high-rise building is more and more, the elevator becomes the standard of the high-rise building, a plurality of potential safety hazards exist when people travel conveniently, people can carry pets, cargoes and the like into the elevator at will in the elevator, and as a plurality of uncertain factors exist in the elevator between the pets and the cargoes, the potential safety hazards exist for the people, for example, the dog rope is rolled into an elevator gap, the cargoes scatter to block the potential safety hazards of an elevator inner door and the like, and the technical problem of low safety in the elevator is caused due to the lack of control over the safety in the elevator in the prior art, the accurate control over the safety of the elevator based on a cloud platform is realized, and the safety in the elevator is improved.
Disclosure of Invention
The application provides a cloud platform-based intelligent elevator safety management method and system, which are used for solving the technical problem of low safety in an elevator caused by lack of management and control of safety in the elevator in the prior art.
In view of the above problems, the application provides a cloud platform-based intelligent elevator safety management method and system.
In a first aspect, the application provides a cloud platform-based intelligent elevator safety management method, which comprises the following steps: collecting multi-frame images in a ladder in a target elevator, and uploading the multi-frame images in the ladder to a cloud platform, wherein a target segmentation model is embedded in the cloud platform and is used for segmenting objects in the multi-frame images in the ladder; object segmentation is carried out on the multi-frame images in the ladder according to the target segmentation model embedded in the cloud platform, and a first identification object and a second identification object are determined, wherein the second identification object is a carrying object of the first identification object; inputting the characteristic information of the first identification object and the second identification object into a potential safety hazard identification model, and calculating the potential safety hazard probability of the first identification object and the second identification object according to the potential safety hazard identification model to obtain a first potential safety hazard probability; when the first hidden danger probability is larger than a first preset hidden danger probability, activating a tracking monitoring instruction; tracking and monitoring the first identification object and the second identification object according to the tracking and monitoring instruction to acquire a tracking multi-frame image; and carrying out feature recognition according to the tracking multi-frame image, outputting a first dynamic hidden danger probability, and generating first safety early warning information when the first dynamic hidden danger probability is larger than a second preset hidden danger probability.
In a second aspect, the present application provides a cloud platform-based intelligent elevator safety management system, the system comprising: the system comprises an image acquisition module, a cloud platform and a target segmentation module, wherein the image acquisition module is used for acquiring multi-frame images in a ladder in a target elevator and uploading the multi-frame images in the ladder to the cloud platform, wherein the cloud platform is embedded with the target segmentation model, and the target segmentation model is used for segmenting objects in the multi-frame images in the ladder; the object segmentation module is used for carrying out object segmentation on the intra-ladder multi-frame image according to the target segmentation model embedded in the cloud platform, and determining a first identification object and a second identification object, wherein the second identification object is a carrying object of the first identification object; the probability calculation module is used for inputting the characteristic information of the first identification object and the second identification object into a potential safety hazard identification model, and carrying out potential safety hazard probability calculation on the first identification object and the second identification object according to the potential safety hazard identification model to obtain first potential safety hazard probability; the instruction activation module is used for activating a tracking monitoring instruction when the first hidden danger probability is larger than a first preset hidden danger probability; the tracking monitoring module is used for tracking and monitoring the first identification object and the second identification object according to the tracking and monitoring instruction to acquire a tracking multi-frame image; the feature recognition module is used for carrying out feature recognition according to the tracking multi-frame images, outputting first dynamic hidden danger probability, and generating first safety early warning information when the first dynamic hidden danger probability is larger than second preset hidden danger probability.
One or more technical schemes provided by the application have at least the following technical effects or advantages:
the intelligent elevator safety management method and system based on the cloud platform provided by the application relate to the technical field of intelligent identification, solve the technical problem of low safety in an elevator caused by lack of control over safety in the elevator in the prior art, realize accurate control over safety of the elevator based on the cloud platform and improve safety in the elevator.
Drawings
Fig. 1 is a schematic flow diagram of a smart elevator safety management method based on a cloud platform;
fig. 2 is a schematic diagram of a first hidden danger probability flow for obtaining the degree of potential safety hazards of a first identification object and a second identification object in a cloud platform-based intelligent elevator safety management method;
fig. 3 is a schematic diagram of a process of acquiring a first hidden danger probability in a cloud platform-based intelligent elevator safety management method;
fig. 4 is a schematic flow chart of a pre-trained safety hidden danger identification model obtained in a cloud platform-based intelligent elevator safety management method;
fig. 5 is a schematic diagram of a process flow of outputting a first dynamic hidden danger probability in the intelligent elevator safety management method based on the cloud platform;
fig. 6 is a schematic structural diagram of a smart elevator safety management system based on a cloud platform.
Reference numerals illustrate: the system comprises an image acquisition module 1, an object segmentation module 2, a probability calculation module 3, an instruction activation module 4, a tracking and monitoring module 5 and a feature identification module 6.
Detailed Description
The application provides a cloud platform-based intelligent elevator safety management method and a cloud platform-based intelligent elevator safety management system, which are used for solving the technical problem that the safety in an elevator is low due to lack of control over the safety in the elevator in the prior art.
Example 1
As shown in fig. 1, an embodiment of the present application provides a cloud platform-based intelligent elevator security management method, which includes:
step S100: collecting multi-frame images in a ladder in a target elevator, and uploading the multi-frame images in the ladder to a cloud platform, wherein a target segmentation model is embedded in the cloud platform and is used for segmenting objects in the multi-frame images in the ladder;
specifically, the intelligent elevator safety management method based on the cloud platform is applied to an intelligent elevator safety management system based on the cloud platform, the intelligent elevator safety management system based on the cloud platform is in communication connection with an image acquisition device, and the image acquisition device is used for acquiring image parameters existing in an elevator.
In order to ensure the safety in the target elevator in the later stage, firstly, the real-time image in the target elevator needs to be acquired, namely, the image acquisition equipment acquires multiple frames of images in the target elevator, wherein the multiple frames of images are image frames of images which are acquired and recorded by the image acquisition equipment in the target elevator and are not empty in the target elevator, further, the multiple frames of images in the extracted target elevator are uploaded to a cloud platform, the cloud platform is a cloud computing platform based on hardware resources and software resources of a computer and used for providing computing, network and storage capacity, a target segmentation model is embedded in the cloud platform, and the target segmentation model is used for segmenting a plurality of objects contained in the multiple frames of images in the target elevator, and the segmentation principle of the target segmentation model is to segment the acquired multiple frames of images into a plurality of subareas with similar color or texture characteristics and enable the images to correspond to different objects or different parts of the objects, so that the intelligent elevator is safely managed based on the cloud platform for the later stage.
Step S200: object segmentation is carried out on the multi-frame images in the ladder according to the target segmentation model embedded in the cloud platform, and a first identification object and a second identification object are determined, wherein the second identification object is a carrying object of the first identification object;
specifically, a target segmentation model embedded in a cloud platform is used as a segmentation tool, object segmentation is carried out on a multi-frame image in an obtained target elevator, namely an I-frame image in the multi-frame image is extracted, the I-frame image is an image of an object carried by people in the target elevator, the extracted I-frame image is further segmented according to a preset segmentation standard, wherein the obtained preset segmentation standard is preset by a related technician according to pixel blocks in the I-frame image, and a target I-frame image block is determined, so that discrete cosine change coefficients are divided into a first discrete cosine change coefficient and a second discrete cosine change coefficient.
Extracting a first direct current coefficient and a first alternating current coefficient in a first discrete cosine transform coefficient, and calculating to obtain the first characteristic value according to the first direct current coefficient and the first alternating current coefficient, wherein a calculation formula is as follows:
；
refers to the first characteristic value,/->Refers to the first DC coefficient, ">Refers to the first AC coefficient, n refers to the first I frame picture,/and>refers to the +.>The sub-blocks, a refers to the influence factor of the first direct current coefficient on the first eigenvalue, and b refers to the influence factor of the first alternating current coefficient on the first eigenvalue.
Further, extracting a second direct current coefficient and a second alternating current coefficient in the second discrete cosine transform coefficient;
and calculating a second characteristic value according to a second direct current coefficient and a second alternating current coefficient, wherein the calculation formula is as follows:
；
refers to the second characteristic value,/->Refers to the second DC coefficient, +.>Refers to the second alternating current coefficient, n+1 refers to the second I frame picture, +.>Refers to the +.>And c refers to an influence factor of the second direct current coefficient on the second characteristic value, and d refers to an influence factor of the second alternating current coefficient on the second characteristic value.
And finally calculating the characteristic value of the adjacent I frame image, namely, the characteristic value of the I frame image is white rectangular pixels and subtracting black rectangular pixels, reflecting the gray change condition of the image, distinguishing and acquiring a first identification object and a second identification object in a target elevator through the characteristic value of the I frame image, wherein the first identification object in the target elevator is a crowd taking the elevator, the second identification object is a carrying object of the first identification object, and the carrying object can be objects such as a pet dog, a large box, goods, a handbag and the like, so as to ensure the safety management of the intelligent elevator based on a cloud platform.
Step S300: inputting the characteristic information of the first identification object and the second identification object into a potential safety hazard identification model, and calculating the potential safety hazard probability of the first identification object and the second identification object according to the potential safety hazard identification model to obtain a first potential safety hazard probability;
further, as shown in fig. 2, step S300 of the present application further includes:
step S310: acquiring characteristic information of the first identification object, wherein the characteristic information comprises body type characteristic information, sex characteristic information and age characteristic information;
step S320: acquiring characteristic information of the second identification object, wherein the characteristic information comprises carried volume characteristic information, static and dynamic characteristic information and constraint characteristic information;
step S330: and taking the characteristic information of the first identification object and the characteristic information of the second identification object as training data, inputting the training data into the pre-trained potential safety hazard identification model for calculation, and obtaining a first potential hazard probability for identifying the potential safety hazard degrees of the first identification object and the second identification object.
Further, as shown in fig. 3, step S300 of the present application further includes:
step S340: when the number of the carrying objects of the first identification object is multiple, obtaining multiple carrying objects;
step S350: establishing a one-to-many mapping relation between the first identification object and the plurality of carrying objects, and outputting a plurality of hidden danger probabilities based on the plurality of carrying objects according to the hidden danger identification model;
step S360: and integrating the hidden danger probabilities and outputting the first hidden danger probability.
Specifically, in order to accurately identify the potential safety hazards existing in the first identification object and the second identification object in the target elevator, therefore, the characteristic information of the first identification object and the characteristic information of the second identification object need to be input into a potential safety hazard identification model constructed by the following steps, further, the potential safety hazard probability calculation is performed on the first identification object and the second identification object according to the potential safety hazard identification model, the characteristic information of the first identification object is acquired, the characteristic information can comprise body type characteristic information, sex characteristic information and age characteristic information of the first identification object, the body type characteristic information can comprise slimming type, shortness type, uniformity type and the like of a human body, the sex characteristic information can comprise male information and female information, the person taking the target elevator can be divided into a child, a young year, a middle-aged year and an elderly person according to the age characteristic information, the characteristic information of the second identification object can be synchronously acquired, the characteristic information can comprise carried volume characteristic information, static characteristic information and constraint characteristic information, the volume characteristic information refers to the volume occupied by the second identification object in the target elevator, the volume occupied by the static characteristic information refers to the second identification object, if the second identification object is a corresponding dynamic characteristic is not present, and the second identification object is constrained according to the corresponding dynamic characteristic.
And then, characteristic information of the first identification object and characteristic information of the second identification object are used as training data, the training data is input into a pre-trained potential safety hazard identification model for calculation, namely, the degree of possible danger of the first identification object is judged according to the body type characteristic information, the sex characteristic information and the age characteristic information of the first identification object, the degree of possible danger of the second identification object is judged according to the volume characteristic information, the static characteristic information and the constraint characteristic information of the second identification object, the degree of possible danger of the second identification object is judged according to the judgment basis, and the average value of the probability of possible danger of the first identification object and the probability of possible danger of the second identification object is taken as the potential safety hazard degree of the first identification object and the second identification object, and meanwhile, the potential safety hazard probability of the first identification object is obtained.
The obtained first hidden danger probability is the probability when the first identification object only carries one second identification object, when the first identification object carries a plurality of the first identification objects, a plurality of hidden danger probability output corresponding to the plurality of the first identification objects is obtained, meanwhile, a one-to-many mapping relation is established between the first identification object and the plurality of the first identification objects, namely, a value is taken among the plurality of the first identification objects, the first identification object has one value only, the first identification object takes one value, the plurality of the first identification objects can have a plurality of values corresponding to the plurality of the first identification objects, the plurality of the first identification objects can take dogs and things, further, the weighting calculation of the hidden danger combination corresponding to the plurality of the first identification objects and the hidden danger corresponding to each of the first identification objects is carried according to the hidden danger identification model, and the weighting duty ratio of the hidden danger corresponding to each of the first identification object can be the influence coefficient of the first identification object: and the influence parameters after the weighting calculation process are the ratio of the influence coefficient of the first identification object to the coefficient, the influence coefficient of each carrying object to the coefficient, the final value of the hidden danger probabilities corresponding to the plurality of carrying objects carried by the first identification object is obtained according to the weighting calculation result, the hidden danger probabilities are finally integrated, the hidden danger probabilities are logically and organically concentrated, and the hidden danger probabilities are output as the first hidden danger probability, so that the basis for carrying out safety management tamping on the intelligent elevator based on the cloud platform is realized.
Further, as shown in fig. 4, step S300 of the present application further includes:
step S370: acquiring an elevator sample data set based on the cloud platform, wherein the elevator sample data set comprises an accident occurrence object and an accident occurrence probability;
step S380: splitting accident occurrence objects in the elevator sample data set to obtain split objects, wherein each split object comprises object characteristic information;
step S390: and performing model training according to the object characteristic information of the split object and the accident occurrence probability to obtain a pre-trained potential safety hazard identification model.
In particular, in order to ensure the accuracy of the obtained first hidden danger probability, a cloud platform is required to be used as a data acquisition source to extract a plurality of sample data of a target elevator, the plurality of sample data of the target elevator can comprise an accident source of an accident of the target elevator and the probability of the accident, the extracted plurality of sample data are integrated and summarized to be used as an elevator sample data set, the elevator sample data set comprises an accident occurrence object and the accident occurrence probability, further, the accident occurrence object in the elevator sample data set is split according to a first identification object and a second identification object, the split objects are obtained according to the split result, and each split object comprises characteristic information of the first identification object or characteristic information of the second identification object, finally, model training is carried out according to the characteristic information of the second identification object and the characteristic information of the first identification object, which are separated, namely the probability of accident that only people take the target elevator, the probability of accident that the combination of people and pets take the target elevator, the probability of accident that the combination of people and goods take the target elevator, the probability of accident that the combination of people and pets take the target elevator, the probability of accident that the combination of goods takes the target elevator, and the probability of accident that the combination of people and pets take the target elevator are calculated through the combination weighted calculation, so that the probability of accident occurrence in the four situations is larger is judged, a pre-trained potential safety hazard identification model is obtained on the basis, and the effect of pre-judging the safety management of the intelligent elevator based on a cloud platform is realized.
Step S400: when the first hidden danger probability is larger than a first preset hidden danger probability, activating a tracking monitoring instruction;
specifically, in order to ensure the safety in the target elevator, the first hidden danger probability output by the hidden danger identification model needs to be judged, when the first hidden danger probability is larger than the first preset hidden danger probability, a tracking monitoring instruction is activated, wherein the first preset hidden danger probability is preset by a relevant technician according to hidden danger probability data quantity in big data, the tracking monitoring instruction is used for dynamically tracking and monitoring a first identification object and a second identification object in the target elevator, and potential safety hazards possibly occurring in the first identification object and the second identification object are avoided to the greatest extent, so that the first hidden danger probability is used as reference data when the intelligent elevator is subjected to safety management based on a cloud platform in the later stage.
Step S500: tracking and monitoring the first identification object and the second identification object according to the tracking and monitoring instruction to acquire a tracking multi-frame image;
specifically, after the tracking monitoring instruction is activated, the first identification object and the second identification object in the target elevator are tracked and monitored dynamically through the tracking monitoring instruction, and meanwhile, based on the characteristic information of the first identification object and the second identification object and the corresponding potential safety hazard probability, and in the process of tracking the first identification object and the second identification, as the influence degree of the second identification object on the potential safety hazard is high, the second identification object needs to be tracked and monitored in a key way, and the method can be used for tracking and monitoring according to 4:6, tracking and monitoring the first identification object and the second identification object to dynamically track and distribute the first identification object and the second identification object, so that multi-frame dynamic images of the first identification object and the second identification object are respectively tracked, and the accuracy of safety management of the intelligent elevator based on the cloud platform is improved.
Step S600: and carrying out feature recognition according to the tracking multi-frame image, outputting a first dynamic hidden danger probability, and generating first safety early warning information when the first dynamic hidden danger probability is larger than a second preset hidden danger probability.
Specifically, the obtained multi-frame image is taken as basic data, the behavior characteristics and the dynamic characteristics of a first identification object and a second identification object contained in the multi-frame image are identified, the movement track and the movement amplitude of the first identification object and the movement amplitude of the second identification object in the elevator are respectively determined, the amplitude of the association influence of the first identification object and the second identification object when the first identification object and the second identification object move is determined according to the movement track, the movement amplitude and the amplitude of the association influence of the first identification object, the movement track and the movement amplitude of the second identification object and the amplitude of the association influence in the elevator at the moment are respectively in a proportional relation with the first dynamic hidden danger probability in the elevator at the moment, the first dynamic hidden danger probability is further judged, when the first dynamic hidden danger probability is larger than the second preset hidden danger probability, the second preset hidden danger probability is preset by relevant technicians according to the limit hidden danger data of the elevator in large data, the target elevator is regarded as the target elevator, and the target elevator is safe hidden danger probability, and the target elevator is accordingly safe, and the first potential safety hazard is synchronously warned, the first potential safety hazard probability is obtained, and the important potential safety hazard warning effect is obtained by the first potential safety warning platform is recorded and is achieved, and the important potential safety warning technology is achieved.
Further, as shown in fig. 5, step S600 of the present application further includes:
step S610: matching and identifying the behavior tracks of the first identification object and the second identification object according to the tracking multi-frame images to obtain track hidden danger indexes;
step S620: performing association recognition on the behavior dynamic relation of the first identification object and the second identification object according to the tracking multi-frame image to obtain a dynamic hidden danger index;
step S630: and outputting a first dynamic hidden danger probability according to the track hidden danger index and the dynamic hidden danger index.
Further, step S610 of the present application includes:
step S611: monitoring the behavior tracks of the first identification object and the second identification object respectively to obtain a first monitoring track and a second monitoring track;
step S612: performing matching degree identification according to the first monitoring track and the second monitoring track to obtain track deviation matching degree and track time matching degree;
step S613: obtaining a first matching degree according to the track deviation matching degree and the track time matching degree;
step S614: and if the first matching degree does not meet the preset matching degree, generating the track hidden danger index.
Further, step S620 of the present application includes:
step S621: calculating the dynamic property of the first identification object according to the tracking multi-frame image to obtain a first dynamic index;
step S622: identifying the behavior dynamic relation of the first identification object and the second identification object based on the first dynamic index to obtain a second dynamic index of the second identification object based on the first dynamic index;
step S623: and when the second dynamic index is larger than a preset dynamic index, generating the dynamic hidden danger index according to the index difference between the second dynamic index and the preset dynamic index.
Specifically, in order to ensure the accuracy of potential safety hazard early warning on the target elevator, the first identification object and the second identification object are firstly required to be subjected to matching recognition according to the tracking multi-frame images, namely the first identification object and the second identification object are respectively monitored in the target elevator to obtain a first monitoring track and a second monitoring track, the first monitoring track and the second monitoring track can comprise a transverse track and a longitudinal track, the transverse track refers to displacement generated by the first identification object and/or the second identification object in the target elevator, the longitudinal track refers to longitudinal tracks such as the first identification object and/or the second identification object moving up and down or jumping in the target elevator, and further, the first monitoring track and the second monitoring track are used as recognition basic data to conduct matching degree recognition between the first identification object and the second identification object, so that the relevance between the first identification object and the second identification object is judged, namely the lower when the running track is not matched, the track deviation degree and the track time matching degree between the first identification object are determined, the first identification object and the second identification object are extracted according to the relevance, and the potential safety hazard is calculated, and the potential safety hazard matching degree between the first identification object and the second identification object is calculated, and the potential safety hazard matching degree is calculated, and the potential safety hazard matching degree between the first identification object and the first identification object is high.
Further, the correlation recognition is performed on the behavior dynamic relationship of the first identification object and the second identification object according to the tracked multi-frame images, the calculation is performed on the multi-frame images of the first identification object on the basis of the tracking, the judgment is performed on the dynamic index of the second identification object according to the dynamic difference between every two adjacent images in the multi-frame images, if the dynamic difference is large, the dynamic of the first identification object is high at the moment, so that the first dynamic index of the first identification object is obtained, the behavior dynamic relationship of the first identification object and the second identification object is recognized on the basis of the first dynamic index, namely, the dynamic relationship of the second identification object is tracked when the dynamic index exists in the first identification object, the relation between the first identification object and the second identification object is judged to be dynamic, so that the second dynamic index of the second identification object is determined under the first dynamic index, the judgment is further performed on the basis of the second dynamic index, and if the second dynamic index is larger than the preset dynamic index, the second dynamic index is made, the judgment is performed on the basis of the second dynamic index, and the potential difference is larger than the preset dynamic index, so that the potential difference is generated on the basis of the potential safety hazard of the dynamic index.
And finally, adding and averaging the track hidden danger indexes obtained by matching and identifying the behavior tracks of the first identification object and the second identification object according to the tracking multi-frame images and the dynamic hidden danger indexes obtained by carrying out association and identification on the behavior dynamic relationship of the first identification object and the second identification object according to the tracking multi-frame images, and recording the average value data as the first dynamic hidden danger probability of the target elevator to output, thereby ensuring that the intelligent elevator is better managed safely based on the cloud platform in the later stage.
In summary, the intelligent elevator safety management method based on the cloud platform provided by the embodiment of the application at least comprises the following technical effects that the safety of an elevator is accurately managed and controlled based on the cloud platform, and the safety in the elevator is improved.
Example two
Based on the same inventive concept as the intelligent elevator safety management method based on the cloud platform in the foregoing embodiment, as shown in fig. 6, the present application provides an intelligent elevator safety management system based on the cloud platform, the system includes:
the system comprises an image acquisition module 1, wherein the image acquisition module 1 is used for acquiring multi-frame images in a ladder in a target elevator and uploading the multi-frame images in the ladder to a cloud platform, wherein a target segmentation model is embedded in the cloud platform and is used for segmenting objects in the multi-frame images in the ladder;
the object segmentation module 2 is used for carrying out object segmentation on the intra-ladder multi-frame image according to the target segmentation model embedded in the cloud platform, and determining a first identification object and a second identification object, wherein the second identification object is a carrying object of the first identification object;
the probability calculation module 3 is configured to input feature information of the first identification object and the second identification object into a potential safety hazard identification model, and calculate the potential safety hazard probability of the first identification object and the second identification object according to the potential safety hazard identification model to obtain a first potential safety hazard probability;
the instruction activation module 4 is configured to activate a tracking monitoring instruction when the first hidden danger probability is greater than a first preset hidden danger probability;
the tracking monitoring module 5 is used for tracking and monitoring the first identification object and the second identification object according to the tracking and monitoring instruction, and acquiring a tracking multi-frame image;
the feature recognition module 6 is configured to perform feature recognition according to the tracking multi-frame image, output a first dynamic hidden danger probability, and generate first safety early warning information when the first dynamic hidden danger probability is greater than a second preset hidden danger probability.
Further, the system further comprises:
the matching recognition module is used for carrying out matching recognition on the behavior tracks of the first identification object and the second identification object according to the tracking multi-frame images to obtain track hidden danger indexes;
the association identification module is used for carrying out association identification on the behavior dynamic relationship of the first identification object and the second identification object according to the tracking multi-frame image to obtain a dynamic hidden danger index;
the first output module is used for outputting a first dynamic hidden danger probability according to the track hidden danger index and the dynamic hidden danger index.
Further, the system further comprises:
the behavior track monitoring module is used for monitoring the behavior tracks of the first identification object and the second identification object respectively to obtain a first monitoring track and a second monitoring track;
the first matching degree identification module is used for carrying out matching degree identification according to the first monitoring track and the second monitoring track to obtain track deviation matching degree and track time matching degree;
the first matching degree identification module is used for obtaining a first matching degree according to the track deviation matching degree and the track time matching degree;
the first judging module is used for generating the track hidden danger index if the first matching degree does not meet the preset matching degree.
Further, the system further comprises:
the first calculation module is used for calculating the dynamic property of the first identification object according to the tracking multi-frame image to obtain a first dynamic index;
the identification module is used for identifying the behavior dynamic relation of the first identification object and the second identification object based on the first dynamic index to obtain a second dynamic index based on the second identification object under the first dynamic index;
and the second judging module is used for generating the dynamic hidden danger index according to the index difference between the second dynamic index and the preset dynamic index when the second dynamic index is larger than the preset dynamic index.
Further, the system further comprises:
the portable object acquisition module is used for acquiring a plurality of portable objects when the number of the portable objects of the first identification object is multiple;
the mapping relation module is used for establishing a one-to-many mapping relation between the first identification object and the plurality of carrying objects, and outputting a plurality of hidden danger probabilities based on the plurality of carrying objects according to the hidden danger identification model;
and the second output module is used for integrating the hidden danger probabilities and outputting the first hidden danger probability.
Further, the system further comprises:
the data set acquisition module is used for acquiring an elevator sample data set based on the cloud platform, wherein the elevator sample data set comprises an accident occurrence object and an accident occurrence probability;
the splitting module is used for splitting the accident occurrence objects in the elevator sample data set to obtain split objects, and each split object comprises object characteristic information;
and the model training module is used for carrying out model training according to the object characteristic information of the split object and the accident occurrence probability to obtain a pre-trained potential safety hazard identification model.
The foregoing detailed description of the intelligent elevator safety management method based on the cloud platform will clearly be known to those skilled in the art, and the intelligent elevator safety management system based on the cloud platform in this embodiment is described more simply for the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, and the relevant points refer to the description of the method section.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the application. Thus, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (6)
1. The intelligent elevator safety management method based on the cloud platform is characterized by comprising the following steps of:
collecting multi-frame images in a ladder in a target elevator, and uploading the multi-frame images in the ladder to a cloud platform, wherein a target segmentation model is embedded in the cloud platform and is used for segmenting objects in the multi-frame images in the ladder;
object segmentation is carried out on the multi-frame images in the ladder according to the target segmentation model embedded in the cloud platform, and a first identification object and a second identification object are determined, wherein the second identification object is a carrying object of the first identification object;
inputting the characteristic information of the first identification object and the second identification object into a potential safety hazard identification model, and calculating the potential safety hazard probability of the first identification object and the second identification object according to the potential safety hazard identification model to obtain a first potential safety hazard probability;
when the first hidden danger probability is larger than a first preset hidden danger probability, activating a tracking monitoring instruction;
tracking and monitoring the first identification object and the second identification object according to the tracking and monitoring instruction to acquire a tracking multi-frame image;
performing feature recognition according to the tracking multi-frame image, outputting a first dynamic hidden danger probability, and generating first safety early warning information when the first dynamic hidden danger probability is larger than a second preset hidden danger probability;
wherein the method comprises the following steps:
acquiring characteristic information of the first identification object, wherein the characteristic information comprises body type characteristic information, sex characteristic information and age characteristic information;
acquiring characteristic information of the second identification object, wherein the characteristic information comprises carried volume characteristic information, static and dynamic characteristic information and constraint characteristic information;
the characteristic information of the first identification object and the characteristic information of the second identification object are used as training data, and are input into the pre-trained potential safety hazard identification model to be calculated, so that a first potential hazard probability for identifying the potential safety hazard degrees of the first identification object and the second identification object is obtained;
wherein the method further comprises:
when the number of the carrying objects of the first identification object is multiple, obtaining multiple carrying objects;
establishing a one-to-many mapping relation between the first identification object and the plurality of carrying objects, and outputting a plurality of hidden danger probabilities based on the plurality of carrying objects according to the hidden danger identification model;
and integrating the hidden danger probabilities and outputting the first hidden danger probability.
2. The method of claim 1, wherein the feature recognition is performed based on the tracked multi-frame image to output a first probability of dynamic risk, the method comprising:
matching and identifying the behavior tracks of the first identification object and the second identification object according to the tracking multi-frame images to obtain track hidden danger indexes;
performing association recognition on the behavior dynamic relation of the first identification object and the second identification object according to the tracking multi-frame image to obtain a dynamic hidden danger index;
and outputting a first dynamic hidden danger probability according to the track hidden danger index and the dynamic hidden danger index.
3. The method of claim 2, wherein the matching recognition of the behavior trace of the first identified object and the second identified object is performed from the tracked multi-frame image, the method comprising:
monitoring the behavior tracks of the first identification object and the second identification object respectively to obtain a first monitoring track and a second monitoring track;
performing matching degree identification according to the first monitoring track and the second monitoring track to obtain track deviation matching degree and track time matching degree;
obtaining a first matching degree according to the track deviation matching degree and the track time matching degree;
and if the first matching degree does not meet the preset matching degree, generating the track hidden danger index.
4. The method of claim 2, wherein the identifying of the association of the behavioral dynamic relationship of the first identified object and the second identified object is based on the tracking multi-frame image, the method comprising:
calculating the dynamic property of the first identification object according to the tracking multi-frame image to obtain a first dynamic index;
identifying the behavior dynamic relation of the first identification object and the second identification object based on the first dynamic index to obtain a second dynamic index of the second identification object based on the first dynamic index;
and when the second dynamic index is larger than a preset dynamic index, generating the dynamic hidden danger index according to the index difference between the second dynamic index and the preset dynamic index.
5. The method of claim 1, wherein the method further comprises:
acquiring an elevator sample data set based on the cloud platform, wherein the elevator sample data set comprises an accident occurrence object and an accident occurrence probability;
splitting accident occurrence objects in the elevator sample data set to obtain split objects, wherein each split object comprises object characteristic information;
and performing model training according to the object characteristic information of the split object and the accident occurrence probability to obtain a pre-trained potential safety hazard identification model.
6. A cloud platform based intelligent elevator safety management system, characterized in that it is applied to the method of any one of claims 1 to 5, said system comprising:
the system comprises an image acquisition module, a cloud platform and a target segmentation module, wherein the image acquisition module is used for acquiring multi-frame images in a ladder in a target elevator and uploading the multi-frame images in the ladder to the cloud platform, wherein the cloud platform is embedded with the target segmentation model, and the target segmentation model is used for segmenting objects in the multi-frame images in the ladder;
the object segmentation module is used for carrying out object segmentation on the intra-ladder multi-frame image according to the target segmentation model embedded in the cloud platform, and determining a first identification object and a second identification object, wherein the second identification object is a carrying object of the first identification object;
the probability calculation module is used for inputting the characteristic information of the first identification object and the second identification object into a potential safety hazard identification model, and carrying out potential safety hazard probability calculation on the first identification object and the second identification object according to the potential safety hazard identification model to obtain first potential safety hazard probability;
the instruction activation module is used for activating a tracking monitoring instruction when the first hidden danger probability is larger than a first preset hidden danger probability;
the tracking monitoring module is used for tracking and monitoring the first identification object and the second identification object according to the tracking and monitoring instruction to acquire a tracking multi-frame image;
the feature recognition module is used for carrying out feature recognition according to the tracking multi-frame images, outputting first dynamic hidden danger probability, and generating first safety early warning information when the first dynamic hidden danger probability is larger than second preset hidden danger probability.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311135363.1A CN116863409B (en) | 2023-09-05 | 2023-09-05 | Intelligent elevator safety management method and system based on cloud platform |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311135363.1A CN116863409B (en) | 2023-09-05 | 2023-09-05 | Intelligent elevator safety management method and system based on cloud platform |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116863409A CN116863409A (en) | 2023-10-10 |
| CN116863409B true CN116863409B (en) | 2023-11-07 |
Family
ID=88229041
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311135363.1A Active CN116863409B (en) | 2023-09-05 | 2023-09-05 | Intelligent elevator safety management method and system based on cloud platform |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116863409B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117439223B (en) * | 2023-10-27 | 2024-06-07 | 南通沃太新能源有限公司 | Safety control method and system of energy storage system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298265A (en) * | 2019-06-10 | 2019-10-01 | 东南大学 | Specific objective detection method in a kind of elevator based on YOLO neural network |
| CN110852198A (en) * | 2019-10-27 | 2020-02-28 | 恒大智慧科技有限公司 | Control method, equipment and storage medium for preventing pet dog attack in smart community |
| CN113291944A (en) * | 2021-05-21 | 2021-08-24 | 南京喜诗图家具有限公司 | Pet safety protection equipment is used to elevator |
| CN114332925A (en) * | 2021-12-20 | 2022-04-12 | 苏州汇川控制技术有限公司 | Method, system and device for detecting pets in elevator and computer readable storage medium |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20230029187A (en) * | 2021-08-24 | 2023-03-03 | 김나은 | Application for pet dog managament based on IOT and pet dog care system |
-
2023
- 2023-09-05 CN CN202311135363.1A patent/CN116863409B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110298265A (en) * | 2019-06-10 | 2019-10-01 | 东南大学 | Specific objective detection method in a kind of elevator based on YOLO neural network |
| CN110852198A (en) * | 2019-10-27 | 2020-02-28 | 恒大智慧科技有限公司 | Control method, equipment and storage medium for preventing pet dog attack in smart community |
| CN113291944A (en) * | 2021-05-21 | 2021-08-24 | 南京喜诗图家具有限公司 | Pet safety protection equipment is used to elevator |
| CN114332925A (en) * | 2021-12-20 | 2022-04-12 | 苏州汇川控制技术有限公司 | Method, system and device for detecting pets in elevator and computer readable storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116863409A (en) | 2023-10-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110473192B (en) | Digestive tract endoscope image recognition model training and recognition method, device and system | |
| Zhang et al. | Real-time sow behavior detection based on deep learning | |
| Guo et al. | Multi-object extraction from topview group-housed pig images based on adaptive partitioning and multilevel thresholding segmentation | |
| Youssif et al. | Automatic facial expression recognition system based on geometric and appearance features | |
| CN116863409B (en) | Intelligent elevator safety management method and system based on cloud platform | |
| CN109241842B (en) | Fatigue driving detection method, device, computer equipment and storage medium | |
| CN110533950A (en) | Detection method, device, electronic equipment and the storage medium of parking stall behaviour in service | |
| CN110058699B (en) | User behavior identification method based on intelligent mobile device sensor | |
| US11887303B2 (en) | Image processing model generation method, image processing method and device, and electronic device | |
| Vipparthi et al. | Directional local ternary patterns for multimedia image indexing and retrieval | |
| CN109117723B (en) | Blind road detection method based on color pattern analysis and semantic segmentation | |
| Chen et al. | Fully automated natural frequency identification based on deep-learning-enhanced computer vision and power spectral density transmissibility | |
| CN114445917A (en) | Network training method and system for face living body recognition and electronic equipment | |
| Farahnakian et al. | Multi-pig pose estimation using DeepLabCut | |
| Tello-Cifuentes et al. | Detection and classification of pavement damages using wavelet scattering transform, fractal dimension by box-counting method and machine learning algorithms | |
| Banerjee et al. | Skin Diseases Detection Using LBP and WLD: An Ensembling Approach | |
| CN116091427A (en) | Dermatosis classification device, classification method and storage medium | |
| CN111582404B (en) | Content classification method, device and readable storage medium | |
| Li et al. | An unsupervised concrete crack detection method based on nnU-Net | |
| Li et al. | Mammogram-based discriminant fusion analysis for breast cancer diagnosis | |
| Bourennane et al. | Deep Feature Extraction with Cubic-SVM for Classification of Brain Tumor | |
| Wu et al. | Fault diagnosis of pumping system based on multimodal attention learning (CBMA Learning) | |
| Garg et al. | ADFAC: automatic detection of facial articulatory features | |
| Melbin et al. | Automated detection and classification of skin diseases using diverse features and improved gray wolf‐based multiple‐layer perceptron neural network | |
| Yancey | Deep Learning for Localization of Mixed Image Tampering Techniques |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |