CN112270333A - Elevator car abnormity detection method and system aiming at electric vehicle identification - Google Patents
Elevator car abnormity detection method and system aiming at electric vehicle identification Download PDFInfo
- Publication number
- CN112270333A CN112270333A CN202010479416.1A CN202010479416A CN112270333A CN 112270333 A CN112270333 A CN 112270333A CN 202010479416 A CN202010479416 A CN 202010479416A CN 112270333 A CN112270333 A CN 112270333A
- Authority
- CN
- China
- Prior art keywords
- electric vehicle
- elevator car
- data set
- detection
- training
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 46
- 238000012549 training Methods 0.000 claims abstract description 39
- 238000003062 neural network model Methods 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims abstract description 6
- 238000005457 optimization Methods 0.000 claims abstract description 4
- 240000007651 Rubus glaucus Species 0.000 claims description 7
- 235000011034 Rubus glaucus Nutrition 0.000 claims description 7
- 235000009122 Rubus idaeus Nutrition 0.000 claims description 7
- 230000002159 abnormal effect Effects 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 5
- 230000001960 triggered effect Effects 0.000 claims description 3
- 230000005856 abnormality Effects 0.000 claims 3
- 238000000034 method Methods 0.000 abstract description 10
- 238000010276 construction Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000002372 labelling Methods 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 125000004432 carbon atom Chemical group C* 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001915 proofreading effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
Images
Classifications
-
- 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
-
- 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
-
- 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
- B66B5/02—Applications of checking, fault-correcting, or safety devices in elevators responsive to abnormal operating conditions
-
- 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
Abstract
The invention discloses an elevator car abnormity detection method and system aiming at electric vehicle identification, which comprises the following steps: constructing a special training data set for the electric vehicle, uploading and downloading pictures of the electric vehicle on a hundred-degree search engine, and marking the position of the electric vehicle in the training set by using an image marking tool to serve as a label of training data; in the model training step, the accuracy of the electric vehicle identification model is improved through 1000 rounds of iterative optimization based on the electric vehicle training data set in the first step. The method is embodied in that a self-built electric vehicle data set is used for training a neural network model, so that the identification precision of the electric vehicle is improved, the safety of a lift car is ensured, and the risk of safety accidents in the later period is reduced; the special electric vehicle data set is constructed, so that the identification and detection precision of the electric vehicle is improved; an edge calculation mode is adopted, and a microcontroller is utilized to solve the problem of time delay depending on a cloud network and even the problem of detection and identification of the electric vehicle with unavailable network.
Description
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to an elevator car abnormity detection method and system aiming at electric vehicle identification.
Background
The object detection means detecting the position and the category of an object in an image or a video, and the traditional method is realized by manually designing features and utilizing a shallow classifier. An end-to-end detection framework based on deep learning realizes object detection through a deep neural network, however, in computer vision, the number, size and posture of objects in each image are different, and meanwhile, the objects in the images are often blocked and cut, which is a challenge of an object detection technology.
In electric motor car discernment field, the scene that gets into elevator car to the violation is bright to be related to, to the detection of single targets such as electric motor car under this scene, helps improving the security that the elevator used.
For different object detection scenes, especially in electric vehicle detection, a large number of special images for electric vehicles are required to construct a special data set to identify and improve the detection performance of the electric vehicles.
Disclosure of Invention
The invention aims to provide an elevator car abnormity detection method and system aiming at electric vehicle identification, so as to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme: an elevator car abnormity detection method and system aiming at electric vehicle identification concretely comprises the following steps:
A. constructing a special training data set for the electric vehicle, loading and unloading electric vehicle pictures on a hundred-degree search engine, and marking the position of the electric vehicle in the training set by using an image marking tool to serve as a label of training data;
B. in the model training step, based on the electric vehicle training data set in the first step, through 1000 rounds of iterative optimization, the accuracy of the electric vehicle identification model is improved, so that the alarm function of abnormal detection of the electric vehicle entering the elevator car is achieved;
C. running electric vehicle detection program
B, based on the trained electric vehicle identification and detection model program in the step B, deploying the electric vehicle identification and detection model program on a raspberry group, and connecting the electric vehicle identification and detection model program with a camera deployed at the top of an elevator car through a wired network so as to ensure that the possible illegal entering behavior of the electric vehicle is identified and detected in real time when the elevator car door acquired by the camera is opened;
D. output of detection result
When the electric car detection program identifies the abnormal condition that the electric car enters the elevator car, an alarm signal is triggered, an alarm prompt is sent until the electric car leaves the car, at the moment, the elevator can normally run, and when the electric car does not appear in the camera of the elevator car, the elevator keeps normally running.
Preferably, the training data set in step a specifically includes acquiring image data of the electric vehicle from 4 open source data sets, such as VOC, COCO, Imagenet, Openimage, and the like.
Preferably, the ratio of the carbon atoms in step B is 3: and 7, dividing the test data set and the training data set according to the proportion, and performing fine tuning training aiming at electric vehicle identification on a network parameter program by using a YOLOV3 neural network model.
Preferably, in the step C, a lightweight and fast deployment mode of the raspberry pi microcontroller is utilized based on the edge calculation mode.
Compared with the prior art, the invention has the following beneficial effects:
1. the method is embodied in that a self-built electric vehicle data set is used for training a neural network model, so that the identification precision of the electric vehicle is improved, the safety of a lift car is ensured, and the risk of safety accidents in the later period is reduced; the special electric vehicle data set is constructed, so that the identification and detection precision of the electric vehicle is improved; an edge calculation mode is adopted, and a microcontroller is utilized to solve the problem of time delay depending on a cloud network and even the problem of detection and identification of the electric vehicle with unavailable network.
Drawings
FIG. 1 is a schematic diagram of a full flow method for electric vehicle detection according to the present invention;
FIG. 2 is a general flow diagram of the electric vehicle inspection of the present invention;
fig. 3 is a schematic diagram of the detection signal control of the electric vehicle according to the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1-3, a method and a system for detecting an abnormal state of an elevator car for electric vehicle identification includes the following steps:
A. constructing a special training data set for the electric vehicle, loading and unloading electric vehicle pictures on a hundred-degree search engine, and marking the position of the electric vehicle in the training set by using an image marking tool to serve as a label of training data;
B. in the model training step, based on the electric vehicle training data set in the first step, through 1000 rounds of iterative optimization, the accuracy of the electric vehicle identification model is improved, so that the alarm function of abnormal detection of the electric vehicle entering the elevator car is achieved;
C. running electric vehicle detection program
B, based on the trained electric vehicle identification and detection model program in the step B, deploying the electric vehicle identification and detection model program on a raspberry group, and connecting the electric vehicle identification and detection model program with a camera deployed at the top of an elevator car through a wired network so as to ensure that the possible illegal entering behavior of the electric vehicle is identified and detected in real time when the elevator car door acquired by the camera is opened;
D. output of detection result
When the electric car detection program identifies the abnormal condition that the electric car enters the elevator car, an alarm signal is triggered, an alarm prompt is sent until the electric car leaves the car, at the moment, the elevator can normally run, and when the electric car does not appear in the camera of the elevator car, the elevator keeps normally running.
In the step a, the training data set specifically includes acquiring image data of the electric vehicle from 4 open source data sets, such as VOC, COCO, Imagenet, Openimage, and the like.
In step B as per 3: and 7, dividing the test data set and the training data set according to the proportion, and performing fine tuning training aiming at electric vehicle identification on a network parameter program by using a Yolov3 neural network model.
And C, based on an edge calculation mode, utilizing a lightweight and rapid deployment mode of the raspberry pi microcontroller.
The method comprises the following steps that a camera is installed in an elevator car, the monitoring direction of the camera is continuously aligned to an elevator door opening in the running process of the elevator, the entering condition of the elevator is collected, data collected by the camera is transmitted to an elevator car control panel, a detection program of the electric vehicle is started, and when the electric vehicle appears in monitoring, an electric vehicle signal is output to give an alarm; when no electric vehicle appears in the video, the electric vehicle runs normally.
The method is concise and convenient for deployment; by adopting an edge calculation mode and utilizing the raspberry group microcontroller, the defect that an identification program depends on cloud service is overcome, and the problem of network delay and even the problem of detection that a network is unavailable are solved.
The camera can be arranged in the elevator car or at the top outside the elevator door as required, the camera is connected with an elevator control panel carrying an electric vehicle detection program through a network cable, the camera monitors the in-out condition of the elevator car at any time during the operation of the elevator, and when the electric vehicle enters the elevator car, the program detects the electric vehicle, sends out an alarm signal and controls the elevator door to be normally open; when no electric vehicle enters the camera, the camera does not send out an alarm signal, and the elevator runs as usual.
3. Electric vehicle data set construction and model training process
The method comprises an electric vehicle image data acquisition part, a manual labeling and proofreading part, an electric vehicle special training data set construction part, an electric vehicle detection program operation part and a detection result output feedback part 5, wherein the electric vehicle image data acquisition part is derived from 4 open source data sets such as VOC, COCO, Imagenet and Openimage, and electric vehicle pictures are loaded and unloaded on a hundred-degree search engine.
In the manual marking calibration, an image marking tool is used for marking the position of the electric vehicle in the electric vehicle image to be used as a label of training data.
Establishing a special training data set for the electric vehicle, and carrying out data acquisition on an open source website, data acquisition on a search engine and manually marked data according to the following steps of 3: and 7, dividing the data into a test data set and a training data set according to the proportion, so that the next step of model test and training is facilitated, and the identification accuracy is improved.
In the running of the electric vehicle detection program, a Yolov3 neural network model is used for carrying out transfer learning training, namely, model parameter fine adjustment is carried out by using electric vehicle data so as to reduce the computational cost.
In the detection result output feedback, the electric vehicle image detected by the Yolov3 neural network model is directly stored into a constructed special data set, the type can be identified when the data is described, the electric vehicle image can be directly used as training data for further enhancing the detection performance, the image with the wrong detection result enters a manual labeling and correcting step, and the data with the wrong detection is further corrected.
And finally, storing the special data set of the electric vehicle, deploying the electric vehicle detection model in the elevator control panel, and completing the identification program training and data set construction aiming at the electric vehicle in the elevator car.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.
Claims (4)
1. An elevator car abnormity detection method and system aiming at electric vehicle identification are characterized in that: the specific implementation comprises the following steps:
A. constructing a special training data set for the electric vehicle, uploading and downloading pictures of the electric vehicle on a hundred-degree search engine, and marking the position of the electric vehicle in the training set by using an image marking tool to serve as a label of training data;
B. in the model training step, based on the electric vehicle training data set in the first step, through 1000 rounds of iterative optimization, the accuracy of the electric vehicle identification model is improved, so that the alarm function of abnormal detection of the electric vehicle entering the elevator car is achieved;
C. running electric vehicle detection program
B, based on the trained electric vehicle identification and detection model program in the step B, deploying the electric vehicle identification and detection model program on a raspberry group, and connecting the electric vehicle identification and detection model program with a camera deployed at the top of an elevator car through a wired network so as to ensure that the possible illegal entering behavior of the electric vehicle is identified and detected in real time when the elevator car door acquired by the camera is opened;
D. output of detection result
When the electric car detection program identifies the abnormal condition that the electric car enters the elevator car, an alarm signal is triggered, an alarm prompt is sent until the electric car leaves the car, at the moment, the elevator can normally run, and when the electric car does not appear in the camera of the elevator car, the elevator keeps normally running.
2. The elevator car abnormality detection method and system for electric vehicle identification according to claim 1, characterized in that: the training data set in the step a specifically includes acquiring image data of the electric vehicle from 4 open source data sets, such as VOC, COCO, Imagenet, Openimage, and the like.
3. The elevator car abnormality detection method and system for electric vehicle identification according to claim 1, characterized in that: in step B, according to 3: and 7, dividing the test data set and the training data set according to the proportion, and performing fine tuning training aiming at electric vehicle identification on a network parameter program by using a Yolov3 neural network model.
4. The elevator car abnormality detection method and system for electric vehicle identification according to claim 1, characterized in that: and C, based on an edge calculation mode, utilizing a lightweight and rapid deployment mode of the raspberry pi microcontroller.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010479416.1A CN112270333A (en) | 2020-05-30 | 2020-05-30 | Elevator car abnormity detection method and system aiming at electric vehicle identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010479416.1A CN112270333A (en) | 2020-05-30 | 2020-05-30 | Elevator car abnormity detection method and system aiming at electric vehicle identification |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112270333A true CN112270333A (en) | 2021-01-26 |
Family
ID=74348493
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010479416.1A Pending CN112270333A (en) | 2020-05-30 | 2020-05-30 | Elevator car abnormity detection method and system aiming at electric vehicle identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112270333A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117007101A (en) * | 2023-09-28 | 2023-11-07 | 广东星云开物科技股份有限公司 | Vehicle monitoring method, device, electronic equipment and storage medium |
-
2020
- 2020-05-30 CN CN202010479416.1A patent/CN112270333A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117007101A (en) * | 2023-09-28 | 2023-11-07 | 广东星云开物科技股份有限公司 | Vehicle monitoring method, device, electronic equipment and storage medium |
| CN117007101B (en) * | 2023-09-28 | 2023-12-26 | 广东星云开物科技股份有限公司 | Vehicle monitoring method, device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20240087107A1 (en) | Systems and Methods for Utilizing Machine-Assisted Vehicle Inspection to Identify Insurance Buildup or Fraud | |
| GB2551001A (en) | Vision-Based Rain Detection Using Deep Learning | |
| CN109249937A (en) | Drive Computer Aided Design analysis system | |
| US20160007018A1 (en) | Failure detection apparatus and failure detection program | |
| US10308260B2 (en) | Method for testing the function of a driver assistance system and control unit and reference device for a driver assistance system | |
| US9875654B2 (en) | Real-time monitoring and diagnostic processing of traffic control data | |
| CN110889351A (en) | Video detection method and device, terminal equipment and readable storage medium | |
| CN111527387B (en) | Method and system for identifying damage to a component | |
| CN114330473A (en) | Perceptual system error detection and re-validation | |
| CN107967445A (en) | A kind of car installs the intelligent checking system and method for skylight additional | |
| CN112270333A (en) | Elevator car abnormity detection method and system aiming at electric vehicle identification | |
| WO2019127085A1 (en) | Processing method, processing apparatus, control device and cloud server | |
| CN113906436A (en) | Method and apparatus for detecting smoke | |
| US20220343158A1 (en) | Method, device, and computer program for creating training data in a vehicle | |
| CN115937575A (en) | Equipment accessory detection method, device, system, equipment and storage medium | |
| CN114966631A (en) | Fault diagnosis and processing method and device for vehicle-mounted laser radar, medium and vehicle | |
| EP4013036A1 (en) | Automatic camera configuration systems and methods | |
| CN110954341B (en) | Test method and system for intelligent networking automobile test scene | |
| CN111654695A (en) | Camera module detection data management system | |
| RU2772235C1 (en) | Method for tuning visualisation sensors, computer software product for tuning visualisation sensors, sensor tuning system, and unit | |
| CN111142904B (en) | Continuous updating method and system for software of automatic driving system | |
| KR102502394B1 (en) | System for predicting error of the vihicle and method thereof | |
| US20240067189A1 (en) | Control device and control method | |
| CN116958764A (en) | Method for detecting data of mounting equipment based on deep learning | |
| US20200125514A1 (en) | Method, device, central device, and system for detecting distribution shift in a data and/or feature distribution of input data |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20210126 |