CN112784759A - Elevator human detection identification method based on artificial intelligence similarity comparison - Google Patents
Elevator human detection identification method based on artificial intelligence similarity comparison Download PDFInfo
- Publication number
- CN112784759A CN112784759A CN202110099868.1A CN202110099868A CN112784759A CN 112784759 A CN112784759 A CN 112784759A CN 202110099868 A CN202110099868 A CN 202110099868A CN 112784759 A CN112784759 A CN 112784759A
- Authority
- CN
- China
- Prior art keywords
- elevator
- human body
- image
- body area
- background image
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Indicating And Signalling Devices For Elevators (AREA)
Abstract
本发明涉及电梯安全管理技术领域，尤其涉及一种基于人工智能相似度比对的电梯人感识别方法。本发明采用如下技术方案：通过先对实时图像与背景图像计算平均结构相似性进行初步筛选，再将不超出阀值的图像利用物体检测模型进行人体区域初步检测得到疑似人体区域坐标，最后再将初步检测出来的疑似人体区域坐标与其对应的背景图像再次计算平均结构相似性进行二次筛查，以此确定电梯内是否有人。本发明的优点在于：无需增加红外传感器设备进行人感识别，可利用摄像头采集的图像进行人感识别，还能有效提高人感识别的正确率，确保电梯的安全高效管理。
The invention relates to the technical field of elevator safety management, in particular to an elevator human sense recognition method based on artificial intelligence similarity comparison. The present invention adopts the following technical scheme: by firstly screening the real-time image and the background image by calculating the average structural similarity, and then using the object detection model to perform preliminary detection of the human body area on the image that does not exceed the threshold value to obtain the coordinates of the suspected human body area, and finally The preliminarily detected coordinates of the suspected human body area and its corresponding background image are calculated again by calculating the average structural similarity for secondary screening, so as to determine whether there is someone in the elevator. The advantages of the present invention are that it does not need to add infrared sensor equipment for human sense recognition, the image collected by the camera can be used for human sense recognition, the accuracy of human sense recognition can be effectively improved, and the safe and efficient management of elevators is ensured.
Description
Technical Field
The invention relates to the technical field of human detection, in particular to an elevator human detection method based on artificial intelligence similarity comparison.
Background
With the popularization of elevators, people pay more and more attention to elevator safety. And the camera can be installed in the elevator car for monitoring in the elevator car, and simultaneously, whether someone exists in the elevator can be monitored by using the camera in the elevator, so that the condition that the elevator is closed is prevented. At present, whether a person is in an elevator is detected by using a camera in the elevator, and human body infrared detection is generally carried out by adding an infrared sensor or human body identification is carried out on a picture shot by the camera in the elevator by using conventional human body identification detection software. The use of the infrared sensor can increase the detection cost, false alarm is easy to occur in the detection of the infrared sensor on a human body, and meanwhile, the infrared sensor is easily interfered by electromagnetic waves of electrical equipment to cause detection errors; however, when the conventional human body identification detection software is used, identification errors can be caused due to low identification rate of the existing human body identification detection software, and the elevator safety management is more unfavorable. Therefore, there is a need for an elevator human recognition method with high recognition rate without adding additional hardware.
Disclosure of Invention
The invention aims to provide an elevator human recognition method, and particularly provides an elevator human recognition method which is based on artificial intelligence similarity comparison and high in recognition rate.
In order to achieve the purpose, the invention adopts the following technical scheme: an elevator human detection identification method based on artificial intelligence similarity comparison comprises the following steps:
s11, firstly, a camera in the elevator is used for collecting images of the elevator in an unmanned state, and the images are stored as background images.
S12, calculating average structural similarity between the real-time image acquired by the camera in the elevator and the background image, if the average structural similarity is larger than a certain threshold value, judging the image as the background image, namely judging that the elevator is in an unmanned state, and otherwise, entering the next step.
And S13, if the average structural similarity between the real-time image and the background image in the step S12 is not larger than a certain threshold value, calling an object detection model to perform human body detection with a lower threshold value to obtain a plurality of suspected human body area coordinates.
And S14, respectively calculating the average structure similarity of the plurality of suspected human body area coordinates obtained in the step S13 and the sub-images of the corresponding positions in the background image, excluding the suspected human body area coordinates with the average structure similarity larger than a certain threshold value, recording the remaining suspected human body area coordinates as human body area coordinates, judging that the elevator is occupied when the human body area coordinates exist, and judging that the elevator is in an unoccupied state when the suspected human body area coordinates are excluded.
Further, before step S11, the method further includes:
and S10, calculating an average gray value of the real-time image acquired by the camera in the elevator, judging that the elevator is in a light-off state if the average gray value is lower than a certain threshold value, judging that the elevator is in an unmanned state, and entering S11-S14 if the average gray value is not lower than the threshold value.
Specifically, the background image in step S11 is an image captured by the camera in a state where the elevator is lighted and no person is present.
Further, the background image in step S11 is re-collected and updated every unit time for comparison with the real-time image, where the unit time is preferably 1min to 30 min.
The invention has the advantages that: the method comprises the steps of firstly carrying out primary screening on the average structural similarity calculated by the real-time image and the background image, then carrying out primary detection on a human body region by using an object detection model, and finally carrying out secondary screening on the average structural similarity calculated by the human body region preliminarily detected and the background image corresponding to the human body region again so as to determine whether a person exists in the elevator. Not only need not to increase infrared sensor, can also effectively improve people and feel the exactness of discernment, ensure the safe high-efficient management of elevator.
Drawings
Fig. 1 is a flowchart of an elevator human detection recognition method of embodiment 1;
fig. 2 is a flowchart of an elevator human detection method of embodiment 2.
Detailed Description
Embodiment 1, referring to fig. 1, a method for identifying an elevator human based on artificial intelligence similarity comparison includes the following steps:
s11, firstly, a camera in the elevator is used for collecting images of the elevator in an unmanned state, and the images are stored as background images.
S12, calculating average structural similarity between the real-time image acquired by the camera in the elevator and the background image, if the average structural similarity is larger than a certain threshold value, judging the image as the background image, namely judging that the elevator is in an unmanned state, and otherwise, entering the next step.
And S13, if the average structural similarity between the real-time image and the background image in the step S12 is not larger than a certain threshold value, calling an object detection model to perform human body detection with a lower threshold value to obtain a plurality of suspected human body area coordinates.
And S14, respectively calculating the average structure similarity of the plurality of suspected human body area coordinates obtained in the step S13 and the sub-images of the corresponding positions in the background image, excluding the suspected human body area coordinates with the average structure similarity larger than a certain threshold value, recording the remaining suspected human body area coordinates as human body area coordinates, judging that the elevator is occupied when the human body area coordinates exist, and judging that the elevator is in an unoccupied state when the suspected human body area coordinates are excluded.
In the embodiment, the average structural similarity is calculated by the real-time image collected by the camera and the elevator background image in an unmanned state, the larger the value of the average structural similarity is, the more similar the real-time image and the background image are, when the average structural similarity is larger than a certain threshold value, the real-time image can be judged to be consistent with the background image, i.e., no person is in the elevator, and on the contrary, when the average structural similarity does not exceed the threshold value, the human body detection of step S13 is performed on the real-time image, wherein the human body detection is to detect the real-time image by using an object detection model, in particular to detect the human body by using the SSD MobileNet-based object detection model, when the object detection model is used for detecting the human body, if a higher threshold value is set for detecting the human body, the situation that the elevator is judged to be unmanned due to the presence of people is easy to occur, and a larger potential safety hazard exists; after human body detection is carried out by using the object detection model, a plurality of suspected human body area coordinates are obtained, then the suspected human body area coordinates are screened for the second time, namely, the average structural similarity of the coordinates of a plurality of suspected human body areas and the subimages at the corresponding positions on the background image is calculated, if the average structural similarity between the image of the suspected human body area coordinates and the sub-image of the corresponding position on the background image is larger than a certain threshold value, the suspected human body area coordinate is actually the background image and is not the real human body area coordinate, only when the average structural similarity between the suspected human body area coordinate and the background image does not exceed the threshold value, the suspected human body area coordinate is confirmed as the real human body area coordinate, and at the moment, the person in the elevator can be judged, so that the efficiency and the accuracy of human body identification can be effectively improved, and the safety of the elevator is improved.
Embodiment 2, referring to fig. 2, a method for identifying an elevator human based on artificial intelligence similarity comparison includes the following steps:
and S10, calculating an average gray value of the real-time image acquired by the camera in the elevator, judging that the elevator is in a light-off state if the average gray value is lower than a certain threshold value, judging that the elevator is in an unmanned state, and entering S11-S14 if the average gray value is not lower than the threshold value.
S11, firstly, a camera in the elevator is used for collecting images of the elevator in an unmanned state, and the images are stored as background images.
S12, calculating average structural similarity between the real-time image acquired by the camera in the elevator and the background image, if the average structural similarity is larger than a certain threshold value, judging the image as the background image, namely judging that the elevator is in an unmanned state, and otherwise, entering the next step.
And S13, if the average structural similarity between the real-time image and the background image in the step S12 is not larger than a certain threshold value, calling an object detection model to perform human body detection with a lower threshold value to obtain a plurality of suspected human body area coordinates.
And S14, respectively calculating the average structure similarity of the plurality of suspected human body area coordinates obtained in the step S13 and the sub-images of the corresponding positions in the background image, excluding the suspected human body area coordinates with the average structure similarity larger than a certain threshold value, recording the remaining suspected human body area coordinates as human body area coordinates, judging that the elevator is occupied when the human body area coordinates exist, and judging that the elevator is in an unoccupied state when the suspected human body area coordinates are excluded.
Compared with the embodiment 1, the difference of the embodiment is that before the average structural similarity calculation is performed on the real-time image and the background image, the average gray value calculation is performed on the real-time image; because the elevator is a specific scene, the lighting lamp in the elevator car is triggered to be turned off or in an energy-saving mode when no person takes the elevator for a long time, the brightness in the elevator car is greatly reduced at the moment, the average gray value of the real-time image is lower at the moment, and the light-off state in the elevator at the moment can be judged by calculating the average gray value of the real-time image when the average gray value is lower than a certain threshold value, namely no person exists in the elevator. By calculating the average gray value of the real-time image, the real-time image in the elevator in a light-out state can be eliminated, and the human recognition efficiency is improved.
Of course, the background image used for the average structural similarity calculation with the real-time image, i.e., the background image in step S11, is an image captured by the camera in a state where there is a light in the elevator and no person is present.
Further, the background image in step S11 is re-collected and updated every unit time for comparison with the real-time image, where the unit time is preferably 1min to 30 min. The background image is refreshed at regular time, so that the image with a bright lamp in the elevator and in an unmanned state can be updated at regular time, the comparison error caused by the change of the elevator background is prevented, and the frequency of the refresh at regular time can be set according to the actual use condition of the elevator, generally 1-30 min, preferably 10 min; the time can also be extended appropriately according to needs, such as to 1h or 2h, and the like, without limitation.
It should be understood that the above-mentioned embodiments are merely preferred embodiments of the present invention, and not intended to limit the scope of the invention, therefore, all equivalent changes in the principle of the present invention should be included in the protection scope of the present invention.
Claims (5)
1. An elevator human detection identification method based on artificial intelligence similarity comparison is characterized in that: the method comprises the following steps:
s11, firstly, carrying out image acquisition on the elevator in an unmanned state by using a camera in the elevator, and storing the image as a background image;
s12, calculating average structural similarity between a real-time image acquired by a camera in the elevator and a background image, if the average structural similarity is larger than a certain threshold value, judging the image as the background image, namely judging that the elevator is in an unmanned state, and otherwise, entering the next step;
s13, if the average structural similarity between the real-time image and the background image in the step S12 is not larger than a certain threshold value, calling an object detection model to perform human body detection with a lower threshold value to obtain a plurality of suspected human body area coordinates;
and S14, respectively calculating the average structure similarity of the plurality of suspected human body area coordinates obtained in the step S13 and the sub-images of the corresponding positions in the background image, excluding the suspected human body area coordinates with the average structure similarity larger than a certain threshold value, recording the remaining suspected human body area coordinates as human body area coordinates, judging that the elevator is occupied when the human body area coordinates exist, and judging that the elevator is in an unoccupied state when the suspected human body area coordinates are excluded.
2. The elevator human recognition method based on artificial intelligence similarity comparison according to claim 1, characterized in that: before the step S11, the method further comprises
And S10, calculating an average gray value of the real-time image acquired by the camera in the elevator, judging that the elevator is in a light-off state if the average gray value is lower than a certain threshold value, judging that the elevator is in an unmanned state, and entering S11-S14 if the average gray value is not lower than the threshold value.
3. The method for identifying the human beings in the elevator based on the artificial intelligence similarity comparison as claimed in claim 2, wherein: the background image in step S11 is an image captured by the camera in a state where the elevator is lighted and no person is present.
4. The method for identifying the human beings in the elevator based on the artificial intelligence similarity comparison, according to claim 3, is characterized in that: the background image in step S11 is re-acquired every unit time and updated for comparison with the real-time image.
5. The method for identifying the human beings in the elevator based on the artificial intelligence similarity comparison, according to claim 4, is characterized in that: the unit time is 1-30 min.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110099868.1A CN112784759A (en) | 2021-01-25 | 2021-01-25 | Elevator human detection identification method based on artificial intelligence similarity comparison |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110099868.1A CN112784759A (en) | 2021-01-25 | 2021-01-25 | Elevator human detection identification method based on artificial intelligence similarity comparison |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112784759A true CN112784759A (en) | 2021-05-11 |
Family
ID=75759110
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110099868.1A Pending CN112784759A (en) | 2021-01-25 | 2021-01-25 | Elevator human detection identification method based on artificial intelligence similarity comparison |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112784759A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115240389A (en) * | 2022-08-03 | 2022-10-25 | 江苏商贸职业学院 | Monitoring and early warning system based on artificial intelligence |
| CN116246215A (en) * | 2023-05-11 | 2023-06-09 | 小手创新(杭州)科技有限公司 | Method for identifying new articles based on visual algorithm, barrel cover and intelligent recycling bin |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103425957A (en) * | 2012-05-14 | 2013-12-04 | 无锡中科智能信息处理研发中心有限公司 | Recognition method for judging whether person image exists in elevator or not through camera |
| CN103771207A (en) * | 2014-01-09 | 2014-05-07 | 深圳市汇川技术股份有限公司 | Elevator car passenger detection system and method |
| CN110002314A (en) * | 2018-08-09 | 2019-07-12 | 浙江新再灵科技股份有限公司 | A kind of stranded number detection system of the elevator passenger based on deep learning |
| CN111325133A (en) * | 2020-02-17 | 2020-06-23 | 深圳龙安电力科技有限公司 | Image processing system based on artificial intelligence recognition |
| CN111783700A (en) * | 2020-07-06 | 2020-10-16 | 中国交通通信信息中心 | Automatic recognition early warning method and system for road foreign matters |
| CN111814720A (en) * | 2020-07-17 | 2020-10-23 | 电子科技大学 | A method for detection and classification of foreign objects in airport runways based on UAV vision |
-
2021
- 2021-01-25 CN CN202110099868.1A patent/CN112784759A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103425957A (en) * | 2012-05-14 | 2013-12-04 | 无锡中科智能信息处理研发中心有限公司 | Recognition method for judging whether person image exists in elevator or not through camera |
| CN103771207A (en) * | 2014-01-09 | 2014-05-07 | 深圳市汇川技术股份有限公司 | Elevator car passenger detection system and method |
| CN110002314A (en) * | 2018-08-09 | 2019-07-12 | 浙江新再灵科技股份有限公司 | A kind of stranded number detection system of the elevator passenger based on deep learning |
| CN111325133A (en) * | 2020-02-17 | 2020-06-23 | 深圳龙安电力科技有限公司 | Image processing system based on artificial intelligence recognition |
| CN111783700A (en) * | 2020-07-06 | 2020-10-16 | 中国交通通信信息中心 | Automatic recognition early warning method and system for road foreign matters |
| CN111814720A (en) * | 2020-07-17 | 2020-10-23 | 电子科技大学 | A method for detection and classification of foreign objects in airport runways based on UAV vision |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115240389A (en) * | 2022-08-03 | 2022-10-25 | 江苏商贸职业学院 | Monitoring and early warning system based on artificial intelligence |
| CN116246215A (en) * | 2023-05-11 | 2023-06-09 | 小手创新(杭州)科技有限公司 | Method for identifying new articles based on visual algorithm, barrel cover and intelligent recycling bin |
| CN116246215B (en) * | 2023-05-11 | 2024-01-09 | 小手创新(杭州)科技有限公司 | Method for identifying new articles based on visual algorithm, barrel cover and intelligent recycling bin |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111144247B (en) | A method for retrograde detection of escalator passengers based on deep learning | |
| CN106373320B (en) | Method for recognizing fire disaster based on flame color dispersion and sequential frame image similarity | |
| CN205507937U (en) | District security protection control system | |
| EP1955285A1 (en) | Detection of abnormal crowd behavior | |
| CN112784759A (en) | Elevator human detection identification method based on artificial intelligence similarity comparison | |
| CN110298837B (en) | Method for detecting fire-fighting road occupation abnormal object based on interframe difference method | |
| CN104743419B (en) | Image monitoring device and elevator monitoring device | |
| CN113936293A (en) | Video information-based off-duty personnel detection method, control device and storage medium | |
| CN107337044A (en) | The detection method and system of elevator components failure | |
| CN110211316B (en) | Monitoring early warning method and system | |
| WO2020248543A1 (en) | Abnormal target detection method and device, and storage medium | |
| CN101316371A (en) | Flame detection method and device | |
| CN109492548B (en) | Method for obtaining region mask picture based on video analysis | |
| CN114519532A (en) | Personnel state detection method based on machine vision and related device | |
| CN108737785B (en) | Indoor automatic detection system that tumbles based on TOF 3D camera | |
| CN115052110B (en) | Security method, security system and computer readable storage medium | |
| CN115367583A (en) | Elevator safety guarantee system based on edge calculation | |
| JP2012212238A (en) | Article detection device and stationary-person detection device | |
| WO2022198507A1 (en) | Obstacle detection method, apparatus, and device, and computer storage medium | |
| CN117132949B (en) | An all-weather fall detection method based on deep learning | |
| CN109896387B (en) | Elevator passenger dangerous behavior detection method and system based on 3D measurement technology | |
| CN119832703A (en) | Security video monitoring system based on big data | |
| CN107145861A (en) | A method for automatic detection of abandoned objects | |
| CN111223255A (en) | Intelligent elevator monitoring system and monitoring equipment | |
| JP5586383B2 (en) | Image monitoring device |
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 |
Application publication date: 20210511 |
|
| WD01 | Invention patent application deemed withdrawn after publication |