CN110589648B - Vertical ladder door label detection method based on deep learning - Google Patents
Vertical ladder door label detection method based on deep learning Download PDFInfo
- Publication number
- CN110589648B CN110589648B CN201910828841.4A CN201910828841A CN110589648B CN 110589648 B CN110589648 B CN 110589648B CN 201910828841 A CN201910828841 A CN 201910828841A CN 110589648 B CN110589648 B CN 110589648B
- Authority
- CN
- China
- Prior art keywords
- gate
- door
- image
- tags
- label
- 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
Images
Classifications
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B3/00—Applications of devices for indicating or signalling operating conditions of elevators
- B66B3/002—Indicators
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/0006—Monitoring devices or performance analysers
- B66B5/0018—Devices monitoring the operating condition of the elevator system
- B66B5/0031—Devices monitoring the operating condition of the elevator system for safety reasons
-
- B—PERFORMING OPERATIONS; TRANSPORTING
- B66—HOISTING; LIFTING; HAULING
- B66B—ELEVATORS; ESCALATORS OR MOVING WALKWAYS
- B66B5/00—Applications of checking, fault-correcting, or safety devices in elevators
- B66B5/0006—Monitoring devices or performance analysers
- B66B5/0037—Performance analysers
Landscapes
- Indicating And Signalling Devices For Elevators (AREA)
Abstract
The invention relates to a method for detecting a label of a vertical ladder door based on deep learning, which comprises the following steps: s1, acquiring a real-time image of an elevator to be detected, judging whether the elevator is in a door closing state or not, and if so, triggering an image analysis request; s2, intercepting the image of the elevator according to the triggering time of the image analysis request; s3, detecting the gate tags according to the image, if the gate tags exist and the number of the gate tags is greater than or equal to 2, comparing the spacing distance between the gate tags with a preset threshold value, and if the spacing distance between the gate tags is smaller than the threshold value, generating a target picture frame containing the gate tags; and S4, judging whether the position of the image pickup equipment for acquiring the image deviates according to the position of the target picture frame in the image, if not, sending a first state result, and otherwise, sending a second state result. The invention solves the defect that a manager cannot master the state of the door label in the elevator in real time, and is beneficial to ensuring the door label to play a warning role and ensuring the personal safety of passengers.
Description
Technical Field
The invention relates to a detection method, in particular to a vertical ladder door label detection method based on deep learning.
Background
With the development of cities, elevators have become an indispensable part of modern buildings. In urban office buildings and community buildings, the convenience of elevators is increasingly prominent, but with the wide use of elevators, more elevator taking accidents also occur. Therefore, the arrangement of the inner door mark of the vertical elevator car also comes along. Through adopting the door label of posting up in the elevator and playing the warning effect, can effectively take precautions against the incident that the stranded trouble of elevator caused of trouble, furthest's guarantee passenger safety, reduction casualties and loss of property. According to the requirement for setting the mark of the elevator car, the mark in the elevator car needs to be standardized and perfected, the advertisement pasted in disorder is timely cleaned, and the mark of failure and irregularity is removed. The elevator door mark is provided with a warning mark for strictly preventing pushing and opening the door. Because the door labels posted on the car doors are very easy to be damaged and torn, the door labels can be found only when being checked by managers, and then the elevator inner door labels are difficult to be repaired in time when being damaged. Cannot play a role in warning in real time.
Disclosure of Invention
The invention aims to provide a method for detecting a vertical ladder door label based on deep learning, which realizes real-time detection of the door label.
In order to achieve the above object, the present invention provides a method for detecting a tag of a vertical ladder door based on deep learning, comprising:
s1, acquiring a real-time image of an elevator to be detected, judging whether the elevator is in a door closing state or not, and if so, triggering an image analysis request;
s2, intercepting the image of the elevator according to the triggering time of the image analysis request;
s3, detecting the gate tags according to the image, if the gate tags exist and the number of the gate tags is greater than or equal to 2, comparing the spacing distance between the gate tags with a preset threshold value, and if the spacing distance between the gate tags is smaller than the threshold value, generating a target picture frame containing the gate tags;
and S4, judging whether the position of the image pickup equipment for acquiring the image deviates according to the position of the target picture frame in the image, if not, sending a first state result, and otherwise, sending a second state result.
According to an aspect of the present invention, in step S3, if the gate tags exist and the number is greater than or equal to 2, comparing the distance between the gate tags with a preset threshold, and if the distance between the gate tags is greater than the threshold, generating an object frame for the gate tag closest to the middle position of the image, and sending a third status result.
According to an aspect of the present invention, in step S3, if the gate tags exist and the number is equal to 1, an object frame is generated for the gate tags, and a fourth status result is issued.
According to one aspect of the invention, in step S3, if the door tag is not present, a fifth status result is issued.
According to one aspect of the invention, the coordinates of the target frame in the image are transmitted to the image capturing device.
According to one aspect of the invention, the first state result, the second state result, the third state result, the fourth state result, and the fifth state result are transmitted to a cloud platform for recording;
the elevator is characterized in that the first state result shows that the elevator is of a double-door type and the door label is normal, the second state result shows that the camera equipment deviates, the third state result shows that the elevator is of a single-door type and the door label is normal, the fourth state result shows that the door label in the elevator is damaged or torn, and the fifth state result shows that the label is not detected in the elevator.
According to an aspect of the present invention, the step of generating the target frame comprises:
s31, generating a label picture frame matched with the door label in shape based on the obtained door label;
and S32, generating the target picture frame based on the label picture frame.
According to one aspect of the invention, the target picture frame is a hexagonal frame.
According to one aspect of the invention, the camera device identifies the opening and closing state of the elevator based on the target picture frame and the position of the door tag.
According to an aspect of the present invention, in the step of performing gate label detection on the image, a network framework model of YOLOv3 is used to detect the image.
According to one scheme of the invention, the image of each elevator is collected through the camera equipment, the label on the elevator door is detected through a target detection algorithm based on deep learning, and the automatic picture frame is carried out according to the detected door label while the state information of the door label is returned, so that the defect that a manager cannot master the state of the door label in the elevator in real time is effectively overcome, and the method is favorable for ensuring that the door label fully plays a warning role and ensuring the personal safety of passengers.
According to the scheme of the invention, the detected door tag is automatically painted, and real-time alarms can be given to the conditions of label part tearing, label position deviation, camera equipment position deviation and the like which are not adhered according to the specification, so that technicians can timely maintain the conditions and the door tag can fully play a warning role.
According to the scheme of the invention, the coordinates of the automatically drawn picture frame are sent to the camera equipment, so that the identification of the elevator door opening and closing by the camera equipment based on the target picture frame is realized, and the timeliness and the identification efficiency of the elevator door opening and closing identification are improved. Meanwhile, the identification precision is also ensured.
Drawings
FIG. 1 schematically illustrates a label view of a vertical lift gate;
FIG. 2 schematically illustrates a vertical lift door label posting location diagram;
FIG. 3 schematically illustrates a block diagram of steps of a vertical lift gate tag detection method according to one embodiment of the present invention;
FIG. 4 schematically shows a flow diagram of a vertical landing door tag detection method according to an embodiment of the invention;
FIG. 5 schematically shows a door tag detection diagram according to an embodiment of the present invention;
FIG. 6 schematically shows a diagram of a yolov3 web framework model, according to an embodiment of the invention;
FIG. 7 schematically illustrates a door tag state diagram corresponding to a first state result according to one embodiment of the present invention;
FIG. 8 schematically illustrates a door tag state diagram corresponding to a second state result in accordance with one embodiment of the present invention;
FIG. 9 schematically illustrates a door tag state diagram corresponding to a third state result in accordance with one embodiment of the present invention;
FIG. 10 schematically illustrates a gate label state diagram corresponding to a fourth state result, in accordance with one embodiment of the present invention;
FIG. 11 schematically illustrates a door tag state diagram corresponding to a fifth state result in accordance with one embodiment of the present invention;
Detailed Description
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments will be briefly described below. It is obvious that the drawings in the following description are only some embodiments of the invention, and that for a person skilled in the art, other drawings can be derived from them without inventive effort.
In describing embodiments of the present invention, the terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like are used in an orientation or positional relationship that is based on the orientation or positional relationship shown in the associated drawings, which is for convenience and simplicity of description only, and does not indicate or imply that the referenced device or element must have a particular orientation, be constructed and operated in a particular orientation, and thus, the above-described terms should not be construed as limiting the present invention.
The present invention is described in detail below with reference to the drawings and the specific embodiments, which are not repeated herein, but the embodiments of the present invention are not limited to the following embodiments.
As shown in fig. 1, the door tags in elevators, which are to be attached to the car doors at a height of typically 1.5-1.8 meters, are capable of warning passengers. If the door opening mode of the straight elevator door is a center opening door (double door), namely the door is opened from the middle to two sides, the detection and adhesion mode is that one door label is attached to the left car door, and the other door label is attached to the right car door, the spacing distance between the two door labels is relatively short, and the reference is shown in fig. 2; if the door opening mode is a side opening door (single door opening), one door label is pasted on the left side of the folding position of the two car doors, the other door label is pasted on the fan-out position of the door, and the spacing distance between the two door labels is relatively far. All tags are clearly identifiable under night emergency lighting conditions.
Referring to fig. 3 and 4, according to an embodiment of the present invention, a method for detecting a vertical ladder gate label based on deep learning includes:
s1, acquiring a real-time image of an elevator to be detected, judging whether the elevator is in a door closing state, and if so, triggering an image analysis request;
s2, intercepting an image of the elevator according to the triggering time of the image analysis request;
s3, detecting the door tags according to the image, if the door tags exist and the number of the door tags is greater than or equal to 2, comparing the spacing distance between the door tags with a preset threshold value, and if the spacing distance between the door tags is smaller than the threshold value, generating a target picture frame containing the door tags;
and S4, judging whether the position of the image pickup equipment for acquiring the image deviates according to the position of the target picture frame in the image, if not, sending a first state result, and otherwise, sending a second state result.
According to one embodiment of the invention, a newly online elevator is taken as an example for explanation, all newly online elevators installed and debugged are stored in the cloud platform, and since the newly online elevator does not pass through an automatic picture frame, a relevant elevator ID can be issued through the cloud platform, and the corresponding camera shooting device in the elevator detects the door tag. The camera equipment starts to acquire real-time images in the elevator car, and the door opening and closing state of the elevator needs to be judged. And if the elevator is not closed, the door tag detection is not carried out, the loop waits for the preset interval time (such as 30S) to acquire at least one image of the condition that the elevator is closed again, and if the elevator is closed in the image, the image analysis request is triggered.
According to one embodiment of the invention, after receiving the image analysis request, the image in the elevator is captured and stored according to the triggering time of the elevator.
Referring to fig. 1, 2 and 5, according to one embodiment of the present invention, a door tag detection is performed on a saved image. For the double door example, label detection is performed by using a network framework model of YOLOv 3. In this embodiment, the network input size 416 x 416 of the network framework model. Because the door label is smaller, and the label has a certain distance from the camera, the depth of field is larger. On the 416 x 416 image, the size of the gate label is 10 x 8-26 x 16 pixel points, and the gate label belongs to small target detection. By adopting the YOLOv3 network framework model, detection can be performed according to different feature scales, detection of the small label targets is facilitated, the training data set can be clustered, and the small target detection performance is enhanced. In the present embodiment, referring to fig. 6, the image to be detected of 416 x 416 is input, the convolution of the first layer 3 x 3 has 32 channels, the original features are extracted, and the second layer 64 convolutions 3 x 3 are subjected to the first down-sampling. And extracting target feature information through 23 reduce blocks to obtain 13 × 13 features. Inputting 13 × 13 features to make a Convolution Set, wherein 512 convolutions of 1 × 1 are used for reducing dimension, reducing the number of channels, performing 1024 convolutions of 3 × 3 on the convolutions of 1 × 1 to extract features, and making 1 × 1 Convolution to reduce dimension after extracting the features, wherein 512 channels are provided. And performing the second 3 x 3 convolution on 512 channels and then reducing the dimension output. After the Convolution Set is done, dimension reduction is carried out on 256 channels through 1 × 1 Convolution, and a tensor with the same scale is spliced with the feature graph of 26 × 26 by Concat. And then, performing Convolution Set again, and performing 1 × 1 dimensionality reduction to 128 channels to perform upsampling and performing Concat splicing on the feature map of 52 × 52 to obtain the tensor with the same scale. After splicing, performing Convolution Set and 3 × 3 Convolution feature extraction once, and finally outputting feature numbers of a class of targets for detecting the labels in a dimensionality reduction mode.
According to an embodiment of the present invention, the step of generating the target frame includes:
s31, generating a label picture frame matched with the shape of the door label based on the acquired door label; as shown in fig. 5, according to an embodiment of the present invention, after the detection of the gate tags, if the number of the gate tags is greater than or equal to 2, a pair of standard gate tags in the image is obtained. And inputting the detected coordinates of the label frame, and further forming a label frame matched with the shape of the label around each label. In this embodiment, if the door tag is a rectangle, the tag frame is also a rectangle.
And S32, generating a target picture frame based on the label picture frame. In the present embodiment, the generated tag frame is used to determine the distance between a pair of gate tags and compare the determined distance with a threshold. Referring to fig. 2, in contrast to a double-door elevator, a pair of tags is respectively attached to two elevator doors according to the specification, and the distance between the two door tags is relatively small. Therefore, if the separation distance between the door tags is less than the threshold value. Then selecting four vertexes on the basis of the label picture frame, wherein the connecting lines of the four vertexes can form a rectangle containing the two gate labels, a point is respectively taken up and down on the center line of the rectangle according to half of the height of the gate label, so that six vertexes can be obtained, and the six vertexes are sequentially covered to form a target picture frame capable of containing the two gate labels, namely the target picture frame is a hexagonal picture frame. The target picture frame is set through the method, and the door tag can be effectively contained in the generated target picture frame.
According to one embodiment of the present invention, after the drawing of the target frame is completed, it is determined whether the position of the target frame in the image is deviated to the leftmost side or the rightmost side, that is, whether the position of the image capturing apparatus capturing the image is deviated according to the position of the target frame in the image. When the newly-on-line elevator is provided with the camera equipment and posts the door tag, the relative positions of the door tag and the camera equipment are already adjusted, and if the position deviation of the target picture frame in the image is found to be large after the detection is finished, the position deviation of the camera equipment can be known, so that the monitoring visual angle is not in the central area of the elevator car. Therefore, by detecting that the position of the target frame in the image is not shifted (see fig. 7), outputting a first state result to the cloud platform; by detecting that the position of the target frame in the image is not shifted (see fig. 8), the second state result is written to the cloud platform.
Referring to fig. 9, in step S3, if there are gate tags and the number of the gate tags is greater than or equal to 2, the distance between the gate tags is compared with a preset threshold, and if the distance between the gate tags is greater than the threshold, an object frame is generated for the gate tag closest to the middle position of the image, and a third status result is sent. In the present embodiment, as described above, when the elevator is single-door, since the distance between the pair of tags is long, the target frame only needs to be generated for the door tag located at the door-open position (i.e., the middle position in the drawing). Firstly, generating a label picture frame for the gate label, selecting four vertexes on the basis of the label picture frame, connecting lines of the four vertexes to form a rectangle containing the two gate labels, respectively taking a point on the middle line of the rectangle according to half of the height of the gate label, further obtaining six vertexes, and sequentially covering the six vertexes to form a target picture frame capable of containing the gate label, namely the target picture frame is a hexagonal picture frame. The target picture frame is set through the method, and the door tag can be effectively contained in the generated target picture frame.
Referring to fig. 10, if the gate tags exist and the number is equal to 1, an object frame is generated for the gate tags and a fourth status result is issued in step S3. In this embodiment, the process of generating the target frame for the gate tag is the same as the above process, and is not described herein again.
Referring to fig. 11, in step S3, if no door tag is present, a fifth status result is issued.
According to one embodiment of the invention, the first status result, the second status result, the third status result, the fourth status result and the fifth status result are all transmitted to the cloud platform for being matched with the corresponding elevator ID. In this embodiment, the first state result indicates that the elevator is a double-door type and the door tag is normal, the second state result indicates that the camera device is deviated, the third state result indicates that the elevator is a single-door type and the door tag is normal, the fourth state result indicates that the door tag in the elevator is damaged or torn, and the fifth state result indicates that the tag in the elevator is not detected. Through the state result, the cloud platform can send the corresponding elevator ID and the state result to a technician, and the technician performs field maintenance.
According to one embodiment of the present invention, coordinates of the target frame in the image are transmitted to the image capturing apparatus. In the embodiment, the camera device identifies the on-off state of the elevator based on the positions of the target frame and the door tag, namely, the tag is not in the picture frame area when the door is opened and is in the picture frame area when the door is closed.
The foregoing is merely exemplary of particular aspects of the present invention and devices and structures not specifically described herein are understood to be those of ordinary skill in the art and are intended to be implemented in such conventional ways.
The above description is only one embodiment of the present invention, and is not intended to limit the present invention, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A method for detecting a label of a vertical ladder door based on deep learning is characterized by comprising the following steps:
s1, acquiring a real-time image of an elevator to be detected, judging whether the elevator is in a door closing state or not, and if so, triggering an image analysis request;
s2, intercepting the image of the elevator according to the triggering time of the image analysis request;
s3, detecting the gate tags according to the image, if the gate tags exist and the number of the gate tags is greater than or equal to 2, comparing the spacing distance between the gate tags with a preset threshold value, and if the spacing distance between the gate tags is smaller than the threshold value, generating a target picture frame containing the gate tags;
and S4, judging whether the position of the image pickup equipment for acquiring the image deviates according to the position of the target picture frame in the image, if not, sending a first state result, and otherwise, sending a second state result.
2. The method as claimed in claim 1, wherein in step S3, if the gate tags exist and the number is greater than or equal to 2, the distance between the gate tags is compared with a preset threshold, and if the distance between the gate tags is greater than the threshold, a target frame is generated for the gate tag closest to the middle position of the image, and a third status result is sent.
3. The method as claimed in claim 2, wherein in step S3, if the gate tags exist and the number is equal to 1, an object frame is generated for the gate tags, and a fourth status result is issued.
4. The method as claimed in claim 3, wherein in step S3, if the door tag is not present, a fifth status result is issued.
5. The method as claimed in claim 4, wherein the coordinates of the target frame in the image are transmitted to the camera.
6. The method for detecting the vertical ladder gate label based on deep learning of claim 5, wherein the first state result, the second state result, the third state result, the fourth state result and the fifth state result are transmitted to a cloud platform for recording;
the elevator is characterized in that the first state result shows that the elevator is of a double-door type and the door label is normal, the second state result shows that the camera equipment deviates, the third state result shows that the elevator is of a single-door type and the door label is normal, the fourth state result shows that the door label in the elevator is damaged or torn, and the fifth state result shows that the label is not detected in the elevator.
7. The method as claimed in any one of claims 1 to 3, wherein the step of generating the target frame comprises:
s31, generating a label picture frame matched with the door label in shape based on the obtained door label;
and S32, generating the target picture frame based on the label picture frame.
8. The method as claimed in claim 7, wherein the target frame is a hexagonal frame.
9. The method as claimed in claim 5, wherein the camera device identifies the open/close state of the elevator based on the target frame and the position of the door tag.
10. The method for detecting the vertical ladder gate label based on deep learning of claim 1, wherein in the step of detecting the gate label according to the image, a network framework model of YOLOv3 is adopted to detect the image.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910828841.4A CN110589648B (en) | 2019-09-03 | 2019-09-03 | Vertical ladder door label detection method based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910828841.4A CN110589648B (en) | 2019-09-03 | 2019-09-03 | Vertical ladder door label detection method based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110589648A CN110589648A (en) | 2019-12-20 |
CN110589648B true CN110589648B (en) | 2021-04-16 |
Family
ID=68857203
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910828841.4A Active CN110589648B (en) | 2019-09-03 | 2019-09-03 | Vertical ladder door label detection method based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110589648B (en) |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2017124906A (en) * | 2016-01-13 | 2017-07-20 | 東芝エレベータ株式会社 | Elevator system |
CN107187980A (en) * | 2017-05-26 | 2017-09-22 | 杭州光珀智能科技有限公司 | The detection method and detection means of elevator door folding in a kind of lift appliance |
CN109592528A (en) * | 2019-01-14 | 2019-04-09 | 浙江新再灵科技股份有限公司 | The elevator safety marker localization method and system of view-based access control model word band |
CN109896386A (en) * | 2019-03-08 | 2019-06-18 | 浙江新再灵科技股份有限公司 | A kind of detection method and system that the elevator door based on computer vision technique switchs repeatedly |
CN110127479A (en) * | 2019-04-17 | 2019-08-16 | 浙江工业大学 | A kind of elevator door switch method for detecting abnormality based on video analysis |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2011002339A (en) * | 2009-06-18 | 2011-01-06 | Panasonic Electric Works Co Ltd | Object detection device |
CN102616613B (en) * | 2012-03-23 | 2015-02-25 | 佛山市智邦电子科技有限公司 | Elevator monitoring system |
CN104777521A (en) * | 2015-03-24 | 2015-07-15 | 广州市地下铁道总公司 | Binocular-vision-based detection system for foreign matter between train door and platform shield gate, as well as detection method for detection system |
-
2019
- 2019-09-03 CN CN201910828841.4A patent/CN110589648B/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2017124906A (en) * | 2016-01-13 | 2017-07-20 | 東芝エレベータ株式会社 | Elevator system |
CN107187980A (en) * | 2017-05-26 | 2017-09-22 | 杭州光珀智能科技有限公司 | The detection method and detection means of elevator door folding in a kind of lift appliance |
CN109592528A (en) * | 2019-01-14 | 2019-04-09 | 浙江新再灵科技股份有限公司 | The elevator safety marker localization method and system of view-based access control model word band |
CN109896386A (en) * | 2019-03-08 | 2019-06-18 | 浙江新再灵科技股份有限公司 | A kind of detection method and system that the elevator door based on computer vision technique switchs repeatedly |
CN110127479A (en) * | 2019-04-17 | 2019-08-16 | 浙江工业大学 | A kind of elevator door switch method for detecting abnormality based on video analysis |
Also Published As
Publication number | Publication date |
---|---|
CN110589648A (en) | 2019-12-20 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110136447B (en) | Method for detecting lane change of driving and identifying illegal lane change | |
CN105844257B (en) | The early warning system and method for road sign are missed based on machine vision travelling in fog day | |
CN106980829B (en) | Abnormal behaviour automatic testing method of fighting based on video analysis | |
CN104847211B (en) | Auxiliary system for safety of platform safety doors and train door sections and implementation method thereof | |
CN108280435A (en) | Passenger abnormal behavior recognition method based on human body posture estimation | |
CN106241533A (en) | Elevator occupant's comprehensive safety intelligent control method based on machine vision | |
CN112347873B (en) | Ladder control method | |
CN109670404A (en) | A kind of road ponding image detection method for early warning based on mixed model | |
CN112016528B (en) | Behavior recognition method and device, electronic equipment and readable storage medium | |
CN106006266A (en) | Machine vision establishment method applied to elevator safety monitoring | |
CN106548147A (en) | A kind of quick noise robustness image foreign matter detection method and TEDS systems | |
CN111339905A (en) | CIM well lid state visual detection system based on deep learning and multi-view angle | |
CN111951410A (en) | Full-automatic ticket checking and checking method and system for park team tickets | |
WO2012005461A2 (en) | Method for automatically calculating information on clouds | |
CN110263623B (en) | Train climbing monitoring method, device, terminal and storage medium | |
CN107644206A (en) | A kind of road abnormal behaviour action detection device | |
CN110255318B (en) | Method for detecting idle articles in elevator car based on image semantic segmentation | |
CN110589647A (en) | Method for real-time fault detection and prediction of elevator door through monitoring | |
CN110175533A (en) | Overpass traffic condition method of real-time, device, terminal and storage medium | |
CN105930774A (en) | Automatic bridge bolt come-off identification method based on neural network | |
CN110070061A (en) | A kind of passengers quantity projectional technique and device | |
CN109001210B (en) | System and method for detecting aging and cracking of sealing rubber strip of civil air defense door | |
CN112216010A (en) | Campus security management method, system, equipment and storage medium | |
CN109409191A (en) | A kind of zebra stripes vehicle evacuation detection method and system based on machine learning | |
CN113450573A (en) | Traffic monitoring method and traffic monitoring system based on unmanned aerial vehicle image recognition |
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 |