CN113139458A - Method and system for identifying opening and closing states of parking garage roller shutter - Google Patents
Method and system for identifying opening and closing states of parking garage roller shutter Download PDFInfo
- Publication number
- CN113139458A CN113139458A CN202110432452.7A CN202110432452A CN113139458A CN 113139458 A CN113139458 A CN 113139458A CN 202110432452 A CN202110432452 A CN 202110432452A CN 113139458 A CN113139458 A CN 113139458A
- Authority
- CN
- China
- Prior art keywords
- convolution
- parking garage
- input image
- identifying
- layer
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 238000005096 rolling process Methods 0.000 claims abstract description 49
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 32
- 230000005540 biological transmission Effects 0.000 claims description 13
- 230000009467 reduction Effects 0.000 claims description 13
- 238000004891 communication Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 7
- 239000000126 substance Substances 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 5
- 230000001960 triggered effect Effects 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims 1
- 238000012544 monitoring process Methods 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 7
- 230000003068 static effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- General Health & Medical Sciences (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Operating, Guiding And Securing Of Roll- Type Closing Members (AREA)
Abstract
The application discloses a method and a system for identifying the opening and closing state of a parking garage roller shutter, which can improve the monitoring efficiency of the parking garage, wherein the method comprises the steps of acquiring an image which is shot by a camera arranged in a parking garage and is right opposite to the parking garage roller shutter and taking the image as an input image; identifying lines along a first direction in the input image and counting the number of the lines by adopting a first convolutional neural network structure, or identifying lines along a second direction in the input image and counting the number of the lines by adopting a second convolutional neural network structure; the opening and closing state of the rolling door is identified according to the number of lines along the first direction or the second direction, the included angle formed by the first direction and the horizontal direction is smaller than a first preset angle, and the included angle formed by the second direction and the vertical direction is smaller than a second preset angle.
Description
Technical Field
The application relates to the technical field of images, in particular to a method and a system for identifying the opening and closing state of a parking garage roller shutter.
Background
The rolling door is also called "rolling door", which is a door that is formed by connecting a plurality of movable joint sheets in series, and rotates up and down in a fixed slideway by taking a scroll above the door as a center, as shown in fig. 1, a schematic view of a common rolling door at present is provided.
The parking garage is one of the places where the rolling gate is applicable. At present, people can install a camera in a parking garage to monitor the safety state of the parking garage.
However, at present, it is mainly determined that the parking garage is in a safe state or is illegally intruded by manually observing videos uploaded by the cameras, so that the efficiency is low.
Disclosure of Invention
Based on this, in order to solve or improve the problems of the prior art, the application provides a method and a system for identifying the opening and closing state of a roller shutter of a parking garage, which can improve the monitoring efficiency of the parking garage.
In a first aspect, a method for identifying the opening and closing state of a parking garage roller shutter is provided, which comprises the following steps:
acquiring an image which is shot by a camera arranged in a parking garage and is right opposite to a rolling gate of the parking garage, and taking the image as an input image;
identifying lines along a first direction in the input image and counting the number of the lines by adopting a first convolutional neural network structure, or identifying lines along a second direction in the input image and counting the number of the lines by adopting a second convolutional neural network structure;
according to the quantity discernment along the lines of first direction or along the second direction the on off state of shutter door, the contained angle that first direction and horizontal direction formed is less than first preset angle, the contained angle that second direction and vertical direction formed is less than the second and presets the angle.
In one embodiment, the step of identifying a line in the input image along a first direction using a first convolutional neural network includes:
normalizing the pixel value of each pixel point of the input image;
carrying out convolution operation on the input image by using an NxN convolution kernel, wherein each convolution operation is carried out by multiplying odd-numbered lines of the NxN convolution kernel by corresponding pixel points of the input image and then adding products;
a line along a first direction in the input image on which the convolution operation is performed is identified.
In one embodiment, the network structure of the first convolutional neural network comprises a plurality of convolutional layers and downsampling layers which are sequentially connected in sequence along the information transmission direction, and a full connection layer; the input end of the first convolution layer is used for accessing an input image, the output end of the last down-sampling layer is connected with the input end of the full-connection layer, and the rest convolution layers except the first convolution layer are all transverse-line convolution layers; wherein the content of the first and second substances,
the first convolution layer is used for extracting features in the input image;
each downsampling layer is used for respectively reducing the dimension of the features extracted from each convolution layer;
each transverse line convolution layer is used for extracting line features along the first direction, convolution operation is carried out on the features subjected to dimension reduction by utilizing convolution kernel, multiplication is carried out on corresponding pixel points of the input image by utilizing odd-numbered lines of the convolution kernel during each convolution operation, and then products are added;
and the output end of the full connection layer is used for outputting the probability whether each row of features subjected to dimensionality reduction is a line along the first direction.
In one embodiment, the step of identifying a line in the input image along the second direction using the second convolutional neural network comprises:
normalizing the pixel value of each pixel point of the input image;
carrying out convolution operation on the input image by using an NxN convolution kernel, wherein each convolution operation is carried out by multiplying odd columns of the NxN convolution kernel by corresponding pixel points of the input image and then adding products;
a line in a second direction in the input image on which the convolution operation is performed is identified.
In one embodiment, the network structure of the second convolutional neural network comprises a plurality of convolutional layers and downsampling layers which are sequentially connected in sequence along the information transmission direction, and a full connection layer; the input end of the first convolution layer is used for accessing an input image, the output end of the last downsampling layer is connected with the input end of the full-connection layer, and the rest convolution layers except the first convolution layer are vertical line convolution layers; wherein the content of the first and second substances,
the first convolution layer is used for extracting features in the input image;
each downsampling layer is used for respectively reducing the dimension of the features extracted from each convolution layer;
each vertical line convolution layer is used for extracting line features along the second direction, convolution operation is carried out on the features subjected to dimension reduction by utilizing convolution kernel, multiplication is carried out on the odd-numbered columns of the convolution kernel and corresponding pixel points of the input image during each convolution operation, and then products are added;
and the output end of the full connection layer is used for outputting the probability whether each row of features subjected to dimensionality reduction is a line along the second direction.
In one embodiment, the convolution kernels applied to each transverse-line convolution layer are sequentially reduced along the information transmission direction by at least one time.
In one embodiment, the convolution kernels applied to each vertical line convolution layer are sequentially reduced along the information transmission direction by at least one time.
In one embodiment, a plurality of strip-shaped structures along the horizontal direction are arranged on the rolling gate;
the step of recognizing the opening and closing state of the roll shutter based on the number of the straight lines in the first direction includes: if the number of the straight lines in the first direction is larger than a first preset number threshold, judging that the rolling door is in a closed state, otherwise, judging that the rolling door is in an open state;
the step of recognizing the opening and closing state of the roll shutter based on the number of the straight lines in the second direction includes: and if the number of the straight lines in the second direction is larger than a second preset number threshold, judging that the rolling door is in an open state, otherwise, judging that the rolling door is in a closed state.
According to the method and the system for identifying the opening and closing states of the parking garage roller shutter door, the opening and closing states of the parking garage roller shutter door are identified by identifying the number of lines in the image at the parking garage roller shutter door by adopting an image identification method, so that the parking garage monitoring efficiency is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments will be briefly introduced below. It is to be understood that the drawings in the following description are illustrative only and are not restrictive of the invention.
FIG. 1 is a schematic view of a conventional rolling shutter door;
FIG. 2 is a schematic flow chart illustrating a method for identifying the opening and closing status of a parking garage roller shutter according to an embodiment of the present application;
FIG. 3 is a schematic structural view of a shutter door according to an embodiment of the present application;
FIG. 4 is a diagram illustrating a convolution operation performed on an input image according to an embodiment of the present application;
FIG. 5 is a schematic diagram of a network structure of a first convolutional neural network according to an embodiment of the present application;
FIG. 6 is a diagram illustrating a convolution operation performed on an input image according to another embodiment of the present application;
FIG. 7 is a schematic diagram of a second convolutional neural network according to an embodiment of the present application;
FIG. 8 is a schematic flow chart illustrating a method for identifying the open/close status of a parking garage roller shutter according to another embodiment of the present application;
FIG. 9 is a schematic flow chart illustrating a method for identifying the open/close status of a parking garage roller shutter according to still another embodiment of the present application;
FIG. 10 is a schematic diagram of a system for identifying the open and closed states of a parking garage roller door in accordance with one embodiment of the present application;
fig. 11 is a schematic structural diagram of an apparatus for recognizing the opening and closing states of a parking garage roller shutter according to an embodiment of the present application.
Detailed Description
The technical solutions in the embodiments of the present application are clearly and completely described below with reference to the accompanying drawings, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
As described in the background art, it is currently mainly determined that the parking garage is in a safe state or is illegally intruded by manually observing videos uploaded by the cameras, which is inefficient.
The embodiment of the application provides a method for identifying the opening and closing states of a rolling door of a parking garage, and the monitoring efficiency of the parking garage can be improved.
Please refer to fig. 2, which is a flowchart illustrating a method for identifying an opening/closing status of a parking garage roller shutter according to an embodiment, including the following steps:
and step 206, identifying the opening and closing state of the roller shutter door according to the number of the lines along the first direction or the second direction, wherein the included angle formed between the first direction and the horizontal direction is smaller than a first preset angle, and the included angle formed between the second direction and the vertical direction is smaller than a second preset angle.
It will be understood that the typical feature of a conventional rolling shutter door is that a plurality of strips are arranged in a horizontal direction (as shown in fig. 3), and the ground outside the rolling shutter door has a vertical line due to the possible presence of plants, utility poles, etc. Therefore, the method for identifying the opening and closing state of the parking garage roller shutter can identify the opening and closing state of the parking garage roller shutter by identifying the number of straight lines in the image at the parking garage roller shutter by adopting an image identification method, and improves the monitoring efficiency of the parking garage.
Working errors may exist in the strip-shaped structures along the horizontal direction of the rolling gate in a real scene, and although plants, telegraph poles and the like outside the rolling gate in the real scene are close to the vertical direction, deviations also exist inevitably. In addition, when the camera performs image processing or shooting, straight lines in the horizontal direction and the vertical direction in the real scene are mistakenly processed into oblique lines. Therefore, when the image recognition is adopted in the embodiment, the straight line forming the preset angle with the horizontal direction and the straight line forming the preset angle with the vertical direction are recognized, and the influence caused by the reasons can be reduced.
In the embodiment of the present application, the direction of the flat ground can be regarded as the horizontal direction. The direction perpendicular to the flat ground can be considered as the vertical direction.
For example, the camera may be installed on the ceiling of the parking garage, and in particular, the camera is installed so that the shooting direction is opposite to the shutter door.
For example, the first preset angle may be within ± 20 °, that is, all the lines in the image forming an angle smaller than 20 ° with the horizontal direction are identified. The second preset angle can be within +/-10 degrees, namely, all straight lines forming an included angle smaller than 10 degrees with the vertical direction in the image are recognized.
In one embodiment, the step of identifying a line in the input image along a first direction using a first convolutional neural network includes:
normalizing the pixel value of each pixel point of the input image;
carrying out convolution operation on the input image by using an NxN convolution kernel, wherein each convolution operation is carried out by multiplying odd-numbered lines of the NxN convolution kernel by corresponding pixel points of the input image and then adding products;
a line along a first direction in the input image on which the convolution operation is performed is identified.
For example, referring to fig. 4, 5 × 5 convolution kernel may be used to perform convolution operation on the input image, and the input image is convolved by using the 1 st, 3 rd and 5 th rows of the 5 × 5 convolution kernel, and the 2 nd and 4 th rows are omitted, and the inventors have found that, as a result, only 3 × 5 to 15 times of computation are required for each kernel computation, and the computation amount is reduced by 40% compared with the case where all rows of the 5 × 5 convolution kernel are involved in convolution operation.
Therefore, by the direction recognition of the line in the first direction in the input image of the present embodiment, the amount of computation can be reduced.
In specific implementation, the method of the embodiment is implemented by using a microprocessor with low computational power, so that the hardware cost can be saved.
In one embodiment, referring to fig. 5, the network structure of the first convolutional neural network includes a plurality of convolutional layers and downsampling layers sequentially connected in sequence along the information transmission direction, and a full connection layer; the input end of the first convolution layer is used for accessing an input image, the output end of the last down-sampling layer is connected with the input end of the full-connection layer, and the rest convolution layers except the first convolution layer are all transverse-line convolution layers; wherein the content of the first and second substances,
the first convolution layer is used for extracting features in the input image;
each downsampling layer is used for respectively reducing the dimension of the features extracted from each convolution layer;
each transverse line convolution layer is used for extracting line features along the first direction, convolution operation is carried out on the features subjected to dimension reduction by utilizing convolution kernel, multiplication is carried out on corresponding pixel points of the input image by utilizing odd-numbered lines of the convolution kernel during each convolution operation, and then products are added;
and the output end of the full connection layer is used for outputting the probability whether each row of features subjected to dimensionality reduction is a line along the first direction.
It will be appreciated that since the network structure includes a plurality of down-sampling layers along the transmission direction for reducing the dimension of the feature, the dimension-reduced features output by the plurality of down-sampling layers are sequentially reduced. Therefore, it is proposed that in one embodiment, in the information transmission direction of the network structure of the first convolutional neural network, the convolutional kernels used by the respective horizontal convolutional layers are sequentially reduced, for example, each time, the reduction may be 1 time, and may also be greater than 1 time. In this way, it is possible to adapt to the change in the size of the input image. In other embodiments, the convolution kernels used for the horizontal convolution layers may have the same size.
It should be noted that, in the implementation of the convolution kernel with the same size or different sizes, the size of the convolution kernel used in the last horizontal convolution layer needs to be smaller than the size of the feature output by the last downsampling layer. Illustratively, the feature size of the output of the last downsampling layer is 10 × 10, and the sizes of convolution kernels adopted by the last horizontal line convolution layer can be 1 × 1, 3 × 3, 5 × 5, 7 × 7 and 9 × 9.
The following specific embodiment of the network structure of the second convolutional neural network is similar to the network structure of the first convolutional neural network, and is not described in detail again.
In one embodiment, the step of identifying a line in the input image along the second direction using the second convolutional neural network comprises:
normalizing the pixel value of each pixel point of the input image;
carrying out convolution operation on the input image by using an NxN convolution kernel, wherein each convolution operation is carried out by multiplying odd columns of the NxN convolution kernel by corresponding pixel points of the input image and then adding products;
a line in a second direction in the input image on which the convolution operation is performed is identified.
For example, referring to fig. 6, 5 × 5 convolution kernel may be used to perform convolution operation on the input image, and the input image is convolved by columns 1, 3, and 5 of the 5 × 5 convolution kernel, and columns 2 and 4 are omitted, and the inventors have found that, as a result, only 3 × 5 to 15 times of computation is required for each kernel computation, and the computation amount is reduced by 40% compared with the case where all columns of the 5 × 5 convolution kernel are involved in convolution operation.
Therefore, by the direction recognition of the line in the second direction in the input image of the present embodiment, the amount of computation can be reduced.
In specific implementation, the method of the embodiment is implemented by using a microprocessor with low computational power, so that the hardware cost can be saved.
In one embodiment, referring to fig. 7, the network structure of the second convolutional neural network includes a plurality of convolutional layers and downsampling layers sequentially connected in sequence along the information transmission direction, and a full connection layer; the input end of the first convolution layer is used for accessing an input image, the output end of the last downsampling layer is connected with the input end of the full-connection layer, and the rest convolution layers except the first convolution layer are vertical line convolution layers; wherein the content of the first and second substances,
the first convolution layer is used for extracting features in the input image;
each downsampling layer is used for respectively reducing the dimension of the features extracted from each convolution layer;
each vertical line convolution layer is used for extracting line features along the second direction, convolution operation is carried out on the features subjected to dimension reduction by utilizing convolution kernel, multiplication is carried out on the odd-numbered columns of the convolution kernel and corresponding pixel points of the input image during each convolution operation, and then products are added;
and the output end of the full connection layer is used for outputting the probability whether each row of features subjected to dimensionality reduction is a line along the second direction.
In one embodiment, the convolution kernels applied to the vertical line convolution layers are sequentially reduced along the information transmission direction of the network structure of the second convolutional neural network.
Other specific embodiments of the network structure of the second convolutional neural network are similar to the network structure of the first convolutional neural network, and are not described again.
In some embodiments, the rolling shutter door adopted by the present application is provided with a plurality of strip-shaped structures along the horizontal direction. Whether the shutter door is closed can be determined by recognizing the number of straight lines in the image in the first direction (the first direction forms an angle smaller than a first preset angle with the horizontal direction). Specifically, referring to fig. 8, the step of recognizing the opening and closing state of the roll shutter according to the number of lines in the first direction includes:
It can be understood that the cross lines are less and less during the opening process of the rolling door. Therefore, specifically, if the number of the first direction straight lines detected within the preset time is gradually reduced from being greater than the first preset number threshold to being less than the first preset number threshold and is maintained for more than the preset time, it is determined that the shutter door is in the open state.
In other embodiments, there may be many vertical line scenes outside the roller shutter door, so whether the roller shutter door is closed can be identified by the number of straight lines in the second direction. Specifically, referring to fig. 9, the step of recognizing the opening and closing state of the roll shutter door according to the number of lines in the second direction includes:
To further improve the identification accuracy, in some embodiments the step of acquiring an image taken by a camera mounted in the parking garage directly opposite the parking garage roller gate is acquiring a plurality of consecutive images; the shooting direction of the camera installed in the parking garage is fixed, namely the shooting area of the camera is always kept unchanged; if the number of the straight lines in the first direction is greater than the first preset number threshold, the step of determining that the shutter door is in the closed state includes: if the number of the straight lines in the first direction in the multiple continuous images is larger than a first preset number threshold value and the positions of the straight lines in the first direction in the multiple continuous images are the same, it is indicated that the straight lines in the first direction in the images are static and are more than the preset number, the rolling door is judged to be in a closed state if the photographed images are likely to be the images of the rolling door, and if the number of the straight lines in the first direction in the multiple continuous images is larger than the first preset number threshold value but the positions of the straight lines in the first direction in more than two images are different, it is indicated that the straight lines in the first direction are shaken or moved, the photographed images are likely to be moving objects rather than the static rolling door images in the closed state, and the rolling door is judged to be in an open state.
In addition, in some embodiments, if the number of the straight lines in the first direction is greater than the first preset number threshold, the step of determining that the rolling door is in the closed state may further include: if the number of the straight lines in the first direction is larger than the first preset number threshold value, and the adjacent straight lines in the first direction are parallel and equidistant, it is indicated that the probability that the straight lines belong to the rolling door is high, and/or the width of the straight lines in the first direction is within the preset width range of the strip-shaped structure of the rolling door, and the rolling door can be judged to be in the closed state under the above conditions.
It can be understood that, in the rolling door in the real scene, the strip-shaped structures are parallel and equidistant and have a certain width, and therefore, in the embodiments, it can be determined that the rolling door is in the closed state by further identifying that the straight lines in the first direction are parallel and equidistant and/or that the straight line width is within the preset width range of the strip-shaped structure of the rolling door. Therefore, the identification accuracy of the rolling door in the closed state is improved.
In some embodiments, in the step of identifying the opening and closing state of the rolling door according to the number of straight lines along the first direction or the second direction, if the rolling door is determined to be in the open state, acquiring an image of the rolling door in the open state, and uploading the image of the rolling door in the open state to a terminal for the parking garage user to log in permission; and when the terminal receives the image at the shutter door, popping up an alarm interface, generating an alarm button on the alarm interface, and if the alarm button is triggered, sending alarm information to a terminal of a public security department by the terminal. Therefore, the property safety of the parking garage is improved.
The embodiment of the present application further provides a system for identifying the opening and closing states of a parking garage roller shutter door, please refer to fig. 10, which includes a processor 630, a memory, and a camera 610 installed in a parking garage, wherein the shooting direction of the camera is opposite to the parking garage roller shutter door 620; the camera 610 is connected with the processor 630;
the camera 610 is used for shooting an image facing the parking garage roller shutter door 620;
the memory stores a computer program which, when executed by the processor 630, causes the processor 630 to perform the steps of the method for identifying the opening and closing states of a parking garage roller shutter according to any of the above embodiments.
In one embodiment, the system for identifying the opening and closing states of the parking garage shutter door further comprises a communication module and a terminal 640 for parking garage user right login, wherein the camera 610 is in communication connection with the terminal 640 through the communication module; the processor 630 is configured to trigger the camera 610 to upload the acquired image of the open state of the shutter door to the terminal 640 for parking garage user right login through the communication module when determining that the shutter door is in the open state; when the terminal 640 receives the image at the shutter door, an alarm interface is popped up, an alarm button is generated on the alarm interface, and if the alarm button is triggered, the terminal sends alarm information to the terminal of the public security department.
The terminal 640 may be implemented in various forms. For example, the terminal 640 described in the present application may include a mobile terminal such as a mobile phone, a tablet computer, a notebook computer, a palmtop computer, a Personal Digital Assistant (PDA), a Portable Media Player (PMP), a navigation device, a wearable device, a smart band, a pedometer, and the like, and a fixed terminal such as a Digital TV, a desktop computer, and the like.
The present application refers to the foregoing embodiments, and the specific limitations of the system for identifying the open/close state of a parking garage roller shutter are not repeated herein.
Fig. 11 is a block diagram of an apparatus for recognizing the open/close state of a parking garage roller shutter according to an embodiment. As shown in fig. 11, the apparatus for recognizing the opening and closing state of the parking garage shutter comprises:
the image acquisition module 710 is used for acquiring an image which is shot by a camera arranged in the parking garage and is right opposite to the parking garage rolling door, and taking the image as an input image;
a straight line recognition module 720, configured to recognize lines in the input image along a first direction by using a first convolutional neural network structure and count the number of the lines, or recognize lines in the input image along a second direction by using a second convolutional neural network structure and count the number of the lines;
the on-off state recognition module 730 of the rolling gate is used for recognizing the on-off state of the rolling gate according to the number of lines along the first direction or the second direction, an included angle formed by the first direction and the horizontal direction is smaller than a first preset angle, and an included angle formed by the second direction and the vertical direction is smaller than a second preset angle.
According to the embodiment of the application, the opening and closing states of the rolling doors are identified by identifying the number of lines in the images at the rolling doors of the parking garage through the image identification method, so that the parking garage monitoring efficiency is improved.
The division of each module in the device for identifying the opening and closing state of the parking garage roller shutter is only used for illustration, and in other embodiments, the device for identifying the opening and closing state of the parking garage roller shutter can be divided into different sub-modules as required to complete all or part of the functions of the device for identifying the opening and closing state of the parking garage roller shutter.
For the specific definition of the means for identifying the opening and closing state of the parking garage roller shutter, reference is made to the above definition of the method for identifying the opening and closing state of the parking garage roller shutter, which is not described herein again. The modules in the device for identifying the opening and closing states of the parking garage shutter can be completely or partially realized by software, hardware and a combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.
The implementation of each module in the device for identifying the opening and closing states of the parking garage roller shutter provided in the embodiment of the application can be in the form of a computer program. The computer program may be run on a terminal or a server. The program modules constituted by the computer program may be stored on the memory of the terminal or the server. Which when executed by a processor, performs the steps of the method described in the embodiments of the present application.
The present application further proposes an electronic device comprising a processor and a memory, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method for identifying the opening and closing states of the parking garage shutter according to any of the above embodiments.
The electronic device may be implemented in various forms. For example, the electronic devices described in the present application may include mobile terminals such as a mobile phone, a tablet computer, a notebook computer, a palmtop computer, a Personal Digital Assistant (PDA), a Portable Media Player (PMP), a navigation device, a wearable device, a smart band, a pedometer, and the like, and fixed terminals such as a Digital TV, a desktop computer, and the like.
The technical features of the embodiments described above may be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the embodiments described above are not described, but should be considered as being within the scope of the present specification as long as there is no contradiction between the combinations of the technical features.
The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.
Claims (10)
1. A method for identifying the opening and closing states of a parking garage roller shutter door is characterized by comprising the following steps:
acquiring an image which is shot by a camera arranged in a parking garage and is right opposite to a rolling gate of the parking garage, and taking the image as an input image;
identifying lines along a first direction in the input image and counting the number of the lines by adopting a first convolutional neural network structure, or identifying lines along a second direction in the input image and counting the number of the lines by adopting a second convolutional neural network structure;
according to the quantity discernment along the lines of first direction or along the second direction the on off state of shutter door, the contained angle that first direction and horizontal direction formed is less than first preset angle, the contained angle that second direction and vertical direction formed is less than the second and presets the angle.
2. The method for identifying the open and closed state of a parking garage roller shutter according to claim 1, wherein said step of identifying a line in a first direction in said input image using a first convolutional neural network comprises:
normalizing the pixel value of each pixel point of the input image;
carrying out convolution operation on the input image by using an NxN convolution kernel, wherein each convolution operation is carried out by multiplying odd-numbered lines of the NxN convolution kernel by corresponding pixel points of the input image and then adding products;
a line along a first direction in the input image on which the convolution operation is performed is identified.
3. The method for identifying the opening and closing states of a parking garage roller shutter according to claim 2, wherein the network structure of the first convolutional neural network comprises a plurality of convolutional layers and downsampling layers which are sequentially connected in sequence along the information transmission direction, and a fully-connected layer; the input end of the first convolution layer is used for accessing an input image, the output end of the last down-sampling layer is connected with the input end of the full-connection layer, and the rest convolution layers except the first convolution layer are all transverse-line convolution layers; wherein the content of the first and second substances,
the first convolution layer is used for extracting features in the input image;
each downsampling layer is used for respectively reducing the dimension of the features extracted from each convolution layer;
each transverse line convolution layer is used for extracting line features along the first direction, convolution operation is carried out on the features subjected to dimension reduction by utilizing convolution kernel, multiplication is carried out on corresponding pixel points of the input image by utilizing odd-numbered lines of the convolution kernel during each convolution operation, and then products are added;
and the output end of the full connection layer is used for outputting the probability whether each row of features subjected to dimensionality reduction is a line along the first direction.
4. The method for identifying the status of a parking garage roller shutter according to claim 1, wherein said step of identifying a line in said input image along a second direction using a second convolutional neural network comprises:
normalizing the pixel value of each pixel point of the input image;
carrying out convolution operation on the input image by using an NxN convolution kernel, wherein each convolution operation is carried out by multiplying odd columns of the NxN convolution kernel by corresponding pixel points of the input image and then adding products;
a line in a second direction in the input image on which the convolution operation is performed is identified.
5. The method for identifying the opening and closing state of a parking garage roller shutter according to claim 4, wherein the network structure of the second convolutional neural network comprises a plurality of convolutional layers and downsampling layers sequentially connected in sequence along the information transmission direction, and a fully-connected layer; the input end of the first convolution layer is used for accessing an input image, the output end of the last downsampling layer is connected with the input end of the full-connection layer, and the rest convolution layers except the first convolution layer are vertical line convolution layers; wherein the content of the first and second substances,
the first convolution layer is used for extracting features in the input image;
each downsampling layer is used for respectively reducing the dimension of the features extracted from each convolution layer;
each vertical line convolution layer is used for extracting line features along the second direction, convolution operation is carried out on the features subjected to dimension reduction by utilizing convolution kernel, multiplication is carried out on the odd-numbered columns of the convolution kernel and corresponding pixel points of the input image during each convolution operation, and then products are added;
and the output end of the full connection layer is used for outputting the probability whether each row of features subjected to dimensionality reduction is a line along the second direction.
6. Method for recognizing the opening and closing state of a parking garage roller shutter according to claim 3, characterized in that the convolution kernel used by each transverse line convolution layer decreases in sequence, at least by a factor of two at a time, in the direction of information transmission.
7. The method for identifying the open-close state of a parking garage roller shutter according to claim 5, characterized in that the convolution kernels applied to each vertical line convolution layer are reduced in sequence by at least one time along the information transmission direction.
8. The method for identifying the opening and closing states of a parking garage roller shutter according to claim 1, wherein said roller shutter is provided with a plurality of horizontally oriented strip-like structures;
the step of recognizing the opening and closing state of the roll shutter based on the number of the straight lines in the first direction includes: if the number of the straight lines in the first direction is larger than a first preset number threshold, judging that the rolling door is in a closed state, otherwise, judging that the rolling door is in an open state;
the step of recognizing the opening and closing state of the roll shutter based on the number of the straight lines in the second direction includes: and if the number of the straight lines in the second direction is larger than a second preset number threshold, judging that the rolling door is in an open state, otherwise, judging that the rolling door is in a closed state.
9. A system for identifying the opening and closing state of a parking garage shutter door is characterized by comprising a processor, a memory and a camera arranged in a parking garage, wherein the shooting direction of the camera is over against the parking garage shutter door; the camera is connected with the processor;
the camera is used for shooting an image which is right opposite to the parking garage roller shutter;
the memory has stored therein a computer program which, when being executed by the processor, causes the processor to carry out the steps of the method of identifying a parking garage roller door switch status according to any one of claims 1 to 8.
10. The system for identifying the opening and closing state of the parking garage roller shutter door according to claim 9, further comprising a communication module and a terminal for parking garage user right login, wherein the camera is in communication connection with the terminal through the communication module;
the processor is used for triggering the camera to upload the acquired image of the shutter in the open state to a terminal for the parking garage user to log in through the communication module when the shutter is judged to be in the open state;
the terminal is used for popping up an alarm interface when receiving the image of the roller shutter door, generating an alarm button on the alarm interface, and sending alarm information to a terminal of a public security department when the alarm button is triggered.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110432452.7A CN113139458B (en) | 2021-04-21 | Method and system for identifying opening and closing states of rolling gate of parking garage |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110432452.7A CN113139458B (en) | 2021-04-21 | Method and system for identifying opening and closing states of rolling gate of parking garage |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113139458A true CN113139458A (en) | 2021-07-20 |
| CN113139458B CN113139458B (en) | 2024-06-28 |
Family
ID=
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105156010A (en) * | 2015-08-28 | 2015-12-16 | 苏州市博得立电源科技有限公司 | Method for controlling garage roller shutter door on basis of image analysis |
| US20170284157A1 (en) * | 2016-03-29 | 2017-10-05 | Bilmar Dock And Door Service Inc. | Overhead door tilt detection package and method |
| CN110197166A (en) * | 2019-06-04 | 2019-09-03 | 西安建筑科技大学 | A kind of car body loading condition identification device and method based on image recognition |
| CN110516615A (en) * | 2019-08-29 | 2019-11-29 | 广西师范大学 | Human and vehicle shunting control method based on convolutional neural networks |
| CN111415323A (en) * | 2019-01-04 | 2020-07-14 | 阿里巴巴集团控股有限公司 | Image detection method and device and neural network training method and device |
| CN111950409A (en) * | 2020-07-31 | 2020-11-17 | 江苏大学 | Intelligent identification method and system for road marking line |
| CN112258471A (en) * | 2020-10-20 | 2021-01-22 | 成都云盯科技有限公司 | Rolling door state detection method and system |
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105156010A (en) * | 2015-08-28 | 2015-12-16 | 苏州市博得立电源科技有限公司 | Method for controlling garage roller shutter door on basis of image analysis |
| US20170284157A1 (en) * | 2016-03-29 | 2017-10-05 | Bilmar Dock And Door Service Inc. | Overhead door tilt detection package and method |
| CN111415323A (en) * | 2019-01-04 | 2020-07-14 | 阿里巴巴集团控股有限公司 | Image detection method and device and neural network training method and device |
| CN110197166A (en) * | 2019-06-04 | 2019-09-03 | 西安建筑科技大学 | A kind of car body loading condition identification device and method based on image recognition |
| CN110516615A (en) * | 2019-08-29 | 2019-11-29 | 广西师范大学 | Human and vehicle shunting control method based on convolutional neural networks |
| CN111950409A (en) * | 2020-07-31 | 2020-11-17 | 江苏大学 | Intelligent identification method and system for road marking line |
| CN112258471A (en) * | 2020-10-20 | 2021-01-22 | 成都云盯科技有限公司 | Rolling door state detection method and system |
Non-Patent Citations (3)
| Title |
|---|
| 刘熙明;杨露溪;叶志伟;: "具有火灾报警功能卷闸门电子控制系统研究与设计", 今日电子, no. 12, 15 December 2017 (2017-12-15) * |
| 周前飞: "机械式立体车库防误入预警系统研究", 《机电工程技术》, vol. 47, no. 12, 2 January 2019 (2019-01-02), pages 26 - 30 * |
| 贺娅莉;谢娟娟;: "微信控制开关在卷闸门中的应用设计", 科技风, no. 10, 10 April 2018 (2018-04-10) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110598558B (en) | Crowd density estimation method, device, electronic equipment and medium | |
| US10181090B2 (en) | Secure and private tracking across multiple cameras | |
| CN111738244B (en) | Image detection method, image detection device, computer equipment and storage medium | |
| CN110765860B (en) | Tumble judging method, tumble judging device, computer equipment and storage medium | |
| TWI738172B (en) | Video processing method and device, electronic equipment, storage medium and computer program | |
| CN112784670A (en) | Object detection based on pixel differences | |
| JP2018508875A (en) | Method and apparatus for biological face detection | |
| CN110674800B (en) | Face living body detection method and device, electronic equipment and storage medium | |
| CN110163057B (en) | Object detection method, device, equipment and computer readable medium | |
| CN111462155B (en) | Motion detection method, device, computer equipment and storage medium | |
| US11816877B2 (en) | Method and apparatus for object detection in image, vehicle, and robot | |
| CN111104841A (en) | Violent behavior detection method and system | |
| CN112884147B (en) | Neural network training method, image processing method, device and electronic equipment | |
| CN112232140A (en) | Crowd counting method and device, electronic equipment and computer storage medium | |
| Patil et al. | Global abnormal events detection in crowded scenes using context location and motion‐rich spatio‐temporal volumes | |
| CN111860590A (en) | License plate voting method and device, computer equipment and storage medium | |
| CN113139458A (en) | Method and system for identifying opening and closing states of parking garage roller shutter | |
| CN113139458B (en) | Method and system for identifying opening and closing states of rolling gate of parking garage | |
| CN112633200A (en) | Human face image comparison method, device, equipment and medium based on artificial intelligence | |
| CN115424335B (en) | Living body recognition model training method, living body recognition method and related equipment | |
| Archana et al. | To detect abnormal event at ATM system by using image processing based on IOT technologies | |
| US20220046315A1 (en) | Thinning video based on content | |
| CN110572618B (en) | Illegal photographing behavior monitoring method, device and system | |
| CN113516016A (en) | Method and system for identifying opening and closing states of parking garage roller shutter | |
| CN114419558A (en) | Fire video image identification method, fire video image identification system, computer equipment and storage medium |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20240528 Address after: Building 9, Building 203B, Phase II, Nanshan Yungu Entrepreneurship Park, No. 2 Pingshan 1st Road, Pingshan Community, Taoyuan Street, Nanshan District, Shenzhen, Guangdong Province, 518033 Applicant after: Core Computing Integrated (Shenzhen) Technology Co.,Ltd. Country or region after: China Address before: Room 1010, 10 / F, building a, Chuangzhi building, Xinjiang Software Park, 455 Kanas Hubei Road, Urumqi Economic and Technological Development Zone, 830000 Xinjiang Uygur Autonomous Region Applicant before: XINJIANG AIHUA YINGTONG INFORMATION TECHNOLOGY Co.,Ltd. Country or region before: China |
|
| GR01 | Patent grant |