CN111860256A - Security detection method and device, computer equipment and storage medium - Google Patents
Security detection method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN111860256A CN111860256A CN202010662974.1A CN202010662974A CN111860256A CN 111860256 A CN111860256 A CN 111860256A CN 202010662974 A CN202010662974 A CN 202010662974A CN 111860256 A CN111860256 A CN 111860256A
- Authority
- CN
- China
- Prior art keywords
- detected
- image
- target
- safety
- cloud node
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 365
- 238000002372 labelling Methods 0.000 claims abstract description 65
- 238000000034 method Methods 0.000 claims abstract description 19
- 230000005540 biological transmission Effects 0.000 claims description 30
- 238000006243 chemical reaction Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 17
- 238000010586 diagram Methods 0.000 description 9
- 238000012549 training Methods 0.000 description 9
- 238000012544 monitoring process Methods 0.000 description 4
- 238000005457 optimization Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 238000010295 mobile communication Methods 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000013480 data collection Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 230000007704 transition Effects 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
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/94—Hardware or software architectures specially adapted for image or video understanding
- G06V10/95—Hardware or software architectures specially adapted for image or video understanding structured as a network, e.g. client-server architectures
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Software Systems (AREA)
- Alarm Systems (AREA)
Abstract
The application relates to a security detection method, a security detection device, computer equipment and a storage medium. The method comprises the following steps: acquiring an image to be detected through an edge cloud node corresponding to the video acquisition equipment; the edge cloud node acquires a target safety detection model, performs safety detection on the image to be detected by using the target safety detection model to obtain a safety detection result, and sends the safety detection result to a terminal so that the terminal performs safety prompt according to the safety detection result; the edge cloud node transmits the image to be detected to a center cloud node, so that the center cloud node obtains a target safety labeling result corresponding to the image to be detected, model parameters of the target safety detection model are updated according to the image to be detected and the target safety labeling result, and the target safety detection model with the updated model parameters is sent to the edge cloud node. By adopting the method, the efficiency of safety detection can be improved.
Description
Technical Field
The present application relates to the field of computer technologies, and in particular, to a security detection method and apparatus, a computer device, and a storage medium.
Background
With the development of computer technology, because the situation that workers do not wear corresponding protection equipment according to requirements often occurs, in order to ensure the safety of the workers, safety detection is usually performed on a work site to ensure that the workers who perform dangerous work wear the corresponding protection equipment, for example, to ensure that the workers on the work site wear safety helmets.
Currently, safety inspection is usually performed manually, for example, monitoring data is analyzed manually to determine whether the wearing of workers meets the requirements. However, the manual safety inspection method consumes a lot of time, resulting in inefficient safety inspection.
Disclosure of Invention
In view of the above, it is necessary to provide a security detection method, a security detection apparatus, a computer device, and a storage medium, which can improve the efficiency of security detection, in order to solve the technical problem of low efficiency of security detection.
A security detection method, the method comprising: acquiring an image to be detected through an edge cloud node corresponding to the video acquisition equipment; the edge cloud nodes acquire a target safety detection model, and the target safety detection model is utilized to perform safety detection on the image to be detected to obtain a safety detection result; the edge cloud node sends the safety detection result to a terminal so that the terminal can carry out safety prompt according to the safety detection result; the edge cloud node transmits the image to be detected to a center cloud node, so that the center cloud node obtains a target safety labeling result corresponding to the image to be detected, model parameters of the target safety detection model are updated according to the image to be detected and the target safety labeling result, and the target safety detection model with the updated model parameters is sent to the edge cloud node.
In some embodiments, the acquiring, by an edge cloud node corresponding to the video capture device, an image to be detected includes: receiving a video to be detected sent by a network conversion device through a first network through an edge cloud node corresponding to a video acquisition device, wherein the video to be detected is sent to the network conversion device by the video acquisition device through a second network; a first data transmission rate corresponding to the first network is greater than a second data transmission rate corresponding to the second network; and selecting the image to be detected from the video to be detected.
In some embodiments, the obtaining the target security detection model comprises: acquiring a stored version number corresponding to the stored security detection model; acquiring a latest version number corresponding to a security detection model from the central cloud node; and when the stored version number is inconsistent with the latest version number, acquiring a security detection model corresponding to the latest version number from the central cloud node as the target security detection model.
In some embodiments, the to-be-detected image is a person-related image, the terminal is an audio output device, and the sending the security detection result to the terminal so that the terminal performs security prompt according to the security detection result includes: acquiring the image to be detected which does not pass the safety detection result and taking the image as a target image to be detected; acquiring the person identity information corresponding to the target image to be detected; and sending the safety detection result and the person identity information corresponding to the target image to be detected to audio output equipment, so that the audio output equipment gives an alarm according to the safety detection result and the person identity information.
In some embodiments, the obtaining a target security labeling result corresponding to the image to be detected, and updating the model parameters of the target security detection model according to the image to be detected and the target security labeling result includes: inputting the image to be detected into the target safety detection model to obtain a predicted safety labeling result corresponding to the image to be detected; calculating the difference between the predicted safety labeling result and the target safety labeling result to obtain a model loss value; and adjusting parameters of the target safety detection model according to the model loss value.
A security detection apparatus, the apparatus comprising: the image acquisition module to be detected is used for acquiring an image to be detected through an edge cloud node corresponding to the video acquisition equipment; a safety detection result obtaining module, configured to obtain a target safety detection model, and perform safety detection on the image to be detected by using the target safety detection model to obtain a safety detection result; the safety prompting module is used for sending the safety detection result to a terminal so that the terminal can perform safety prompting according to the safety detection result; and the image transmission module to be detected is used for transmitting the image to be detected to a central cloud node so that the central cloud node obtains a target safety labeling result corresponding to the image to be detected, updating model parameters of the target safety detection model according to the image to be detected and the target safety labeling result, and sending the target safety detection model with the updated model parameters to the edge cloud node.
A computer device comprising a memory storing a computer program and a processor implementing the steps of the above-described security detection method when the processor executes the computer program.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned security detection method.
According to the safety detection method, the device, the computer equipment and the storage medium, the image to be detected is obtained through the edge cloud node corresponding to the video acquisition equipment, the edge cloud node obtains the target safety detection model, the image to be detected is subjected to safety detection by utilizing the target safety detection model, and the safety detection result is obtained. The edge cloud node sends the safety detection result to the terminal, so that the terminal carries out safety prompt according to the safety detection result, automatic safety prompt is realized, and the efficiency of carrying out safety prompt is improved. The edge cloud node transmits the image to be detected to the center cloud node, so that the center cloud node obtains a target security labeling result corresponding to the image to be detected, model parameters of the target security detection model are updated according to the image to be detected and the target security labeling result, iterative optimization of the security detection model is achieved, accuracy of the security detection model is improved, and due to the fact that the edge cloud node is limited in storage resources generally, the security detection model is updated by the center cloud node instead of the edge cloud node, burden of the edge cloud node can be reduced. And sending the target safety detection model with the updated model parameters to the edge cloud node, so that the edge cloud node can detect the image to be detected by using the target safety detection model with the updated model parameters, and the accuracy of safety detection is improved.
A security detection method, the method comprising: acquiring an image to be detected transmitted by an edge cloud node, wherein the edge cloud node performs security detection on the image to be detected by using a target security detection model; acquiring a target safety labeling result corresponding to the image to be detected; and updating model parameters of the target safety detection model according to the image to be detected and the target safety labeling result, and sending the target safety detection model with the updated model parameters to the edge cloud node.
A security detection apparatus, the apparatus comprising: the system comprises an image to be detected obtaining module, an edge cloud node and a target safety detection module, wherein the image to be detected obtaining module is used for obtaining an image to be detected transmitted by the edge cloud node, and the edge cloud node is used for carrying out safety detection on the image to be detected by using a target safety detection model; the target safety labeling result acquisition module is used for acquiring a target safety labeling result corresponding to the image to be detected; and the parameter updating module is used for updating the model parameters of the target safety detection model according to the image to be detected and the target safety labeling result and sending the target safety detection model with the updated model parameters to the edge cloud node.
A computer device comprising a memory storing a computer program and a processor implementing the steps of the above-described security detection method when the processor executes the computer program.
A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the steps of the above-mentioned security detection method.
A security detection system, the system comprising: the edge cloud nodes are used for acquiring images to be detected from corresponding video acquisition equipment; obtaining a target safety detection model, and carrying out safety detection on the image to be detected by using the target safety detection model to obtain a safety detection result; sending the safety detection result to a terminal; the terminal is used for carrying out safety prompt according to the safety detection result; the edge cloud node is also used for transmitting the image to be detected to the central cloud node; the central cloud node is used for acquiring a safety labeling result corresponding to the image to be detected, updating the model parameters of the target safety detection model according to the image to be detected and the safety labeling result, and sending the target safety detection model with updated model parameters to the edge cloud node.
In some embodiments, the system further comprises: the network conversion equipment is used for receiving the video to be detected sent by the video acquisition equipment through a second network; the network conversion equipment is further used for sending the video to be detected to the edge cloud node through a first network; a first data transmission rate corresponding to the first network is greater than a second data transmission rate corresponding to the second network; the edge cloud node is further used for selecting the image to be detected from the video to be detected.
According to the safety detection system, the image to be detected is obtained through the edge cloud node corresponding to the video acquisition equipment, the edge cloud node obtains the target safety detection model, the image to be detected is subjected to safety detection through the target safety detection model, and a safety detection result is obtained. The edge cloud node sends the safety detection result to the terminal, so that the terminal carries out safety prompt according to the safety detection result, automatic safety prompt is realized, and the efficiency of carrying out safety prompt is improved. The edge cloud node transmits the image to be detected to the center cloud node, so that the center cloud node obtains a target security labeling result corresponding to the image to be detected, model parameters of the target security detection model are updated according to the image to be detected and the target security labeling result, iterative optimization of the security detection model is achieved, accuracy of the security detection model is improved, and due to the fact that the edge cloud node is limited in storage resources generally, the security detection model is updated by the center cloud node instead of the edge cloud node, burden of the edge cloud node can be reduced. And sending the target safety detection model with the updated model parameters to the edge cloud node, so that the edge cloud node can detect the image to be detected by using the target safety detection model with the updated model parameters, and the accuracy of safety detection is improved.
Drawings
FIG. 1 is a diagram of an environment in which a security detection method may be used in some embodiments;
FIG. 2 is a schematic flow chart of a security detection method in some embodiments;
FIG. 3 is a schematic flow chart of a security detection method in some embodiments;
FIG. 4 is a schematic flow chart of steps for obtaining a target security detection model in some embodiments;
FIG. 5 is a flow diagram illustrating steps of a security prompt in some embodiments;
FIG. 6 is a schematic flow chart of a security detection method in some embodiments;
FIG. 7 is a system block diagram of a security detection system in some embodiments;
FIG. 8A is a schematic diagram of a security detection system in some embodiments;
FIG. 8B is a schematic diagram of the edge cloud working in conjunction with the cloud platform in some embodiments;
FIG. 9 is a block diagram of a security detection device in some embodiments;
FIG. 10 is a diagram of the internal structure of a computer device in some embodiments.
Detailed Description
In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.
The security detection method provided by the application can be applied to the application environment shown in fig. 1. The application environment includes a video capture device 102, an edge cloud node 104, a center cloud node 106, and a terminal 108. The video acquisition device 102 and the edge cloud node 104 communicate with each other through a network, the edge cloud node 104 and the center cloud node 106 communicate with each other through a network, and the terminal 108 and the edge cloud node 104 communicate with each other through a network.
Specifically, the edge cloud node 104 acquires an image to be detected from the corresponding video acquisition device 102, the edge cloud node 104 acquires a target security detection model, the image to be detected is subjected to security detection by using the target security detection model to obtain a security detection result, the edge cloud node 104 sends the security detection result to the terminal 108, the terminal 108 performs security prompt according to the security detection result, the edge cloud node 104 transmits the image to be detected to the center cloud node 106, the center cloud node 106 acquires a security labeling result corresponding to the image to be detected, model parameters of the target security detection model are updated according to the image to be detected and the security labeling result, and the target security detection model with updated model parameters is sent to the edge cloud node 104.
The video capture device 102 may be a camera, the terminal 108 may be, but is not limited to, various audio output devices, video output devices, a personal computer, a notebook computer, a smart phone, a tablet computer, and a portable wearable device, and the edge cloud node 104 and the center cloud node 106 may be implemented by independent servers or a server cluster formed by a plurality of servers.
In some embodiments, as shown in fig. 2, a security detection method is provided, which is described by taking the method as an example applied to the edge cloud node 104 in fig. 1, and includes the following steps:
And S202, acquiring an image to be detected through the edge cloud node corresponding to the video acquisition equipment.
Specifically, the edge cloud node may establish a connection between the video capture device and the center cloud node, and a physical distance between the edge cloud node corresponding to the video capture device and the video capture device is smaller than a preset physical distance. The preset physical distance may be set as desired, for example, 3 km. The edge cloud node is, for example, a 5G (fifth generation mobile communication technology) edge cloud node. 5G is a cellular network that can provide high bandwidth, low latency, and massive connectivity. The 5G Edge cloud is, for example, a 5G MEC (Multi-access Edge Computing, 5G Multi-access Edge cloud), and the 5G MEC is calculated based on a 5G mobile Edge. The edge cloud node corresponding to the video acquisition device may be a private edge cloud node of the video acquisition device, that is, the edge cloud node corresponding to the video acquisition device only provides service for the video acquisition device. The video acquisition equipment can transmit data to the edge cloud node according to the network address of the edge cloud node.
The central cloud node may be a cloud computing server, such as a cloud server, which has sufficient computing resources and may perform at least one of model training, model publishing, and model updating, and the cloud computing server may provide centralized management resources (resources are, for example, computing capacity, storage capacity, and network capacity) and has the capability of providing resources on demand. The physical distance between the center cloud node and the video acquisition equipment can be any, and generally, the physical distance between the center cloud node and the video acquisition equipment is larger than the physical distance between the edge cloud node and the video acquisition equipment. The central cloud node may be located in a different region than the video capture device. The central cloud node and the edge cloud node may be in different network environments.
The video capture device may be any of a variety of video capture devices located at a location where security detection is desired. The video capture device may be a camera. The location where the security check is required may include at least one of a construction site, a factory workshop, or a traffic lane. The video capture device may support a standard video Streaming Protocol, which may be, for example, RTSP (Real Time Streaming Protocol).
The image to be detected refers to an image for security detection. The security detection may include at least one of a hard hat detection, a hard coat detection, or a face detection.
In some embodiments, the video capture device transmits the video data captured in real time to the corresponding edge cloud node, and the edge cloud node may obtain the image to be detected from the video data captured in real time. Specifically, the edge cloud node may determine a location where the video acquisition device is located according to the location information of the video acquisition device, determine a condition (denoted as an image detection condition) that an image related to security detection needs to meet according to a scene where the video acquisition device is located, and select a corresponding image area from a video frame of the video data according to the image detection condition to obtain an image to be detected. Wherein the location of the video capture device may include at least one of a construction site, a factory floor, or a traffic lane. The image detection conditions may be different for different locations. For example, when the place where the video capture device is located is a building site, whether a worker wears a safety helmet or not can be detected, the corresponding image detection condition can be related to the head of the human body, and the edge cloud node can select an image area related to the head of the human body from the video frame as an image to be detected.
S204, the edge cloud nodes acquire a target safety detection model, and safety detection is performed on the image to be detected by using the target safety detection model to obtain a safety detection result.
Specifically, the edge cloud node may store a plurality of types of security detection models, and the security detection models may include at least one of a helmet detection model, a safety clothing detection model, or a vehicle detection model, and the helmet detection model may detect whether a person wears safety clothing, the safety clothing detection model may detect whether the person wears safety clothing, and the vehicle detection model may detect whether a vehicle violates a traffic rule. The security detection result may include pass and fail. For the safety helmet detection model, when the safety detection result is pass, the person can be determined to wear the safety helmet, and when the safety detection result is fail, the person can be determined not to wear the safety helmet.
In some embodiments, the edge cloud node may determine the type (denoted as a target type) of the security detection model corresponding to the image to be detected according to the location to which the video acquisition device belongs, and use the security detection model corresponding to the target type as the target security detection model.
In some embodiments, the edge cloud node may obtain a trained security detection model from a training node of the security detection model and store the trained security detection model, and when security detection is required, the stored security detection model is used for security detection. The training nodes of the safety detection model can continuously optimize the trained safety detection model, and the optimized safety detection model is obtained. The training nodes of the safety detection model can distinguish the safety detection model before and after optimization according to the version number. The edge cloud node may use the security detection model corresponding to the latest version number as the target security detection model.
And S206, the edge cloud node sends the safety detection result to the terminal so that the terminal can carry out safety prompt according to the safety detection result.
Specifically, the terminal may be at least one of an audio output device or a video output device. The terminal may be located at the same site as the video capture device, e.g., both site a devices. The security prompt may include at least one of a voice alert or a video alert. The edge cloud node can send the security detection result to the terminal, wherein the security detection result is a failed security detection result, and the terminal can perform security prompt according to the failed security detection result.
In some embodiments, the edge cloud node may obtain image information of an image to be detected, send the image information and a corresponding security detection result to the terminal, and the terminal may perform security prompt according to the image information and the corresponding security detection result. The image information may include at least one of character information or vehicle information in the image. The personal information may include at least one of personal identification information or personal location information, and the vehicle information may include at least one of vehicle identification or vehicle location information.
In some embodiments, the edge cloud node may obtain an image to be detected whose security detection result is failed, determine position information of the image to be detected in a corresponding video frame, mark the image to be detected in the corresponding video frame, and transmit the marked video frame to a terminal for display.
For example, the video acquisition device is a device of a building site a, the image to be detected is an image related to the head of a human body, when the safety detection model is a helmet detection model, the edge cloud node can determine the information of the person in the image to be detected, and the information of the person and the safety detection result are transmitted to the audio output device of the building site a, so that the audio output device can perform voice prompt according to the information of the person and the safety detection result, and timely warning can be timely given to a worker who does not wear a helmet in the building site a, and the life safety of the worker is ensured.
And S208, the edge cloud node transmits the image to be detected to the center cloud node, so that the center cloud node obtains a target safety labeling result corresponding to the image to be detected, updates the model parameters of the target safety detection model according to the image to be detected and the target safety labeling result, and sends the target safety detection model with the updated model parameters to the edge cloud node.
Specifically, the safety labeling result may include at least one of a pass safety detection result or a fail safety detection result, and the target safety labeling result may be manually labeled and is a real safety labeling result corresponding to the image to be detected. For example, if the image to be detected is an image without a safety helmet, the corresponding target safety annotation result may be that the safety detection result is failed. The central cloud node may train and optimize the security detection model. The central cloud node can carry out image preprocessing on the obtained image to be detected, and the processed image to be detected is used for updating the model parameters of the target safety detection model. Certainly, the central cloud node can also select an image to be detected with higher definition from the acquired images to be detected, and the selected image to be detected with higher definition is used for updating the model parameters of the target safety detection model.
In some embodiments, the edge cloud node may store the images to be detected, and transmit the images to be detected to the center cloud node when the number of the stored images to be detected reaches a preset number. The preset number may be set as desired, for example, 1000. Certainly, the edge cloud node can transmit the image to be detected acquired within the preset time length to the center cloud node. The preset time period may be set as desired, for example, 24 hours.
In some embodiments, the central cloud node may input the image to be detected into the target security detection model, and update the model parameters of the target security detection model according to the output result of the target security detection model and the difference between the target security labeling results.
In some embodiments, the edge cloud node receives the target security detection model after model parameters are updated, which is sent by the central cloud node, and replaces the locally stored target security detection model with the target security detection model after model parameters are updated, so as to perform security detection by using the target security detection model after model parameters are updated.
In the safety detection method, the image to be detected is obtained through the edge cloud node corresponding to the video acquisition equipment, the edge cloud node obtains the target safety detection model, the image to be detected is subjected to safety detection through the target safety detection model, and a safety detection result is obtained. The edge cloud node sends the safety detection result to the terminal, so that the terminal carries out safety prompt according to the safety detection result, automatic safety prompt is realized, and the efficiency of carrying out safety prompt is improved. The edge cloud node transmits the image to be detected to the center cloud node, so that the center cloud node obtains a target security labeling result corresponding to the image to be detected, model parameters of the target security detection model are updated according to the image to be detected and the target security labeling result, iterative optimization of the security detection model is achieved, accuracy of the security detection model is improved, and due to the fact that the edge cloud node is limited in storage resources generally, the security detection model is updated by the center cloud node instead of the edge cloud node, burden of the edge cloud node can be reduced. The target safety detection model with the updated model parameters is sent to the edge cloud nodes, so that the edge cloud nodes can detect the image to be detected by using the target safety detection model with the updated model parameters, and the accuracy of safety detection is improved.
In some embodiments, as shown in fig. 3, the step S202 of acquiring the image to be detected by the edge cloud node corresponding to the video capture device includes:
s302, receiving a video to be detected sent by a network conversion device through a first network through an edge cloud node corresponding to a video acquisition device, wherein the video to be detected is sent to the network conversion device by the video acquisition device through a second network; the first data transmission rate corresponding to the first network is greater than the second data transmission rate corresponding to the second network.
S304, selecting the image to be detected from the video to be detected.
In particular, the network switching device may establish communication connections between different types of networks. The type of network may include at least one of a mobile communication network (e.g., 3G, 4G, and 5G) or wifi (wireless network). For example, the network transition device may establish a communication connection between the 5G network and the wireless network. The network switching device may be, for example, a 5G CPE (Customer Premise Equipment), and the 5G CPE may implement wifi network and 5G network switching devices.
The first network refers to a network adopted by the edge cloud node to communicate with the network conversion equipment. The first network is for example 5G. The first data transmission rate refers to a data transmission rate of the first network. The second network refers to a network through which the video capture device communicates with the network conversion device. The second network is for example a wireless network. The second data transmission rate refers to a data transmission rate of the second network.
In some embodiments, the video capture device may send the video data captured in real time to the network conversion device through the second network, the network conversion device may send the received video data to the edge cloud node through the first network, and the edge cloud node may select an image to be detected from the video to be detected.
In the above embodiment, because the first data transmission rate corresponding to the first network is greater than the second data transmission rate corresponding to the second network, the data acquired by the video acquisition device is sent to the edge cloud node through the first network by using the network conversion device, which is equivalent to increasing the rate of data transmission between the video acquisition device and the edge cloud node.
In some embodiments, as shown in fig. 4, obtaining the target security detection model includes:
s402, acquiring a storage version number corresponding to the stored security detection model.
S404, obtaining the latest version number corresponding to the security detection model from the central cloud node.
S406, when the stored version number is inconsistent with the latest version number, the security detection model corresponding to the latest version number is obtained from the central cloud node and is used as the target security detection model.
Specifically, the stored version number refers to a version number corresponding to a security detection model stored in the edge cloud node, and the latest version number refers to a version number corresponding to the latest security detection model stored in the center cloud node. The type of the security detection model corresponding to the latest version number is the same as that of the security detection model corresponding to the stored version number, for example, both the security detection models are security cap detection models. The more recent version corresponds to a higher accuracy security detection model.
In some embodiments, the edge cloud node may compare the stored version number with the latest version number, and when the stored version number is the same as the latest version number, may determine that the stored security detection model is the latest security detection model, and may therefore use the stored security detection model as the target security detection model; when the stored version number is different from the latest version number, the stored security detection model is determined not to be the latest security detection model, so that the security detection model corresponding to the latest version number can be obtained from the central cloud node, and the security detection model corresponding to the latest version number is used as the target security detection model.
In the above embodiment, when the stored version number is not consistent with the latest version number, the security detection model corresponding to the latest version number is obtained from the central cloud node and is used as the target security detection model, so that the target security detection model obtained by the edge cloud node can be ensured to be latest, and since the accuracy of the security detection model corresponding to the latest version number is higher, the accuracy of security detection can be improved by using the security detection model corresponding to the latest version number as the target security detection model.
In some embodiments, as shown in fig. 5, the image to be detected is a human-related image, the terminal is an audio output device, and the step S206 is to send the security detection result to the terminal, so that the terminal performs security prompt according to the security detection result, including:
and S502, acquiring the image to be detected with the security detection result of failure as a target image to be detected.
S504, obtaining the person identity information corresponding to the target image to be detected.
S506, the security detection result and the person identity information corresponding to the target image to be detected are sent to the audio output device, so that the audio output device gives an alarm according to the security detection result and the person identity information.
Specifically, the target image to be detected refers to an image to be detected, which has failed in the safety detection result. The personal identification information may include at least one of a job number, a name, or a title. When the image to be detected is a person-related image, the edge cloud node can acquire the identity information of the person in the image to be detected, and sends the security detection result and the person identity information corresponding to the target image to be detected to the audio output device, and the audio output device can give an alarm according to the security detection result and the person identity information.
In the above embodiment, the person identity information corresponding to the image to be detected of the target and the safety detection result is sent to the audio output device, so that the audio output device can give an alarm according to the safety detection result and the person identity information, and can give an alarm to related personnel in time, for example, timely give an alarm to personnel who do not wear a safety helmet, and the alarm efficiency is improved.
In some embodiments, obtaining a target security labeling result corresponding to an image to be detected, and updating a model parameter of the target security detection model according to the image to be detected and the target security labeling result includes: inputting an image to be detected into a target safety detection model to obtain a prediction safety labeling result corresponding to the image to be detected; calculating the difference between the predicted safety labeling result and the target safety labeling result to obtain a model loss value; and adjusting parameters of the target safety detection model according to the model loss value.
Specifically, the predicted security labeling result is a security labeling result output by the target security detection model, and the central cloud node may adjust parameters of the target security detection model in a direction of decreasing the model loss value by a gradient descent method according to the model loss value to obtain adjusted parameters.
In the above embodiment, the parameters of the target safety detection model are adjusted according to the model loss value, so that the parameters of the target safety detection model are adjusted toward the direction of decreasing the model loss value, and the accuracy of the safety detection model corresponding to the adjusted parameters is improved.
In some embodiments, as shown in fig. 6, a security detection method is provided, which is described by taking the method as an example applied to the central cloud node 106 in fig. 1, and includes the following steps:
s602, acquiring an image to be detected transmitted by the edge cloud node, and performing security detection on the image to be detected by the edge cloud node by using the target security detection model.
And S604, acquiring a target safety labeling result corresponding to the image to be detected.
And S606, updating model parameters of the target safety detection model according to the image to be detected and the target safety labeling result, and sending the target safety detection model with the updated model parameters to the edge cloud nodes.
Specifically, please refer to the related description information of the security detection method corresponding to fig. 2 for detailed description, which is not repeated herein.
In some embodiments, as shown in fig. 7, there is provided a security detection system comprising:
the edge cloud nodes are used for acquiring images to be detected from corresponding video acquisition equipment; acquiring a target safety detection model, and performing safety detection on an image to be detected by using the target safety detection model to obtain a safety detection result; and sending the safety detection result to the terminal.
And the terminal is used for carrying out safety prompt according to the safety detection result.
And the edge cloud node is also used for transmitting the image to be detected to the central cloud node.
And the central cloud node is used for acquiring a safety labeling result corresponding to the image to be detected, updating the model parameters of the target safety detection model according to the image to be detected and the safety labeling result, and sending the target safety detection model with the updated model parameters to the edge cloud node.
In some embodiments, the system further comprises:
the network conversion equipment is used for receiving the video to be detected sent by the video acquisition equipment through a second network;
the network conversion equipment is also used for sending the video to be detected to the edge cloud node through the first network; a first data transmission rate corresponding to the first network is greater than a second data transmission rate corresponding to the second network;
And the edge cloud node is also used for selecting an image to be detected from the video to be detected.
In some embodiments, as shown in fig. 8A, a security detection system is provided that includes a front-end device layer, a network transport layer, an edge layer, and a cloud platform layer. The front-end equipment layer comprises monitoring camera equipment, and the network transmission layer comprises a 5GCPE device. The edge layer comprises a 5G MEC, an AI (Artificial Intelligence) helmet wearing detection client (hereinafter, referred to as a client) is deployed on the 5G MEC, and the 5G MEC edge cloud is a cloud server deployed in a local environment of a client and has limited computing power and storage resources. And the cloud platform layer is used for model training and updating.
Specifically, the network transmission layer realizes monitoring video stream data transmission, the camera equipment of the front-end equipment layer transmits the video stream data to the 5G CPE device through the WiFi network, and the 5G CPE device transmits the video stream data to the 5G MEC edge cloud of the edge layer through the 5G network. The cloud platform layer can obtain a safety helmet detection model based on Mxnet framework and neural network Darknet53 training, and the file format of the obtained safety helmet detection model is a model file format supported by Mxnet and mainly takes params and json as extension names. The cloud platform layer can send the trained safety helmet detection model to the client of the edge layer, and synchronizes the updated safety helmet detection model to the edge cloud. The client on the 5G MEC edge cloud utilizes the safety helmet detection model to detect and analyze the video stream in real time, for example, people who do not wear the safety helmet in the video stream are detected, the detection result is displayed in real time through the corresponding client interface, and local voice alarm can be realized.
In some embodiments, the client may include a data collection module, a model scheduling module, a model storage module (including a model version number), a model service module, and a front-end display module. As shown in fig. 8B, a flowchart of cooperative work (edge cloud cooperation) between the edge layer (or the client) and the cloud platform layer is shown, the cloud platform layer is trained through a training data set to obtain a model file, the model file can be sent to the client, and a model scheduling module of the client can automatically check a version number of a model stored in the client, so that video detection is performed by using a model of the latest version, and a detection result is displayed at the front end through a model service module. The client can acquire the video data through the data acquisition module and store the video data into a picture format, and the stored picture can be transmitted to the cloud platform layer through the model scheduler so as to be used for iterative training of the model of the cloud platform.
The traditional helmet wearing detection method has the following defects: the video data capacity is large, so that the transmission delay is large, the real-time model detection cannot be realized, and the detection practicability is influenced; the model cannot adapt to the requirements of rapidly changing external application scenes in both expansibility and accuracy, so that the actual application effect is influenced. In the embodiment, the 5G network is used for transmitting the monitoring video stream data to the edge cloud layer, the situation that a person wears the safety helmet can be detected in real time on the edge layer based on the trained neural network model (safety helmet detection model), the detection real-time performance is improved, the edge layer transmits the video data to the cloud platform layer, the safety helmet detection model is subjected to iterative training on the cloud platform layer, the model is updated, and the accuracy of model detection is improved.
It should be understood that, although the steps in the flowcharts of the above embodiments are shown in sequence as indicated by the arrows, the steps are not necessarily executed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts of the above embodiments may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least a part of the steps or stages in other steps.
In some embodiments, as shown in fig. 9, there is provided a security detection apparatus comprising: an image to be detected acquisition module 902, a security detection result obtaining module 904, a security prompt module 906, and an image to be detected transmission module 908, wherein:
and an image to be detected acquiring module 902, configured to acquire an image to be detected through an edge cloud node corresponding to the video acquisition device.
A safety detection result obtaining module 904, configured to obtain a target safety detection model, and perform safety detection on an image to be detected by using the target safety detection model to obtain a safety detection result.
And the safety prompting module 906 is configured to send the safety detection result to the terminal, so that the terminal performs safety prompting according to the safety detection result.
The to-be-detected image transmission module 908 is configured to transmit an image to be detected to the central cloud node, so that the central cloud node obtains a target security labeling result corresponding to the image to be detected, updates the model parameters of the target security detection model according to the image to be detected and the target security labeling result, and sends the target security detection model with the updated model parameters to the edge cloud node.
In some embodiments, the image acquisition module 902 to be detected comprises:
and the video receiving unit to be detected is used for receiving the video to be detected sent by the network conversion equipment through the first network through the edge cloud node corresponding to the video acquisition equipment, and the video to be detected is sent to the network conversion equipment by the video acquisition equipment through the second network. The first data transmission rate corresponding to the first network is greater than the second data transmission rate corresponding to the second network.
And the to-be-detected image selecting unit is used for selecting the to-be-detected image from the to-be-detected video.
In some embodiments, the security detection result obtaining module 904 includes:
and the storage version number acquiring unit is used for acquiring the storage version number corresponding to the stored security detection model.
And the latest version number acquiring unit is used for acquiring the latest version number corresponding to the security detection model from the central cloud node.
And the safety detection model acquisition unit is used for acquiring a safety detection model corresponding to the latest version number from the central cloud node as a target safety detection model when the stored version number is inconsistent with the latest version number.
In some embodiments, the image to be detected is a human-related image, the terminal is an audio output device, and the security prompt module 906 includes:
and the target image to be detected obtaining unit is used for obtaining the image to be detected with the failed safety detection result as the target image to be detected.
And the person identity information acquisition unit is used for acquiring the person identity information corresponding to the target image to be detected.
And the person identity information sending unit is used for sending the safety detection result and the person identity information corresponding to the target image to be detected to the audio output equipment so that the audio output equipment gives an alarm according to the safety detection result and the person identity information.
In some embodiments, the image transmission module to be detected 908 comprises:
and the predicted safety labeling result obtaining unit is used for inputting the image to be detected into the target safety detection model to obtain a predicted safety labeling result corresponding to the image to be detected.
And the model loss value obtaining unit is used for calculating the difference between the predicted safety labeling result and the target safety labeling result to obtain a model loss value.
And the parameter adjusting unit is used for adjusting the parameters of the target safety detection model according to the model loss value.
A security detection apparatus, the apparatus comprising: the system comprises an image to be detected obtaining module, an edge cloud node and a target safety detection module, wherein the image to be detected obtaining module is used for obtaining an image to be detected transmitted by the edge cloud node, and the edge cloud node carries out safety detection on the image to be detected by using a target safety detection model; the target safety labeling result acquisition module is used for acquiring a target safety labeling result corresponding to the image to be detected; and the parameter updating module is used for updating the model parameters of the target safety detection model according to the image to be detected and the target safety labeling result, and sending the target safety detection model with the updated model parameters to the edge cloud nodes.
For the specific definition of the safety detection device, reference may be made to the above definition of the safety detection method, which is not described herein again. The modules in the security detection device can be implemented in whole or in part 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.
In some embodiments, a computer device is provided, which may be an edge cloud node and a central cloud node, and the internal structure diagram of which may be as shown in fig. 10. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a security detection method.
Those skilled in the art will appreciate that the architecture shown in fig. 10 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
In some embodiments, a computer device is provided, comprising a memory and a processor, the memory having stored therein a computer program, the processor implementing the steps of the above-described security detection method when executing the computer program.
In some embodiments, a computer-readable storage medium is provided, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the above-mentioned security detection method.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database or other medium used in the embodiments provided herein can include at least one of non-volatile and volatile memory. Non-volatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical storage, or the like. Volatile Memory can include Random Access Memory (RAM) or external cache Memory. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), among others.
The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as 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 (11)
1. A security detection method, the method comprising:
acquiring an image to be detected through an edge cloud node corresponding to the video acquisition equipment;
the edge cloud nodes acquire a target safety detection model, and the target safety detection model is utilized to perform safety detection on the image to be detected to obtain a safety detection result;
the edge cloud node sends the safety detection result to a terminal so that the terminal can carry out safety prompt according to the safety detection result;
The edge cloud node transmits the image to be detected to a center cloud node, so that the center cloud node obtains a target safety labeling result corresponding to the image to be detected, model parameters of the target safety detection model are updated according to the image to be detected and the target safety labeling result, and the target safety detection model with the updated model parameters is sent to the edge cloud node.
2. The method according to claim 1, wherein the acquiring the image to be detected through the edge cloud node corresponding to the video capture device comprises:
receiving a video to be detected sent by a network conversion device through a first network through an edge cloud node corresponding to a video acquisition device, wherein the video to be detected is sent to the network conversion device by the video acquisition device through a second network; a first data transmission rate corresponding to the first network is greater than a second data transmission rate corresponding to the second network;
and selecting the image to be detected from the video to be detected.
3. The method of claim 1, wherein the obtaining a target security detection model comprises:
Acquiring a stored version number corresponding to the stored security detection model;
acquiring a latest version number corresponding to a security detection model from the central cloud node;
and when the stored version number is inconsistent with the latest version number, acquiring a security detection model corresponding to the latest version number from the central cloud node as the target security detection model.
4. The method according to claim 1, wherein the image to be detected is a person-related image, the terminal is an audio output device, and the sending the security detection result to the terminal so that the terminal performs security prompt according to the security detection result comprises:
acquiring the image to be detected which does not pass the safety detection result and taking the image as a target image to be detected;
acquiring the person identity information corresponding to the target image to be detected;
and sending the safety detection result and the person identity information corresponding to the target image to be detected to audio output equipment, so that the audio output equipment gives an alarm according to the safety detection result and the person identity information.
5. The method according to claim 1, wherein the obtaining of the target security labeling result corresponding to the image to be detected and the updating of the model parameters of the target security detection model according to the image to be detected and the target security labeling result comprise:
Inputting the image to be detected into the target safety detection model to obtain a predicted safety labeling result corresponding to the image to be detected;
calculating the difference between the predicted safety labeling result and the target safety labeling result to obtain a model loss value;
and adjusting parameters of the target safety detection model according to the model loss value.
6. A security detection method, the method comprising:
acquiring an image to be detected transmitted by an edge cloud node, wherein the edge cloud node performs security detection on the image to be detected by using a target security detection model;
acquiring a target safety labeling result corresponding to the image to be detected;
and updating model parameters of the target safety detection model according to the image to be detected and the target safety labeling result, and sending the target safety detection model with the updated model parameters to the edge cloud node.
7. A security detection system, the system comprising:
the edge cloud nodes are used for acquiring images to be detected from corresponding video acquisition equipment; obtaining a target safety detection model, and carrying out safety detection on the image to be detected by using the target safety detection model to obtain a safety detection result; sending the safety detection result to a terminal;
The terminal is used for carrying out safety prompt according to the safety detection result;
the edge cloud node is also used for transmitting the image to be detected to the central cloud node;
the central cloud node is used for acquiring a safety labeling result corresponding to the image to be detected, updating the model parameters of the target safety detection model according to the image to be detected and the safety labeling result, and sending the target safety detection model with updated model parameters to the edge cloud node.
8. The system of claim 7, further comprising:
the network conversion equipment is used for receiving the video to be detected sent by the video acquisition equipment through a second network;
the network conversion equipment is further used for sending the video to be detected to the edge cloud node through a first network; a first data transmission rate corresponding to the first network is greater than a second data transmission rate corresponding to the second network;
the edge cloud node is further used for selecting the image to be detected from the video to be detected.
9. A security detection apparatus, the apparatus comprising:
the image acquisition module to be detected is used for acquiring an image to be detected through an edge cloud node corresponding to the video acquisition equipment;
A safety detection result obtaining module, configured to obtain a target safety detection model, and perform safety detection on the image to be detected by using the target safety detection model to obtain a safety detection result;
the safety prompting module is used for sending the safety detection result to a terminal so that the terminal can perform safety prompting according to the safety detection result;
and the image transmission module to be detected is used for transmitting the image to be detected to a central cloud node so that the central cloud node obtains a target safety labeling result corresponding to the image to be detected, updating model parameters of the target safety detection model according to the image to be detected and the target safety labeling result, and sending the target safety detection model with the updated model parameters to the edge cloud node.
10. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor, when executing the computer program, implements the steps of the method of any of claims 1 to 5 or 6.
11. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 5 or 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010662974.1A CN111860256A (en) | 2020-07-10 | 2020-07-10 | Security detection method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010662974.1A CN111860256A (en) | 2020-07-10 | 2020-07-10 | Security detection method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111860256A true CN111860256A (en) | 2020-10-30 |
Family
ID=73153278
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010662974.1A Pending CN111860256A (en) | 2020-07-10 | 2020-07-10 | Security detection method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111860256A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112527849A (en) * | 2020-12-23 | 2021-03-19 | 珠海星客合创科技有限公司 | Intelligent data annotation method based on cloud-edge mixed Internet of things system |
| CN112650580A (en) * | 2020-11-02 | 2021-04-13 | 北京迅达云成科技有限公司 | Industrial big data monitoring system based on edge calculation |
| CN112861809A (en) * | 2021-03-22 | 2021-05-28 | 南京大学 | Classroom new line detection system based on multi-target video analysis and working method thereof |
| CN114065193A (en) * | 2021-11-23 | 2022-02-18 | 北京邮电大学 | Deep learning safety method applied to image task in edge cloud environment |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110490077A (en) * | 2019-07-18 | 2019-11-22 | 北京工业大数据创新中心有限公司 | A kind of intelligence is broken rules and regulations recognition methods and system |
| CN110505446A (en) * | 2019-07-29 | 2019-11-26 | 西安电子科技大学 | The hotel's video security protection system calculated based on mist |
| CN110580182A (en) * | 2019-08-19 | 2019-12-17 | 北京邮电大学 | Inter-cloud computing unloading method and device in edge computing |
| CN110619314A (en) * | 2019-09-24 | 2019-12-27 | 杭州宇泛智能科技有限公司 | Safety helmet detection method and device and electronic equipment |
| CN110688893A (en) * | 2019-08-22 | 2020-01-14 | 成都通甲优博科技有限责任公司 | Detection method for wearing safety helmet, model training method and related device |
| WO2020019673A1 (en) * | 2018-07-25 | 2020-01-30 | 深圳云天励飞技术有限公司 | Construction site monitoring method and device based on image analysis, and readable storage medium |
| CN111191581A (en) * | 2019-12-27 | 2020-05-22 | 深圳供电局有限公司 | Safety helmet detection method and device based on electric power construction and computer equipment |
| AU2020100705A4 (en) * | 2020-05-05 | 2020-06-18 | Chang, Jiaying Miss | A helmet detection method with lightweight backbone based on yolov3 network |
-
2020
- 2020-07-10 CN CN202010662974.1A patent/CN111860256A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2020019673A1 (en) * | 2018-07-25 | 2020-01-30 | 深圳云天励飞技术有限公司 | Construction site monitoring method and device based on image analysis, and readable storage medium |
| CN110490077A (en) * | 2019-07-18 | 2019-11-22 | 北京工业大数据创新中心有限公司 | A kind of intelligence is broken rules and regulations recognition methods and system |
| CN110505446A (en) * | 2019-07-29 | 2019-11-26 | 西安电子科技大学 | The hotel's video security protection system calculated based on mist |
| CN110580182A (en) * | 2019-08-19 | 2019-12-17 | 北京邮电大学 | Inter-cloud computing unloading method and device in edge computing |
| CN110688893A (en) * | 2019-08-22 | 2020-01-14 | 成都通甲优博科技有限责任公司 | Detection method for wearing safety helmet, model training method and related device |
| CN110619314A (en) * | 2019-09-24 | 2019-12-27 | 杭州宇泛智能科技有限公司 | Safety helmet detection method and device and electronic equipment |
| CN111191581A (en) * | 2019-12-27 | 2020-05-22 | 深圳供电局有限公司 | Safety helmet detection method and device based on electric power construction and computer equipment |
| AU2020100705A4 (en) * | 2020-05-05 | 2020-06-18 | Chang, Jiaying Miss | A helmet detection method with lightweight backbone based on yolov3 network |
Non-Patent Citations (1)
| Title |
|---|
| 施辉;陈先桥;杨英;: "改进YOLO v3的安全帽佩戴检测方法", 计算机工程与应用, no. 11 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112650580A (en) * | 2020-11-02 | 2021-04-13 | 北京迅达云成科技有限公司 | Industrial big data monitoring system based on edge calculation |
| CN112650580B (en) * | 2020-11-02 | 2021-07-20 | 北京迅达云成科技有限公司 | Industrial big data monitoring system based on edge calculation |
| CN112527849A (en) * | 2020-12-23 | 2021-03-19 | 珠海星客合创科技有限公司 | Intelligent data annotation method based on cloud-edge mixed Internet of things system |
| CN112861809A (en) * | 2021-03-22 | 2021-05-28 | 南京大学 | Classroom new line detection system based on multi-target video analysis and working method thereof |
| CN112861809B (en) * | 2021-03-22 | 2024-03-12 | 南京大学 | Classroom head-up detection system based on multi-target video analysis and working method thereof |
| CN114065193A (en) * | 2021-11-23 | 2022-02-18 | 北京邮电大学 | Deep learning safety method applied to image task in edge cloud environment |
| CN114065193B (en) * | 2021-11-23 | 2024-05-07 | 北京邮电大学 | Deep learning security method applied to image task in edge cloud environment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111860256A (en) | Security detection method and device, computer equipment and storage medium | |
| CN210605801U (en) | Construction engineering project construction informatization supervision system | |
| CN111918039B (en) | Artificial intelligence high risk operation management and control system based on 5G network | |
| JP6125035B2 (en) | Video surveillance system using mobile terminals | |
| EP3756375B1 (en) | Method and system for automated dynamic network slice deployment using artificial intelligence | |
| CN105072425A (en) | Video monitoring method and video monitoring device | |
| CN108334831A (en) | A kind of monitoring image processing method, monitoring terminal and system | |
| CN108574696A (en) | A kind of command dispatching system and working method based on GIS map | |
| WO2021027844A1 (en) | Model training method and apparatus, system, prediction method, and computer readable storage medium | |
| CN110225143A (en) | A kind of wisdom security protection data system | |
| CN110781735A (en) | Alarm method and system for identifying on-duty state of personnel | |
| CN115393781A (en) | Video monitoring data processing method and device | |
| CN115168024A (en) | Early warning method, device and equipment based on edge calculation | |
| CN109214326A (en) | A kind of information processing method, device and system | |
| CN111988397B (en) | Earthquake-proof disaster-reduction disaster-relief method and system based on edge calculation | |
| KR20050065199A (en) | Spatial imagery information delivery system of disaster management and method | |
| WO2020073555A1 (en) | Paging method and apparatus, equipment, and storage medium | |
| CN112367397A (en) | Monitoring and early warning method and system for field work, computer equipment and storage medium | |
| CN112200995A (en) | Low-power consumption field security system based on hybrid architecture | |
| KR102454920B1 (en) | Surveillance system and operation method thereof | |
| CN112016380A (en) | Wild animal monitoring method and system | |
| CN115620179A (en) | Fire detection method and device based on unmanned aerial vehicle, electronic equipment and medium | |
| CN115442739A (en) | Subway construction temporary communication processing method and device, electronic equipment and medium | |
| CN112669353B (en) | Data processing method, data processing device, computer equipment and storage medium | |
| CN113901873A (en) | Power transmission line early warning method, device and system |
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 |