LED screen broadcast control safety management system and method, and LED screen

Abstract

The invention relates to the field of screen playing, and particularly discloses an LED screen playing control safety management system and method, and an LED screen. A main control unit receives original video data, sends the original video data to a storage unit for storage, extracts each frame of image from the stored original video data, and sends the extracted image to a computing power unit for safety information identification; and sending the images after the security information identification to a storage unit for storage, and restoring all the images after the security information identification into video data for playing. According to the data real-time detection and management technology based on the edge side, edge side data monitoring and management are achieved, high bandwidth is not needed, the network transmission risk is avoided, product deployment is flexible, expandability is high, and the spreading risk of violation information is effectively reduced on the premise that the real-time performance is not affected.

Description

Technical Field

[0001] The present invention relates to the field of screen broadcasting, and in particular to a LED screen broadcast control safety management system and method, and an LED screen. Background Art

[0002] With the rapid development of information technology, LED screens have become an important channel for information dissemination. The safe broadcast control system is an important technical means to ensure the safety of broadcast content and prevent broadcast accidents. Existing safe broadcast control technologies mainly use pre-encryption technology and automatic identification and filtering technology. The former uses encryption algorithms such as AES and RSA to encrypt the transmission signal to prevent the broadcast content from being illegally intercepted or tampered with during the transmission process. The latter is mainly deployed on the server side, relying on powerful algorithm processing capabilities and high network bandwidth. There is still a risk of data interception and tampering during the layered network transmission process.

[0003] Existing technologies mainly rely on the architecture of terminal, network and cloud, all of which are indispensable, and have high requirements on network bandwidth and cloud processing capabilities. They have obvious disadvantages in edge scenarios with high requirements on big data processing and real-time performance, and have high deployment costs, making them unsuitable for LED screen broadcasting equipment. Summary of the invention

[0004] To solve the above problems, the present invention provides an LED screen broadcast control security management system and method, and an LED screen, which are based on real-time data detection and management technology on the edge side to achieve edge side data monitoring and management, without the need for high bandwidth and avoiding network transmission risks. The product has flexible deployment and strong scalability, and can effectively reduce the risk of dissemination of illegal information without affecting real-time performance.

[0005] In a first aspect, the technical solution of the present invention provides a LED screen broadcast control safety management system, comprising: Storage unit: used to store original video data and image data after security information identification; Computing unit: used to pre-train the security information recognition neural network model, and use the security information recognition neural network model to detect and process security information on images; Main control unit: used to receive original video data, send the original video data to the storage unit for storage, extract each frame of image from the stored original video data and send it to the computing unit for security information identification, send the image after security information identification to the storage unit for storage, and restore all images after security information identification to video data for playback.

[0006] In an optional implementation, the main control unit adopts the ARM architecture RK35XX platform and LPDDR5 memory, integrates the NPU and the extended computing power card as the computing power unit, and integrates the SSD solid state drive as the storage unit.

[0007] In an optional implementation, the main control unit also integrates a multi-channel data interface to implement multi-node network communication for receiving original video data.

[0008] In a second aspect, the technical solution of the present invention provides a LED screen broadcast control safety management method, which is implemented based on any of the above-mentioned systems, and the method includes the following steps: Receiving raw video data; Sending the original video data to a storage unit for storage; Extract each frame of image from the stored raw video data and send it to the computing unit for security information recognition; Sending the image after security information recognition to a storage unit for storage; All images after security information recognition are restored to video data for playback.

[0009] In an optional implementation, extracting each frame of image from the stored original video data and sending it to the computing unit for security information recognition specifically includes: Decode the original video data and extract the image frame by frame in sequence; Sequentially mark each frame of the extracted image; Each frame of the sequentially marked image is sent to the computing unit for security information recognition.

[0010] In an optional implementation, each frame of the sequentially marked image is sent to a computing unit for security information recognition, specifically including: Input the current image into a pre-trained security information recognition neural network model for feature recognition; Detect whether there is any illegal information based on the feature recognition results; If there is any illegal information, the current image will be processed securely.

[0011] In an optional implementation, the security information recognition neural network model includes a convolution layer, an activation function and a fully connected layer. The convolution layer extracts local features of the input image through a convolution kernel filter, the activation function realizes the recognition and classification of local features of the image, and outputs feedback results. The fully connected layer performs multiple fitting on the local features of the image to strengthen the feature information. When training the neural network model for security information recognition, first classify the violation information and create data annotation samples. Create the size and number of CNN convolution kernels based on the data classification and samples, confirm the activation function, select the Adam optimizer to train the CNN network, and increase training diversity by scaling, rotating, and flipping the image.

[0012] In an optional implementation, performing security processing on the current image includes blurring or deleting frames of the current image.

[0013] In an optional implementation, all images after security information recognition are restored to video data for playback, specifically including: Encode according to the sequential markings of the image, and restore the image after all security information is identified to video data for playback.

[0014] In a third aspect, the technical solution of the present invention provides an LED screen, configured with any of the systems described above, and executing any of the methods described above.

[0015] The LED screen broadcast control security management system and method, and the LED screen provided by the present invention have the following beneficial effects compared to the prior art: a system including a storage unit, a computing unit, and a main control unit is constructed on the edge side LED screen, and a security information recognition neural network model is trained, data storage is realized based on the storage unit, security information recognition using the model is realized based on the computing unit, and edge side security management is realized based on the main control unit. The present invention is based on the real-time data detection and management technology on the edge side, realizes edge side data monitoring and management, does not require high bandwidth and avoids network transmission risks, and the product deployment is flexible and scalable, and the risk of dissemination of illegal information is effectively reduced without affecting real-time performance. BRIEF DESCRIPTION OF THE DRAWINGS In order to more clearly illustrate the embodiments of the present invention or the technical solutions of the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a structural schematic diagram of an LED screen broadcast control safety management system provided by an embodiment of the present invention.

[0017] Figure 2 It is a flow chart of a LED screen broadcast control safety management method provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to enable those skilled in the art to better understand the scheme of the present invention, the present invention is further described in detail below in conjunction with the accompanying drawings and specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0020] Figure 1 This is a schematic diagram of the structure of a LED screen broadcast control safety management system provided by an embodiment of the present invention. Figure 1 The system achieves real-time detection and management of data based on the edge side.

[0021] like Figure 1 As shown, the system includes a storage unit, a computing unit and a main control unit.

[0022] Storage unit: used to store original video data and image data after security information identification.

[0023] Computing power unit: used to pre-train the security information recognition neural network model, and use the security information recognition neural network model to detect and process security information on images.

[0024] Main control unit: used to receive original video data, send the original video data to the storage unit for storage, extract each frame of image from the stored original video data and send it to the computing unit for security information identification, send the image data after security information identification to the storage unit for storage, and restore all images after security information identification to video data for playback.

[0025] In this embodiment, the main control unit adopts the ARM architecture RK35XX platform and the high-bandwidth LPDDR5 memory solution, integrates the high-computing power NPU and the expandable computing power card as the computing power unit, and integrates the high-bandwidth SSD solid-state hard disk as the storage unit. The SSD solid-state hard disk adopts the M.2 interface SSD solid-state hard disk. In the edge application scenario, the network communication module is mostly concentrated in the main control unit. The main control unit also integrates a multi-channel data interface to support multi-node network communication for receiving original video data. In addition, the main control unit also provides a terminal device system layer communication API interface protocol to support the management of the broadcast control system on the application side.

[0026] The software can be based on the Linux operating system and an operating environment that can support the deployment of CNN model architecture. In addition, the main control unit also provides a system-level API interface protocol to support data communication and system management on the application side.

[0027] An embodiment of an LED screen broadcast control safety management system is described in detail above. Based on the LED screen broadcast control safety management system described in the above embodiment, an embodiment of the present invention also provides an LED screen broadcast control safety management method corresponding to the method.

[0028] Figure 2 It is a flow chart of a LED screen broadcast control safety management method provided by an embodiment of the present invention. Figure 2 The execution subject may be the LED screen broadcast control safety management system of the above embodiment. According to different requirements, the order of the steps in the flow chart may be changed, and some may be omitted.

[0029] like Figure 2 As shown, the method includes the following steps.

[0030] S1, receiving original video data.

[0031] In this embodiment, the main control unit receives the video image processing task and the original video data to be processed. It should be noted that the main control unit refers to the main control of the screencasting device, and the data can be transmitted through other common methods such as the network. This embodiment mainly performs risk control on the data received by the screencasting terminal device, and the risk of the transmission process does not affect this solution.

[0032] S2, sending the original video data to a storage unit for storage.

[0033] In this embodiment, the main control unit sends the received original video data to the SSD solid state drive for caching. The high-bandwidth SSD hard drive development can be applied to the caching of video images and other data of edge computing devices.

[0034] S3, extracts each frame of image from the stored original video data and sends it to the computing unit for security information recognition.

[0035] In this embodiment, when the main control unit receives a data task, it stores the pre-processed and processed data through the integrated large-capacity SSD solid-state hard drive, performs CNN model training and image feature recognition through the high-computing power NPU and extended computing power card integrated in the scheduling system, and performs specific processing on the identified problem image frames.

[0036] The computing unit pre-trains the security information recognition neural network model, which performs security information recognition on images, including the detection and processing of security information. The main control unit extracts images from the stored original video data and transmits them to the computing unit, which executes the security information recognition neural network model to perform security information recognition, specifically including the following steps.

[0037] S3.1, decode the original video data and extract the image frame by frame in sequence.

[0038] This embodiment extracts the original video data into images of all frames, and performs security information recognition on the images of each frame.

[0039] S3.2, sequentially mark each extracted frame of the image.

[0040] The purpose of this step is to mark each frame of the image sequentially so that it can be restored to video form after the subsequent security information recognition processing is completed, and restored according to the sequential marking to ensure the correctness of the restored video.

[0041] S3.3, each frame of the sequentially marked image is sent to the computing unit for security information recognition.

[0042] S3.3.1, input the current image into a pre-trained security information recognition neural network model for feature recognition.

[0043] S3.3.2, detect whether there is any illegal information based on the feature recognition results.

[0044] S3.3.3, if there is any illegal information, perform security processing on the current image.

[0045] The computing unit uses the security information recognition neural network model to identify security information for each frame of the image, and performs security processing on the image when illegal information is identified. The security information recognition neural network model of this embodiment has a convolution layer, an activation function, and a fully connected layer. Among them, the convolution layer extracts local features of the input image through a convolution kernel filter, the activation function realizes the recognition and classification of local features of the image, and outputs feedback results. The fully connected layer performs multiple fittings on the local features of the image to strengthen the feature information. The activation function can use an activation function with the mathematical form of f(x)=Max(ax,x).

[0046] When training the model, configure appropriate training parameters (including the number of data samples q, the upper limit of training u, the coefficient of the proportion of different feature data samples s, and the numerical range needs to be configured according to different scenarios) to train and tune the model. First, classify the preset confidential, sensitive and other illegal information and create data annotation samples. Then, create the size and number of CNN convolution kernels based on the data classification and samples, and confirm the activation function f(x)=Max(ax,x). Select the Adam optimizer to train the CNN network, and increase the training diversity by scaling, rotating, flipping and other operations on the image. The CNN network outputs the type of illegal information identified and the processing results for each frame of the image. In an optional implementation, a model optimization and upgrade interface can be provided at the system level.

[0047] In this embodiment, performing security processing on the current image includes blurring or deleting frames of the current image, and blurring or deleting frames of a specific area (an area with illegal information) may be performed.

[0048] S4, sending the image after security information recognition to a storage unit for storage.

[0049] In this embodiment, for images with illegal information, they are sent to the storage unit for storage after being blurred or frame deleted, and for images without illegal information, they are directly sent to the storage unit for storage.

[0050] S5, restore all images after security information recognition to video data for playback.

[0051] In this embodiment, encoding is performed according to the sequence mark of the image, and all images after security information recognition are restored to video data for playback. The main control unit encodes all images after security information recognition in the original order. The image order has been marked when the frame is extracted. The encoding here mainly refers to the restoration of the original video after frame extraction, and then outputs it to the display end for playback.

[0052] This embodiment is based on real-time data detection and management technology on the edge side to achieve edge-side data monitoring and management. It does not require high bandwidth and avoids network transmission risks. The product has flexible deployment and strong scalability, and can effectively reduce the risk of dissemination of illegal information without affecting real-time performance.

[0053] This embodiment provides an LED screen, configured with the LED screen broadcast control safety management system of the above embodiment, and executes the LED screen broadcast control safety management method of the above embodiment.

[0054] The above disclosure is only a preferred embodiment of the present invention, but the present invention is not limited thereto. Any non-creative changes that can be thought of by a person skilled in the art, as well as several improvements and modifications made without departing from the principle of the present invention, should fall within the protection scope of the present invention.

Claims

1. A LED screen broadcast control safety management system, characterized in that: include, Storage unit: used to store original video data and image data after security information identification; Computing unit: used to pre-train the security information recognition neural network model, and use the security information recognition neural network model to detect and process security information on images; Main control unit: used to receive original video data, send the original video data to the storage unit for storage, extract each frame of image from the stored original video data and send it to the computing unit for security information identification, send the image after security information identification to the storage unit for storage, and restore all images after security information identification to video data for playback.

2. The LED screen broadcast control safety management system according to claim 1 is characterized in that: The main control unit adopts the ARM architecture RK35XX platform and LPDDR5 memory, integrates NPU and extended computing power card as computing power unit, and integrates SSD solid state drive as storage unit.

3. The LED screen broadcast control safety management system according to claim 2 is characterized in that: The main control unit also integrates multiple data interfaces to achieve multi-node network communication for receiving original video data.

4. A LED screen broadcast control safety management method, characterized in that: Based on the system implementation according to any one of claims 1 to 3, the method comprises the following steps: Receiving raw video data; Sending the original video data to a storage unit for storage; Extract each frame of image from the stored raw video data and send it to the computing unit for security information recognition; Sending the image after security information recognition to a storage unit for storage; All images after security information recognition are restored to video data for playback.

5. The LED screen broadcast control safety management method according to claim 4 is characterized in that: Extract each frame of the image from the stored raw video data and send it to the computing unit for security information recognition, including: Decode the original video data and extract the image frame by frame in sequence; Sequentially mark each frame of the extracted image; Each frame of the sequentially marked image is sent to the computing unit for security information recognition.

6. The LED screen broadcast control safety management method according to claim 5 is characterized in that: Each frame of the sequentially marked image is sent to the computing unit for security information recognition, including: Input the current image into a pre-trained security information recognition neural network model for feature recognition; Detect whether there is any illegal information based on the feature recognition results; If there is any illegal information, the current image will be processed securely.

7. The LED screen broadcast control safety management method according to claim 6 is characterized in that: The neural network model for security information recognition includes convolutional layers, activation functions, and fully connected layers. The convolutional layer extracts local features of the input image through convolution kernel filters. The activation function recognizes and classifies local features of the image and outputs feedback results. The fully connected layer performs multiple fittings on local features of the image to enhance feature information. When training the neural network model for security information recognition, first classify the violation information and create data annotation samples. Create the size and number of CNN convolution kernels based on the data classification and samples, confirm the activation function, select the Adam optimizer to train the CNN network, and increase training diversity by scaling, rotating, and flipping the image.

8. The LED screen broadcast control safety management method according to claim 6 or 7, characterized in that: The security processing of the current image includes blurring or deleting frames of the current image.

9. The LED screen broadcast control safety management method according to claim 5 is characterized in that: All images after security information recognition are restored to video data for playback, including: Encode according to the sequential markings of the image, and restore the image after all security information is identified to video data for playback.

10. An LED screen, characterized in that: The system described in any one of claims 1 to 3 is configured to perform the method described in any one of claims 4 to 9.

Images