Management Method for Public Landscape Camera Devices in Smart City, Internet of Things System, Management Device and Medium

Abstract

The embodiments of this specification provide a management method for a smart city public landscape camera device and an Internet of Things system. This method is implemented based on the Internet of Things system for managing smart city public landscape camera devices, and this Internet of Things system for managing smart city public landscape camera devices includes a user platform, a service platform, a management platform, and an object platform; this method includes: statistically counting the number of viewers of the landscape images corresponding to different user platforms within a preset future time length based on the service platform; and sending the number of viewers to the management platform; wherein, the number of viewers within the preset future time length is determined based on the viewing numbers of the landscape images in a preset historical time period by a third prediction model, and the third prediction model is a machine learning model; determining the camera devices to be cancelled based on the management platform, and the camera devices to be cancelled are the camera devices corresponding to the landscape images whose viewing numbers do not meet the preset conditions.

Description

[0001] Division Explanation

[0002] This application is a divisional application filed in response to a Chinese application with an application date of July 11, 2022, an application number of 202210807951.4, and an invention title of "A Method for Controlling Live Broadcast of Public Landscapes in a Smart City and an Internet of Things System". Technical Field

[0003] This specification relates to the field of the Internet of Things and cloud platforms, and particularly to a method for managing public landscape camera devices in a smart city and an Internet of Things system. Background Art

[0004] Live broadcast of public landscapes is an important part of the construction of a smart city, which enables citizens to visit various scenic views of the scenic spots without leaving their homes. Due to the large number of scenic spots and limited camera acquisition devices, it is impossible to meet the viewing needs of all citizens.

[0005] Therefore, it is necessary to provide a method for managing public landscape camera devices in a smart city and an Internet of Things system, which uses the Internet of Things and cloud platforms to improve the efficiency of public landscape live broadcast control while meeting the viewing needs of most citizens. Summary of the Invention

[0006] One embodiment of this specification provides a method for managing public landscape camera devices in a smart city, which is implemented based on an Internet of Things system for managing public landscape camera devices in a smart city. The Internet of Things system for managing public landscape camera devices in a smart city includes a user platform, a service platform, a management platform, and an object platform. Among them, the number of user platforms is multiple, and different user platforms correspond to playing landscape pictures collected by different camera devices. The number of object platforms is multiple, and different object platforms are correspondingly arranged at different camera positions of the camera devices. The method includes: based on the service platform, counting the number of viewers of the landscape pictures corresponding to different user platforms within a preset future time period; and sending the number of viewers to the management platform. Among them, the number of viewers within the preset future time period is determined based on a third prediction model for the number of viewers of the landscape pictures in a preset historical time period, and the third prediction model is a machine learning model; based on the management platform, determining the camera devices to be cancelled, where the camera devices to be cancelled are the camera devices corresponding to the landscape pictures whose number of viewers does not meet the preset conditions.

[0007] One embodiment of this specification provides an Internet of Things system for managing public landscape camera devices in a smart city. The system includes a user platform, a service platform, a management platform, and an object platform. Among them, the number of user platforms is multiple, and different user platforms play landscape images collected by different camera devices. The number of object platforms is multiple, and different object platforms are correspondingly arranged with the camera devices at different camera positions. The object platform is used to: collect the landscape images based on the camera devices. The user platform is used to: play the landscape images, and the landscape images are collected by the corresponding object platform. The service platform is used to: count the number of viewers of the landscape images corresponding to different user platforms within a preset future time period; and send the number of viewers to the management platform. Among them, the number of viewers within the preset future time period is determined based on a third prediction model for the number of viewers of the landscape images in a preset historical time period, and the third prediction model is a machine learning model. The management platform is used to: determine the camera devices to be cancelled, and the camera devices to be cancelled are the camera devices corresponding to the landscape images whose number of viewers does not meet the preset conditions.

[0008] One embodiment of this specification provides a device for managing public landscape camera devices in a smart city. The device includes at least one processor and at least one memory. The at least one memory is used to store computer instructions. The at least one processor is used to execute at least part of the computer instructions to implement the method for managing public landscape camera devices in a smart city as described in any of the above embodiments.

[0009] One embodiment of this specification provides a computer-readable storage medium. The storage medium stores computer instructions, and when a computer reads the computer instructions in the storage medium, the computer executes the method for managing public landscape camera devices in a smart city. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] This specification will be further described by way of exemplary embodiments, and these exemplary embodiments will be described in detail through the drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where:

[0011] Figure 1 is a schematic diagram of the application scenario of the Internet of Things system for live broadcast control of public landscape in a smart city shown in some embodiments of this specification;

[0012] Figure 2 is an exemplary diagram of the Internet of Things system for live broadcast control of public landscape in a smart city shown in some embodiments of this specification;

[0013] Figure 3is an exemplary flowchart of a live broadcast control method for a smart city public landscape shown in some embodiments of this specification;

[0014] Figure 4 is an exemplary flowchart of determining the overall user intention shown in some embodiments of this specification;

[0015] Figure 5 is an exemplary flowchart of determining the intention weight value shown in some embodiments of this specification;

[0016] Figure 6 is an exemplary structural diagram of a first prediction model shown in some embodiments of this specification;

[0017] Figure 7 is an exemplary flowchart of an adjustment method for a camera device shown in some embodiments of this specification. Detailed implementation manners

[0018] To more clearly illustrate the technical solutions of the embodiments of this specification, the accompanying drawings required for the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without creative efforts, this specification can also be applied to other similar scenarios based on these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the figures represent the same structure or operation.

[0019] It should be understood that the "system", "device", "unit" and / or "module" used herein is a way to distinguish different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the said words can be replaced by other expressions.

[0020] As shown in this specification and the claims, unless the context clearly indicates an exception, words such as "a", "an", "one" and / or "the" are not specifically singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of the clearly identified steps and elements, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0021] Flowcharts are used in this specification to illustrate the operations performed by the systems according to the embodiments of this specification. It should be understood that the operations before or after may not be precisely executed in sequence. On the contrary, the steps can be processed in reverse order or simultaneously. At the same time, other operations can also be added to these processes, or one or several steps can be removed from these processes.

[0022] Figure 11 is a schematic diagram of an application scenario for a smart city public landscape live broadcast control IoT system according to some embodiments of this specification. In some embodiments, application scenario 100 may include a server 110, a storage device 120, a user terminal 130, a sensor device 140, an IoT gateway 150, a network 160, and a landscape image 170.

[0023] In some embodiments, server 110 may be a single server or a server group. The server group may be centralized or distributed. For example, server 110 may be a distributed system. In some embodiments, server 110 may be local or remote. In some embodiments, server 110 may be implemented on a cloud platform. In some embodiments, server 110 or a portion of server 110 may be integrated into sensor device 140.

[0024] In some embodiments, server 110 may include a processing device 112. Processing device 112 may be used to obtain information and analyze and process the collected information to perform one or more functions described herein. For example, processing device 112 may obtain voting information, viewing time, number of viewers, and other information from user terminal 130, and perform weighted calculations to determine the user's overall willingness. For another example, processing device 112 may generate a camera parameter control strategy based on the user's overall willingness, issue control instructions to sensor device 140, and control sensor device 140 to capture new landscape images.

[0025] In some embodiments, processing device 112 may include one or more processing engines (eg, single-chip processing engines or multi-chip processing engines). By way of example only, processing device 112 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or any combination thereof.

[0026] The storage device 120 may be used to store data and / or instructions. For example, the storage device 120 may be used to store landscape images captured by the sensor device 140. The storage device 120 may obtain data and / or instructions from, for example, the server 110 or the user terminal 130. In some embodiments, the storage device 120 may store data and / or instructions that the processing device 112 executes or uses to perform the exemplary methods described herein.

[0027] The user terminal 130 may refer to a terminal used by a user to play a landscape image, input voting information, viewing requirements, and other information. For example, the user terminal 130 may include, but is not limited to, a smart phone 130-1, a tablet computer 130-2, a laptop computer 130-3, a processor 130-4, and one or any combination of other devices with input and / or output functions. In some embodiments, the user using the user terminal 130 may be one or more users, including users directly using the service and other related users.

[0028] The sensing device 140 may refer to a device for acquiring a landscape image. For example, the sensing device 140 may include, but is not limited to, a landscape photographing device 140-1 and a drone photographing device 140-2. In some embodiments, the landscape photographing device 140-1 may be an infrared camera or a high-definition digital camera. In some embodiments, the drone photographing device 140-2 may be an unpiloted aircraft controlled by a radio remote control device. For example, the drone photographing device 140-2 may include a multi-rotor drone, an unmanned helicopter, a solar drone, etc. In some embodiments, the sensing device 140 may be configured as one or more object sub-platforms of the Internet of Things. The landscape photographing device 140-1 is a landscape photographing device sub-platform, and the drone photographing device 140-2 is a drone photographing device sub-platform.

[0029] The Internet of Things gateway 150 may refer to a data channel and gateway for the user terminal 130 and / or the sensing device 140 to upload monitoring data. For example, the Internet of Things gateway 150 may include, but is not limited to, a landscape photographing device Internet of Things gateway 150-1, a drone photographing device Internet of Things gateway 150-2, and a user terminal Internet of Things gateway 150-3. In some embodiments, the landscape photographing device 140-1 may upload a landscape image through the landscape photographing device Internet of Things gateway 150-1. The drone photographing device 140-2 may upload a landscape image through the drone photographing device Internet of Things gateway 150-2. The user terminal 130 may upload a landscape image through the user terminal Internet of Things gateway 150-3. In some embodiments, the server 110 may issue a control instruction through the landscape photographing device Internet of Things gateway 150-1 and control the operation of the landscape photographing device 140-1. In some embodiments, the server 110 may issue a control instruction through the drone photographing device Internet of Things gateway 150-2 and control the operation of the drone photographing device 140-2.

[0030] Network 160 can provide a channel for information and / or data exchange. In some embodiments, information can be exchanged between server 110, storage device 120, user terminal 130, sensing device 140, IoT gateway 150, and landscape image 170 via network 160. For example, server 110 can receive voting information sent by user terminal 130 via network 160. As another example, server 110 can obtain landscape image information uploaded by sensing device 140 via network 160 and store it in storage device 120.

[0031] Landscape image 170 can refer to images of various landscapes collected by sensing device 140. For example, landscape images can include various natural landscape images or cultural landscape images. In some embodiments, landscape image 170 can be a landscape image collected by sensing device 140 from a certain shooting angle. Different landscape images 170 can correspond to the same landscape or different landscapes. By way of example only, landscape image 170 can include landscape image 170-1 collected from the front of a rockery, landscape image 170-2 collected from the back of the rockery, and landscape image 170-3 collected from the side of the rockery. As another example, landscape image 170 can include landscape image 170-1 of a zoo, landscape image 170-2 of a natural scenic area, and landscape image 170-3 of the urban architectural style. For different landscapes or landscape images obtained from different angles, the landscape acquisition parameters can be different to enable the acquired landscape effect to meet the actual needs of users.

[0032] It should be noted that the application scenarios are provided for illustrative purposes only and are not intended to limit the scope of this specification. Those of ordinary skill in the art can make various modifications or changes according to the description of this specification. For example, the application scenario can also include a database. As another example, the application scenario can be implemented on other devices to achieve similar or different functions. However, the changes and modifications will not depart from the scope of this specification.

[0033] The IoT system is an information processing system that includes some or all of the platforms such as the object platform, sensing network platform, management platform, service platform, and user platform. The management platform can coordinate the connections and collaborations between various functional platforms (such as the sensing network platform and the object platform), gather the information of the IoT operation system, and provide sensing management and control management functions for the IoT operation system. The sensing network platform can connect the management platform and the object platform and play the functions of sensing communication for perception information and sensing communication for control information. The object platform is a functional platform that executes the generated perception information and control information. The service platform refers to a platform that provides input and output services for users. The user platform refers to a platform led by users, including obtaining user needs and feeding back information to users.

[0034] The processing of information in the Internet of Things system can be divided into the processing flow of sensed information and the processing flow of control information. The control information can be information generated based on the sensed information. Among them, the processing of sensed information is that the object platform obtains the sensed information and transmits it to the management platform through the sensing network platform. The control information is sent from the management platform to the object platform through the sensing network platform, so as to realize the control of the corresponding object.

[0035] In some embodiments, when the Internet of Things system is applied to urban management, it can be called a smart city Internet of Things system.

[0036] Figure 2 It is an exemplary schematic diagram of the smart city public landscape live broadcast control Internet of Things system shown in some embodiments of this specification.

[0037] As Figure 2 shown, the smart city public landscape live broadcast control Internet of Things system 200 may include a user platform 210, a service platform 220, a management platform 230, a sensing network platform 240, and an object platform 250.

[0038] The user platform 210 refers to a user-led platform, including a platform for obtaining user needs and feeding back information to the user. For example, the user platform can watch the landscape picture through a user terminal (for example, user terminal 130). For another example, the user platform can obtain the voting information of the user through the user terminal, and then control the landscape shooting device (for example, landscape shooting device 140-1) and / or the drone shooting device (for example, drone shooting device 140-2). For another example, the user platform can feed back information such as the user's viewing duration to the server 110.

[0039] In some embodiments, the user platform is configured to obtain at least one user intention based on a willingness acquisition strategy and determine the overall user intention corresponding to the user platform; and determine the camera parameters of the object platform corresponding to the user platform according to the overall user intention. Among them, different object platforms correspond to different camera devices, and the landscape pictures played by the user platform are collected by the corresponding object platforms. At least one user intention includes adjustment opinions on the landscape picture. For more descriptions on determining the overall user intention, refer to the content of step 330.

[0040] The service platform 220 is a platform that provides input and output services for users. The service platform is configured to sequentially transmit the camera parameters to the target platform corresponding to the user platform based on the management platform and the sensor network platform. For example, the service platform can obtain voting information sent by users through the user platform and feedback the voting results to the users. In some embodiments, the service platform may include multiple service sub-platforms, and the service platform uses different service sub-platforms to store, process, and / or transmit data sent by different user platforms.

[0041] The Public Landscape Live Broadcast Control and Management Platform 230 coordinates and coordinates the connections and collaborations between various functional platforms, aggregating all IoT information and providing perception, management, and control capabilities for the IoT operating system. For example, the Public Landscape Live Broadcast Control and Management Platform can obtain all user voting information for the current time period within a preset area, determine camera parameters based on overall user preferences, and adjust the landscape imagery captured by the object platform 250.

[0042] In some embodiments, the public landscape live broadcast control management platform may include a comprehensive management information database (i.e., a general management database) and multiple management sub-platforms. In some embodiments, each management sub-platform may also be equipped with its corresponding sub-database for storing data and instructions received by its corresponding management sub-platform.

[0043] The sensor network platform 240 is a functional platform that connects the public landscape live broadcast control and management platform and the object platform, and performs sensory communication of perception information and sensory communication of control information. In some embodiments, the sensor network platform can be configured as an Internet of Things gateway (e.g., Internet of Things gateway 150). It can be used to establish a channel for uploading perception information and distributing control information between user terminals (e.g., user terminal 130) and / or sensor devices (e.g., sensor device 140) and the public landscape live broadcast control and management platform.

[0044] In some embodiments, sensor network platform 240 can be configured as a standalone structure. This means that the sensor network platform utilizes different sensor network sub-platforms (also known as sensor network sub-platforms or sensor network sub-platforms) to store, process, and / or transmit data from different target platforms. For example, each sensor network sub-platform can correspond to multiple target platforms, and sensor network platform 240 can acquire landscape images uploaded by each target platform and upload them to the management platform.

[0045] The object platform 250 refers to the functional platform where perception information is generated and control information is finally executed. In some embodiments, the object platform may be configured as a landscape shooting device or a drone shooting device. In some embodiments, the object platform is configured to obtain a landscape image and transmit the landscape image to the corresponding user platform based on the sensing network platform, the management platform, and the service platform in sequence. For more descriptions of the landscape image, reference may be made to the relevant content of step 310. In some embodiments, the object platform is also configured to obtain a new landscape image according to the shooting parameters. In some embodiments, the object platform may be classified into multiple object platforms based on different sensing devices, and each object platform corresponds to a viewing position for viewing image and data collection.

[0046] Some embodiments of this specification also provide a computer-readable storage medium. The storage medium stores computer instructions, and when the computer instructions are executed by a processor, a smart city public landscape live broadcast control method is implemented.

[0047] It should be noted that the above description of a smart city public landscape live broadcast control Internet of Things system and its internal modules is only for convenience of description and does not limit this specification to the scope of the exemplified embodiments. It can be understood that for those skilled in the art, after understanding the principle of the system, they may, without departing from this principle, make any combination of the various modules, or form a subsystem and connect it with other modules. In some embodiments, Figure 2 the user platform 210, the service platform 220, the management platform 230, the sensing network platform 240, and the object platform 250 disclosed in

[0048] Figure 3 is an exemplary flowchart of the smart city public landscape live broadcast control method shown in some embodiments of this specification. In some embodiments, process 300 may be executed by the smart city public landscape live broadcast control Internet of Things system 200.

[0049] As Figure 3 shown, process 300 includes the following steps:

[0050] Step 310, obtain a landscape image. In some embodiments, step 310 may be executed by the object platform 250.

[0051] A landscape can refer to the view of a certain area presented to the public. For example, a landscape can include public activity places (such as parks, zoos), cultural scenic spots (such as the Forbidden City, the Great Wall, etc.). In some embodiments, a landscape can also include the view of specific items or buildings such as exhibits (such as museum collections, etc.), landmark facilities and buildings.

[0052] In some embodiments, to facilitate civil services, a landscape can also include the working view of social service institutions. For example, it can include the working conditions of publicly accessible places such as relief stations.

[0053] A landscape picture can be the specific view presented by a landscape. For example, a landscape picture can include a real-time image of a building taken from a specific position and angle. For another example, a landscape picture can include the display picture of a handicraft on a display stand.

[0054] In some embodiments, a landscape picture can be obtained through the sensing devices (such as sensing device 140) built into the object platform. For example, the sensing device can obtain the real-time video file and real-time audio file of a specific landscape as the landscape picture of that landscape. For another example, the sensing device can obtain the video file of a specific landscape (such as a collection) and store it in a storage device (such as storage device 120), and call the corresponding video file as the landscape picture when the user needs it.

[0055] In some embodiments, a landscape picture can be determined based on user needs. For example, the user platform can correspond to each landscape, and the user can determine the landscape they are viewing by selecting the user platform (such as the live broadcast rooms of each landscape).

[0056] Step 320, transfer the landscape picture to the corresponding user platform.

[0057] In some embodiments. The landscape pictures obtained by the object platform can be forwarded to the corresponding user platforms based on the sensing network platform, management platform, and service platform in sequence. For example, the user platform can correspond to the landscape identification information (such as landscape ID). After obtaining the landscape picture, it can determine the user platform (such as the live broadcast room of that landscape) corresponding to the landscape according to the landscape ID of the obtained landscape, and transfer the landscape picture to the corresponding user platform along the transmission path.

[0058] In some embodiments, sensing devices are deployed at multiple points of a public landscape. The picture taken by each sensing device corresponds to a sub-picture, and a landscape picture of an object can be composed of multiple sub-pictures. For example, a landscape picture can include the live broadcast picture of the trial of a publicly accessible civil case, and this landscape picture can include sub-pictures of the judge, both defense lawyers, both parties, the court recorder, etc.

[0059] In some embodiments, each sub - screen can be presented in the corresponding user platform according to a preset presentation rule. For example, each sub - screen can be presented in the corresponding user platform according to a preset presentation area, where the sub - screen with the largest presentation area can be called the main view. For another example, the user can adjust the presentation area of each sub - screen by himself. Exemplarily, for the live broadcast screen of a civil case trial, the preset main view can be the sub - screen of the judge, and on the user platform, the user can adjust the presentation area of each sub - screen by himself to re - determine the main view or hide some sub - screens.

[0060] In some embodiments, the user platform can present only some sub - screens. For example, when the user platform responds to a user request to present a landscape screen, it can first present the main view in the sub - screens. The user can actively adjust the presentation of each sub - screen. In some embodiments, the user platform can also set different sub - screens in different threads. After the user enters the landscape live broadcast, the main view screen can be presented first, and other sub - screens can be presented in other threads.

[0061] In some embodiments, the main view can be determined according to the actual viewing situation of each user platform. For example, based on the number of viewers of each sub - screen, the sub - screen with the largest number of viewers can be determined as the main view. For another example, based on the average stay time of the audience on each sub - screen, the sub - screen with the longest average stay time can be used as the main view. For another example, various indicators of the actual viewing situation can also be weighted to determine the main view.

[0062] Step 330, obtain at least one user intention based on the intention acquisition strategy and determine the overall user intention corresponding to the user platform. In some embodiments, step 330 can be executed by the user platform 210.

[0063] The user intention can be the user's active feedback on the landscape screen, where the user intention includes the adjustment opinion on the landscape screen. In some embodiments, the user intention can include various types of adjustment opinions. For example, it can include the adjustment opinion on each landscape and the adjustment opinion on the sub - screen of the landscape.

[0064] The adjustment opinion can refer to the adjustment suggestion for the landscape screen. For example, the adjustment opinion can include the suggestion of increasing or decreasing the sensing device, the shooting situation suggestion of the landscape screen, the landscape screen that the user wants to view, etc. Exemplarily, the user's adjustment opinion can be reflected as the hope that the picture is enlarged, the camera angle is adjusted, the hope to see the picture of "feeding food", etc.

[0065] The overall user preference can be the statistical result of the user preferences of all users watching the current landscape video. Exemplarily, in the user preference, the adjustment opinions on the video of photographing the giraffe house in the zoo landscape can be counted, and the overall user preference of the user platform corresponding to playing the video of the giraffe house can be determined.

[0066] In some embodiments, a user feedback button can be set on the user platform, and users can feedback their user preferences through the user feedback button. Among them, the feedback button can be provided with feedback options and a text input box. The feedback options can list common adjustment opinions, and the text input box can be used to input literal feedback opinions. After the user feedbacks their user preference through the feedback button, the user platform can determine the user preference of the user according to the selection of the feedback options and the literal feedback opinions.

[0067] The preference acquisition strategy can refer to the method of determining the overall user preference. For example, the preference acquisition strategy can include a statistical strategy, and the user platform 210 can process the user preferences based on the statistical strategy to determine the overall user preference. Exemplarily, for a specific landscape video, its user preferences can include the adjustment opinions of users A, B, and C. Among them, user A wants to zoom in on the video, and users B and C want to zoom out the video. Then the overall user preference can include 1 person wanting to zoom in on the video and 2 people wanting to zoom out the video.

[0068] In some embodiments, the preference acquisition strategy includes the preference acquisition frequency.

[0069] The preference acquisition frequency can refer to the number of times of determining the overall user preference per unit time. For example, the preference acquisition frequency can be once every 5 minutes, then the user preferences are acquired every 5 minutes, and the overall user preference is determined based on the user preferences.

[0070] In some embodiments, the preference acquisition frequency is related to at least one of the number of viewers of the landscape video and the viewer activity.

[0071] The number of viewers of the landscape video can refer to the number of people watching a specific landscape video or sub-video. For example, it can be determined by the number of users of the user platform watching each landscape video. In some embodiments, the number of viewers of the landscape video can also refer to the active number of people, that is, the number of viewers of the landscape video does not include the number of users who watch the landscape video but have not performed other operations for a long time (such as users who have not performed operations on the viewing terminal for 20 minutes).

[0072] The viewer activity can refer to the participation degree of the viewing users during the playback process. Among them, the viewer activity can be determined by the number of interactions of the user with the landscape live broadcast (such as interaction behaviors such as comments, likes, rewards, feedback adjustment opinions, etc.). The more the number of interactions, the higher the viewer activity.

[0073] In some embodiments, the higher the number of viewers of the landscape video and the higher the viewer activity, the smaller the opinion acquisition frequency. For example, when the number of viewers is 10,000 and the viewer activity is high (such as more than 5 million likes per minute), the opinion acquisition frequency can be once per minute. When the number of viewers is 100 and the viewer activity is low (such as less than 10 likes per minute), the opinion acquisition frequency can be once per 10 minutes.

[0074] In some embodiments, the overall user intention may further include an overall adjustment opinion. The overall adjustment opinion may be an opinion on the processing and adjustment of the landscape video. For example, the overall adjustment opinion may be embodied as the adjustment opinion with the largest number of supporters among all adjustment opinions. Exemplarily, when the overall user intention includes 1 person wanting to zoom in on the video and 2 people wanting to zoom out on the video, and the total number of people wanting to zoom out is more than the total number of people wanting to zoom in, the overall adjustment opinion may be embodied as zooming out on the video.

[0075] In some embodiments, the intention acquisition strategy includes a user intention weighting rule. Each adjustment opinion can be processed based on the user intention weighting rule, and the overall adjustment opinion can be determined according to the weighted adjustment opinions. For more descriptions of the user intention weighting rule, reference can be made to Figure 4 the relevant content.

[0076] Step 340: Determine the camera parameters of the target platform corresponding to the user platform according to the overall user intention. In some embodiments, step 340 may be executed by the user platform 210.

[0077] In some embodiments, different target platforms correspond to different camera devices, and the landscape videos played by the user platform are collected by their corresponding target platforms. When adjusting the landscape video according to the overall user intention, it can be achieved by adjusting the camera parameters of the corresponding target platform. The corresponding relationship between the landscape video and the camera device may be pre-stored in the smart city public landscape live broadcast control IoT system 200, and the camera device (such as the unique device ID of the camera device) can be directly determined according to the corresponding relationship during adjustment.

[0078] The camera parameters may refer to the parameters when the camera device captures the landscape image. For example, the camera parameters may include relevant parameters such as the camera ID, camera type, camera position, camera angle, camera content, zoom ratio, exposure and white balance, and resolution.

[0079] When determining the camera parameters according to the overall user intention, the corresponding camera parameters can be determined according to the specific content of the overall adjustment opinion. For example, when the overall adjustment opinion includes zooming in on the video, it can be achieved by adjusting the zoom ratio and / or the camera position. When the overall adjustment opinion includes increasing the resolution, it can be achieved by adjusting the camera lens or replacing the camera device.

[0080] In some embodiments, when the current camera device cannot meet the overall user's desired camera parameters, the camera device can be replaced to meet the camera parameters. For example, when the current camera device cannot meet the clarity requirement in the camera parameters, a camera device with higher resolution can be replaced.

[0081] In step 350 , the service platform sequentially transmits the camera parameters to the object platform corresponding to the user platform based on the management platform and the sensor network platform.

[0082] In some embodiments, after the user platform determines the camera parameters, the camera parameters can be transmitted to the corresponding target platform of the user platform through the service platform, the management platform, and the sensor network platform. For example, the user platform can have a one-to-one correspondence with the camera device (e.g., the user platform can be bound to the unique identification code of the camera device), and the camera parameters can be sent to the corresponding target platform based on the corresponding relationship.

[0083] In some embodiments, the camera parameters may include camera parameters of multiple camera devices. For example, the camera parameters may include camera parameters of a landscape image and each sub-image.

[0084] In some embodiments, when the current camera device cannot achieve the corresponding camera parameters, the camera device can be replaced. For example, the correspondence between the camera device and the user platform can be cancelled, and a camera device that can meet the camera parameters can be associated with the user platform to capture and transmit the corresponding landscape image.

[0085] Step 360 , obtaining a new landscape image according to the camera parameters. In some embodiments, step 360 may be performed by the object platform 250 .

[0086] In some embodiments, the new landscape image may be a landscape image captured by the camera device after the camera parameters are updated.

[0087] A smart city public landscape live broadcast control method provided based on some embodiments of this specification can conduct statistics and analysis on user intentions to determine the overall user intentions, thereby adjusting the landscape picture, improving user satisfaction with public landscape live broadcasts, and taking into account the user intentions of the vast majority of users.

[0088] Figure 4 FIG4 is an exemplary flow chart of determining overall user intention according to some embodiments of the present specification. In some embodiments, process 400 may be executed by the user platform 210.

[0089] like Figure 4 As shown, process 400 may include the following steps:

[0090] Step 410: Acquire at least one user intention based on the intention acquisition frequency.

[0091] In some embodiments, a willing acquisition period may be determined according to the willing acquisition frequency as desired, and the user willing within the willing acquisition period may be counted with the willing acquisition period as the statistical range. Among them, the willing acquisition period may be the interval time for counting user willing, and the willing acquisition period may be the reciprocal of the willing acquisition frequency. For example, if the willing acquisition frequency is once every 5 minutes, the willing acquisition period may be 5 minutes.

[0092] Step 420, determine the willing weight value of each user corresponding to each user willing among at least one user willing based on the user willing weighting rule.

[0093] The willing weight value may reflect the influence of the adjustment opinion in the overall adjustment opinion. In some embodiments, the weight may be any value between 0 and 1. In some embodiments, the willing weight value may include the weight of each adjustment opinion in the user willing.

[0094] The willing weighting rule may refer to the method of determining the weight of the adjustment opinion. Among them, the willing weighting rule may process the relevant data of the user willing to determine the willing weight value of the user willing. For example, the willing weighting rule may determine the willing weight value according to the identity of each user. Among them, for the general audience (such as users without special identities), their willing weight value may be a conventional value (such as 0.5), and for the audience with special identities, their willing weight value may not be a conventional value. For example, for the feedback personnel invited officially, their willing weight value may be greater than the conventional value (such as 1), and for the personnel who disrupt the order of the live broadcast room, their willing weight value may be less than the conventional value (such as 0.1 or 0).

[0095] In some embodiments, the user willing weighting rule may include a viewing duration rule. Among them, in the viewing duration rule, the willing weight value of the user is related to the duration of the user viewing the landscape picture. That is, the longer the duration of the user viewing the landscape picture, the higher the willing weight value of the user. For example, the audience with the longest viewing duration of the landscape picture within a preset time period may be used as the benchmark, and the weights of other users may be converted according to the viewing time. Exemplarily, the preset time period may be 8:00 - 9:00, then the longest viewing duration during this period may be 1 hour, and the willing weight value of the audience who views the landscape picture for 1 hour may be set to 1, and the willing weight values of other viewing durations may be the viewing duration / 1 hour. For example, if only viewing for 0.5 hours, then its willing weight value is 0.5.

[0096] In some embodiments, the user willing weighting rule includes an execution ratio rule. In the execution ratio rule, the willing weight value of the user is related to the ratio of the user willing being successfully executed. That is, the higher the ratio of the user willing being successfully executed, the higher the willing weight value of the user.

[0097] The successful execution of the user's will can refer to the adjustment opinions or user's will put forward by the user in the past being accepted by the user platform and regarded as the overall user's will. The ratio of the successful execution of the user's will can refer to the proportion of the accepted adjustment opinions or user's will in the total number put forward by the user.

[0098] In some embodiments, the user will weighting rule includes an anti-community weighting rule, and the will weight value of the anti-community weighting rule is the anti-community weighting value obtained based on the contact map. Among them, the anti-community weighting rule can refer to a weighting rule to avoid the influence of user groups on the result, and the anti-community weighting rule can be realized by reducing the weights of similar users in the same group. For more content about the anti-community weighting rule, reference can be made to 5 and its related content.

[0099] In some embodiments, the will acquisition strategy can be predicted and determined based on the first prediction model. For example, each will acquisition strategy can be input into the first prediction model to determine the impact of executing each candidate will acquisition strategy on the number of viewers, so as to determine the target will acquisition strategy among the candidate will acquisition strategies. The first prediction model can be a machine learning model. For more descriptions about the first prediction model, reference can be made to Figure 6 and its related content.

[0100] Step 430, based on the at least one user will and the will weight value, determine the overall user will corresponding to the user platform.

[0101] In some embodiments, the user will can be weighted based on the will weight value to obtain the weighted user will, and the weighted results of various types of user wills can be determined through each weighted user will, and the user will with the largest value in the weighted results can be used as the overall user will. For example, the user will of 5 users can be to indent the focal length, and the user will of 10 users can be to stretch the focal length. The weighted results of each user will can be 4.2 people indenting the focal length and 7.1 people stretching the focal length, then stretching the focal length can be selected as the overall user will.

[0102] Based on the method for determining the overall user will provided in some embodiments of this specification, the user wills of each user can be fully considered, and the influence of each user on the overall user is quantitatively analyzed through weights. Thus, the overall user will can reflect the will of most users, and further improve the live broadcast effect of the landscape live broadcast.

[0103] Figure 5 It is an exemplary flowchart for determining the will weight value shown in some embodiments of this specification. In some embodiments, process 500 can be executed by user platform 210.

[0104] As Figure 5 shown, process 500 can include the following steps:

[0105] Step 510: Obtain a connection graph, wherein the connection graph includes edges between nodes.

[0106] A connection graph may refer to a database having a graph structure, wherein the connection graph may include nodes, node features, edges between nodes, and edge features.

[0107] Nodes can correspond to individual users. Figure 5 As shown, the contact graph may include nodes A through G, where each node corresponds to a user viewing the live broadcast via the user platform. In some embodiments, the contact graph may include nodes corresponding to all users who currently and historically viewed the live broadcast via the user platform. In some embodiments, to determine the current overall user intent, the user currently expressing the user intent may be extracted from all users as a node in the contact graph.

[0108] In some embodiments, a node may include node features. Node features may reflect user-related information. For example, node features may include the user's address, work unit, social media interactions, viewing habits, etc.

[0109] In some embodiments, a node can be determined by the user identification information of the corresponding user (such as user ID, ID number, mobile phone number, etc.), and the node characteristics can be retrieved from the relevant database (such as social platform, etc.) based on the user identification information as the node characteristics. For example, when a user enters the public landscape live broadcast, the smart city public landscape live broadcast control IoT system can obtain the user ID from the user platform (such as the social network that the user logged in to) and request to retrieve the corresponding relevant data (such as mobile phone number, location, work unit, social software usage, etc.) from the social network to determine the node characteristics.

[0110] Edges can reflect the mutual influence between the connected nodes. For example, the users corresponding to the two nodes connected by an edge may be influenced by each other, and the user intentions of both parties may be similar or the same. Figure 5 As shown, the contact graph may have edges AB, BG, AG, CD, and EF, which means that the user intentions of users A, B, and G may be influenced by each other and be the same or similar, the user intentions of users C and D may be influenced by each other and be the same or similar, and the user intentions of users E and F may be influenced by each other and be the same or similar.

[0111] Edge features can reflect the mutual influence between nodes. For example, edge features can reflect the likelihood of mutual influence between nodes. In some embodiments, edge features can be characterized by the closeness between nodes. The closeness can be determined by the similarity of node features (or partial features) between two nodes.

[0112] In some embodiments, edge features may include overall closeness, address closeness, social media interaction, unit closeness, and viewing habit closeness. In some embodiments, edge features may also be described by the similarity of historical user intentions between nodes, where the similarity of historical user intentions may be the similarity between user intentions provided by each node during the historical intention acquisition cycle.

[0113] Address closeness can be determined by whether the nodes are located in the same location. For example, "location consistency" can mean that two nodes are located in the same neighborhood, building, or unit. Different situations will result in different levels of address closeness. For example, the address closeness corresponding to being in the same unit is greater than the address closeness corresponding to being in the same neighborhood.

[0114] The degree of social media interaction can be determined based on the number of social media interactions between two nodes. For example, a high ratio of likes to comments indicates a high degree of social media interaction.

[0115] Unit closeness can be determined by the similarity between the units of two nodes and the length of time they have worked together. The longer they have worked together in the same unit, the higher the unit closeness. The length of time worked together can be determined based on the date of joining the company, and the unit of a node can be determined based on the social security contributions paid by the node.

[0116] The degree of similarity in viewing habits can be determined by the historical viewing records of the two nodes. The higher the similarity in the historical viewing records, the higher the degree of similarity in viewing habits.

[0117] The overall closeness may be a statistical result determined by processing address closeness, social media interaction, unit closeness, and viewing habit closeness using statistical methods such as weighted summation and average.

[0118] In some embodiments, when determining a contact graph, the nodes of the contact graph can be first constructed based on the user, and then each node can be traversed and the overall closeness between each node can be determined based on the node characteristics, and edges can be constructed between nodes whose overall precision is greater than a threshold, and each type of closeness can be used as the edge feature of the corresponding edge.

[0119] Step 520: Obtain user intentions of each node based on the contact graph.

[0120] The user intentions of a node may refer to the user intentions of the user corresponding to each node in the contact graph. For example, the user intentions of a node may include the user intentions of the corresponding user in the current intention acquisition cycle. In some embodiments, the user intentions of each node may also include the historical user intentions of the corresponding user in historical data (e.g., historical intention acquisition cycles).

[0121] In some embodiments, user intentions can be described by a voting vector. The voting vector can include multiple elements and the attributes of the corresponding elements. Each element corresponds to various types of user intentions, and the corresponding element value can be the specific situation of the corresponding intention.

[0122] In some embodiments, various types of user intentions can refer to various adjustment methods of the landscape picture. For example, zooming in / out the landscape picture, whether to change the content of the landscape picture (the user hopes to see other landscape pictures), etc. The specific meaning of the element value can be related to the corresponding picture adjustment method. For example, when the element represents zooming in / out the landscape picture, the corresponding element value can include -2, -1, 0, 1, 2. Among them, 0 can represent no change, positive numbers can reflect zooming in the focal length, negative numbers can reflect zooming out the focal length, and the absolute value of the element value can reflect the degree of change.

[0123] In some embodiments, the user intentions of each node can include the current user voting vector and the historical user voting vector. The current user voting vector can be a voting vector determined based on the user intention of the user corresponding to the node in the current intention acquisition period, and the historical user voting vector can be a voting vector determined based on the historical data of the user corresponding to the node (such as the user intention in the historical intention acquisition period of the current landscape picture).

[0124] In some embodiments, various types of user intentions can be characterized as votes for different options. Each vote can be regarded as selecting one or more from multiple options, and each option can be represented by a number. For example, "1" represents selecting the first option, "2" represents selecting the second option, and so on. Exemplarily, in a certain intention acquisition period, the user intention can include that option 1 can be to zoom in the landscape picture, option 2 can be to zoom out the landscape picture, and option 3 can be to keep the landscape picture unchanged. Then the user voting vector can include the option voting situation of the user in each intention acquisition period. For example, the user voting vector can include the option votes of the user in the recent five times, then the voting vector can be a vector of five elements, where each element corresponds to each vote. Exemplarily, when the user voting vector V is (1, 2, 4, 1, 3), it means that the user's voting behavior in the recent five votes is: selecting the 1st option, selecting the 2nd option, selecting the 4th option, selecting the 1st option, and selecting the 3rd option.

[0125] In some embodiments, the voting on different options can be determined based on the statistics of user preferences. For example, the specific preferences of various types of user preferences of the user can be counted, and each type of user preference can be regarded as a vote, and the corresponding specific preference is used as the corresponding option. In some embodiments, the voting on different options can also be determined based on the user preferences periodically collected by the user platform. For example, the user platform can periodically pop up a user preference questionnaire, which can include at least one vote and its corresponding options, and the user can fill in the questionnaire to determine the options for the user's current vote.

[0126] Step 530: Determine the willingness similarity between nodes based on the user preferences of each node.

[0127] The willingness similarity between nodes can refer to the degree of similarity of user preferences between each node where there is an edge in the contact graph.

[0128] In some embodiments, the willingness similarity between nodes can be determined according to the distance between the voting vectors of the nodes. For example, the willingness similarity between nodes can be inversely proportional to the distance between the voting vectors of the nodes. For example, the farther the distance between the voting vectors, the smaller the similarity.

[0129] In some embodiments, the willingness similarity between nodes can be used as the node similarity. In some embodiments, the node similarity can be determined based on a normalization function, where the function value of the normalization function is within the interval [0,1], when the distance of the voting vector is 0, the function value is 1, the normalization function is monotonically decreasing, and the function value is 0 when the distance of the voting vector approaches infinity.

[0130] For example, the node similarity can be determined according to the following formula.

[0131] S ij =1 / 1+|V i -V j |

[0132] Where S ij is the node similarity between node i and node j, V i can refer to the voting vector of node i, V j can refer to the voting vector of node j, |V i -V j | can reflect the distance between V i and V j (such as the Euclidean distance).

[0133] Step 540: Iteratively update the willingness weight value of the node based on the willingness similarity between nodes and the contact graph to determine the target willingness weight value of the node.

[0134] The target willingness weight value of a node can be the overall output of process 500, and the target willingness weight value of the node can be used as the willingness weight value of each user in process 400.

[0135] In some embodiments, the iteration of the willingness weight value of a node can be implemented based on a preset algorithm. In some embodiments, the preset algorithm can include a pre-trained machine learning algorithm. By iterating the contact graph, the willingness weight value of the node with an edge in the contact graph is reduced according to the edge feature. For example, the willingness weight value of a node can be iterated by a Graph Neural Network (GNN) according to the contact graph and the willingness similarity between nodes.

[0136] In some embodiments, the preset algorithm can include an anti-community iteration algorithm. The anti-community iteration algorithm can iterate the willingness weight value based on the following formula:

[0137] U i ′ = U i - U i × f(∑ k∈M(i) S ik U k R ik )

[0138] Wherein, U i ′ represents the iterative value of the willingness weight value of node i after this iteration, and U i represents the initial value of the willingness weight value of node i before this iteration. For example, the initial value of the willingness weight value in the first round of iteration can take a default value, etc., and the initial value of the willingness weight value in the second round of iteration is the iterative value of the willingness weight value obtained in the first round of iteration.

[0139] M(i) represents the first-degree adjacent nodes of node i. The first-degree adjacent nodes are other nodes directly connected to this node by an edge. S ik represents the willingness similarity between node k and node i among the first-degree adjacent nodes of node i. R ik represents the edge feature between node k and node i among the first-degree adjacent nodes of node i. Wherein, R ik can be characterized by the closeness (such as the overall closeness) between node i and node k. For the description of the closeness, refer to the content corresponding to step 510. f(·) can be a mapping function that maps the independent variable to the weight region [0, 1]. For example, f(x) = 1 / (1 + x^2).

[0140] In some embodiments, the iterative algorithm based on U i can be iterated for a specified number of rounds (such as 3 rounds) to determine the iterated U i ′ as the target willingness weight value of node i.

[0141] In some embodiments, before iterating on the willingness weight values of each node, the initial values of the willingness weight values of each node can be determined first.

[0142] In some embodiments, the initial values of the willingness weight values can be preset. For example, the willingness weight values of each node can all be preset values (such as 0.5). In some embodiments, the initial values of the willingness weight values can be determined based on other user willingness weighting rules. For example, user willingness weighting rules such as viewing duration rules, execution ratio rules, user identity rules, etc. can be used to process node features to determine the initial values of the willingness weight values of the nodes.

[0143] Based on the method for determining willingness weight values provided in some embodiments of this specification, it is possible to prevent users from forming groups to vote uniformly, disrupting the free voting environment, and thus affecting the landscape live broadcast environment. Thereby improving the stability of the landscape live broadcast.

[0144] Figure 6 It is an exemplary structural diagram of the first prediction model shown in some embodiments of this specification.

[0145] As Figure 6 shown, when determining the target willingness acquisition strategy 650 based on the first prediction model 610, the to-be-tested willingness acquisition strategy 620 and the number of live broadcast viewers 630 of the live broadcast room corresponding to the to-be-tested willingness acquisition strategy (which can also be understood as the current number of users using the corresponding user platform) can be input into the first prediction model 610. After being processed by the first prediction model 610, the predicted change in the number of live broadcast viewers 640 of the live broadcast room after adopting the to-be-tested willingness acquisition strategy 620 is output. And based on the possible changes in the number of viewers 640 caused by each to-be-tested willingness acquisition strategy 620, the target willingness acquisition strategy 650 is determined from the to-be-tested willingness acquisition strategies 620.

[0146] The first prediction model 610 can be a trained machine learning model. For example, the first prediction model 610 can be a trained deep neural network (DNN), convolutional neural network (CNN), recursive neural network (RNN), or a similar machine learning model.

[0147] In some embodiments, the first prediction model 610 can include algorithms related to determining the overall user willingness. Inputting the to-be-tested willingness acquisition strategy 620 into the first prediction model 610 can be to adjust the specific parameters of the algorithms related to the overall user willingness.

[0148] The to-be-tested willingness acquisition strategy 620 can be various candidate willingness acquisition strategies. The to-be-tested willingness acquisition strategy 620 can determine the willingness weight values of each user so as to determine the overall user willingness. In some embodiments, the to-be-tested willingness acquisition strategy 620 can also include relevant data such as willingness acquisition frequency, willingness acquisition period, etc.

[0149] In some embodiments, at least one corresponding overall user willingness can be determined based on each to-be-tested willingness acquisition strategy 620. Inputting the to-be-tested willingness acquisition strategy 620 into the first prediction model 610 can be understood as inputting at least one overall user willingness corresponding to the to-be-tested willingness acquisition strategy 620 into the first prediction model 610, so that the first prediction model 610 estimates the change in the number of viewers 640 according to the overall user willingness. At this time, the first prediction model 610 can estimate the change in the number of viewers 640 based on the overall user willingness and the number of viewers in the live broadcast room 630.

[0150] The number of viewers in the live broadcast room 630 can refer to the number of viewers before the current scene is adjusted based on the overall user willingness. For example, the number of viewers in the live broadcast room 630 can be the total number of viewers within a willingness acquisition period.

[0151] The change in the number of viewers 640 can refer to the change in the number of people in the live broadcast room within a preset time after the to-be-tested willingness acquisition strategy 620 is executed. For example, the change in the number of viewers 640 can refer to the total change in the number of viewers within a willingness acquisition period after the to-be-tested willingness acquisition strategy 620 is executed.

[0152] In some embodiments, considering that the natural growth rate of the live broadcast room has little relation with the to-be-tested willingness acquisition strategy 620, the change in the number of viewers 640 can specifically refer to the change in the number of people in the live broadcast room caused by the execution of the to-be-tested willingness acquisition strategy 620. For example, the change in the number of viewers 640 can reflect the situation of the current audience leaving the live broadcast room within the willingness acquisition period after the to-be-tested willingness acquisition strategy 620 is executed.

[0153] The target willingness acquisition strategy 650 can be the strategy with the highest user satisfaction among the to-be-tested willingness acquisition strategies 620. In some embodiments, among the changes in the number of viewers 640 caused by each to-be-tested willingness acquisition strategy 620, the to-be-tested willingness acquisition strategy with the largest positive change (or the smallest negative change) in the number of viewers in the change in the number of viewers 640 can be taken as the willingness acquisition strategy to be executed.

[0154] In some embodiments, the first prediction model 610 can be determined by training an initial first prediction model based on historical data. Among them, the initial first prediction model can be a first prediction model without set parameters.

[0155] Historical data may include training samples and training labels. Among them, the training samples may include the number of viewers in the live broadcast room at a historical moment and the willingness acquisition strategy at the historical moment, and the training labels may include the change in the number of people in the live broadcast room within a preset time at the historical moment. During training, the training samples may be input into the initial first prediction model to determine the model output, the model output and the training labels may be input into the loss function, and the initial first prediction model may be iteratively adjusted based on the loss function until the training is completed. The initially trained first prediction model is used as the first prediction model 610. Among them, the completion of training may include that the number of iterations exceeds a threshold, the deviation value of the model output converges, etc.

[0156] Based on the method for determining the willingness acquisition strategy based on the first prediction model provided by some embodiments of this specification, appropriate willingness acquisition strategies can be automatically selected for different live broadcast situations, solving the problem of the adaptation of the same willingness acquisition strategy to different live broadcast situations. Furthermore, the live broadcast effect of the landscape live broadcast is improved.

[0157] Figure 7 It is an exemplary flowchart of the adjustment method of the imaging device shown according to some embodiments of this specification. In some embodiments, process 700 may be executed by the user platform 210.

[0158] As Figure 7 shown, process 700 may include the following steps:

[0159] Step 710, based on the service platform, count the number of viewers of the landscape images corresponding to different user platforms within a preset time length, and send the number of viewers to the management platform.

[0160] The number of viewers of the landscape image may include the total number of non-repeating viewers of the landscape image within a preset time length, the remaining number of viewers, etc. In some embodiments, the number of viewers of the landscape image may also include the number of viewers of each sub-image in the landscape image.

[0161] In some embodiments, the preset time length may include at least one of a past time period and a future time period. Among them, the number of viewers in the past time period may be determined through historical data. The number of viewers in the future time period may be determined through estimated data.

[0162] In some embodiments, the preset time length may divide the historical time period and the future time period through a start time, a current time, and an end time.

[0163] In some embodiments, when both the start time and the end time are times before the current time, the preset time length only includes a historical time period, and the number of viewers can be the total number of viewers at the start time and the end time in the historical data. In some embodiments, when both the start time and the end time are times after the current time, the preset time length only includes a future time period, and the number of viewers can be determined based on a machine learning model trained with historical data. In some embodiments, when the current time is between the start time and the end time, the preset time length can be split into a historical time period and a future time period to determine the number of viewers separately and then combined.

[0164] In some embodiments, the number of viewers of the landscape images collected at the existing camera positions in a preset future time period can be predicted based on a third prediction model processing the number of viewers of the landscape images collected at the existing camera positions in a preset historical time period. Among them, the input of the third prediction model can be the number of viewers of the landscape images collected at the camera positions in a preset historical time period. The output can be the number of viewers of the landscape images collected at the existing camera positions in a preset future time period.

[0165] In some embodiments, the third prediction model can be a trained machine learning model. For example, the third prediction model can be a trained deep neural network (DNN), convolutional neural network (CNN), recursive neural network (RNN), or a similar machine learning model.

[0166] In some embodiments, the third prediction model can be determined by training an initial third prediction model based on historical data. Among them, the initial third prediction model can be a third prediction model without set parameters.

[0167] The historical data can include training samples and training labels. Among them, the training samples can include the number of viewers of the landscape images collected at the camera positions to be predicted in a preset historical time period (such as the number of viewers from January 1st to January 7th, 2020), and the training labels can include the number of viewers of the landscape images collected at the camera positions at a certain historical time point or time period (such as January 8th, 2020). During training, the training samples can be input into the initial third prediction model to determine the model output, and the model output and the training labels can be input into a loss function. Based on the loss function, the initial third prediction model is iteratively adjusted until the training is completed. The trained initial third prediction model is used as the third prediction model. Among them, the completion of training can include that the number of iterations exceeds a threshold, the deviation value of the model output converges, etc.

[0168] Step 720, determine the camera devices to be cancelled based on the management platform.

[0169] The preset condition may be a condition that should be met when maintaining the landscape screen. For example, the preset condition may be a threshold condition for the number of viewers (e.g., 100 people). When the number of viewers is less than the threshold condition, the thread or the landscape screen does not need to continue to be open, and the camera device corresponding to the landscape screen may be canceled. For example, when the number of viewers of a sub-screen in the landscape screen does not meet the preset condition, the thread of the sub-screen may be closed, and the camera device that produced the sub-screen may be determined (e.g., the ID of the camera device may be determined based on the corresponding relationship), and the camera device may be canceled to save thread traffic.

[0170] like Figure 7 As shown, the process 700 may further include the step of adding a camera device and a sub-screen.

[0171] Step 730 : The management platform predicts the number of viewers of the landscape images captured at the candidate camera locations based on the number of viewers corresponding to the landscape images captured at the existing camera locations.

[0172] The candidate points may refer to locations where a new camera device can be added. The new camera device can capture landscape images from the candidate points.

[0173] In some embodiments, candidate locations can be determined based on user feedback. For example, a user can provide user feedback on the candidate locations they want to add. For another example, candidate locations can be determined based on other user wishes that are not included in the overall user wishes. For example, for a live broadcast of a panda breeding scene, in a process of determining the overall user wishes, most users want to see the image of panda A, and this is selected as the overall user wish. A small number of users do not want to see the image of panda B, so the observation point of panda B can be selected as a candidate location based on the users of panda B.

[0174] In some embodiments, the number of viewers can be estimated based on the number of user feedback. For example, if a candidate location has multiple user feedback, the number of viewers of the landscape image captured by the candidate location can be estimated based on the number of people who provided feedback about the candidate location. For example, if 100 people provided feedback that they wanted to add candidate location A, the number of viewers for candidate location A could be 10,000.

[0175] In some embodiments, the coordinates of existing camera locations and their corresponding number of viewers, as well as the coordinates of candidate locations, may be processed based on the second prediction model to predict the number of viewers of the landscape image captured by the candidate locations.

[0176] In some embodiments, the second prediction model may be a trained machine learning model. For example, the second prediction model may be a trained Deep Neural Networks (DNN), Convolutional Neural Networks (CNN), Recursive Neural Network (RNN), or a similar machine learning model.

[0177] In some embodiments, the second prediction model may be determined by training an initial second prediction model based on historical data. The initial second prediction model may be a second prediction model without set parameters.

[0178] The historical data may include training samples and training labels. The training samples may include the number of viewers of other landscape pictures (and their sub-pictures) before the new camera position at a historical moment, the corresponding landscape acquisition point coordinates, and the coordinates of the new camera position to be added. The training labels may include the number of viewers of the new camera position after the new camera position is added at the historical moment. The labels may be manually annotated by humans based on the historical viewership statistics data of each point. During training, the training samples may be input into the initial second prediction model to determine the model output. The model output and the training labels may be input into a loss function, and the initial second prediction model may be iteratively adjusted based on the loss function until the training is completed. The trained initial second prediction model is used as the second prediction model. Completion of training may include the number of iterations exceeding a threshold, convergence of the deviation value of the model output, etc.

[0179] Step 740: Determine the candidate points where the number of viewers meets the preset conditions as the new camera positions, and set corresponding camera devices based on the new camera positions.

[0180] In some embodiments, the preset condition may be a condition of a viewer number threshold. When the estimated number of viewers at a candidate point is greater than the viewer number threshold condition, a camera device may be set at the candidate point to obtain the corresponding landscape picture (or a sub-picture of the landscape picture). After setting the camera device, the new camera device (such as the device ID) may be bound to the sensor network sub-platform, the management sub-platform, and the user platform, and the landscape picture obtained by the camera device may be sent to the corresponding user platform.

[0181] It should be noted that the above descriptions of processes 300, 400, 500, and 700 are only for illustration and explanation, and do not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to processes 300, 400, 500, and 700 under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0182] The basic concepts have been described above. Obviously, for those skilled in the art, the above detailed disclosure is only an example and does not constitute a limitation to this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification, so such modifications, improvements, and corrections still fall within the spirit and scope of the exemplary embodiments of this specification.

[0183] Meanwhile, this specification uses specific terms to describe the embodiments of this specification. Such as "one embodiment", "an embodiment", and / or "some embodiments" mean a certain feature, structure, or characteristic related to at least one embodiment of this specification. Therefore, it should be emphasized and noted that the "one embodiment" or "an embodiment" or "an alternative embodiment" mentioned twice or more at different positions in this specification does not necessarily refer to the same embodiment. In addition, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0184] In addition, unless clearly stated in the claims, the order of the processing elements and sequences, the use of numerical letters, or the use of other names in this specification is not used to limit the order of the processes and methods of this specification. Although some currently considered useful invention embodiments are discussed through various examples in the above disclosure, it should be understood that such details only serve the purpose of illustration. The appended claims are not limited to the disclosed embodiments. On the contrary, the claims are intended to cover all modifications and equivalent combinations that conform to the essence and scope of the embodiments of this specification. For example, although the system components described above can be implemented by hardware devices, they can also be implemented only through software solutions, such as installing the described system on existing servers or mobile devices.

[0185] Similarly, it should be noted that, in order to simplify the expression of the disclosure of this specification and thus help the understanding of one or more invention embodiments, in the previous description of the embodiments of this specification, sometimes multiple features are merged into one embodiment, drawing, or description thereof. However, this disclosure method does not mean that the features required by the subject of this specification are more than those mentioned in the claims. In fact, the features of the embodiments are less than all the features of the individual embodiments disclosed above.

[0186] In some embodiments, numbers are used to describe components and the quantity of attributes. It should be understood that such numbers used in the description of embodiments are modified by the modifiers "about", "approximate" or "substantially" in some examples. Unless otherwise specified, "about", "approximate" or "substantially" indicate that the said numbers allow a variation of ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, and such approximate values may vary according to the characteristics required by individual embodiments. In some embodiments, the numerical parameters should consider the specified significant digits and adopt the method of retaining the general number of digits. Although the numerical ranges and parameters used in some embodiments of this specification to confirm the breadth of their scope are approximate values, in specific embodiments, such numerical settings are as precise as possible within the feasible range.

[0187] For each patent, patent application, patent application publication, and other materials cited in this specification, such as articles, books, specifications, publications, documents, etc., their entire contents are hereby incorporated into this specification by reference. This excludes the application history documents that are inconsistent with or conflict with the content of this specification, as well as the documents that limit the broadest scope of the claims of this specification (currently or subsequently appended to this specification). It should be noted that if there are inconsistencies or conflicts between the descriptions, definitions, and / or uses of terms in the supplementary materials of this specification and the content described in this specification, the descriptions, definitions, and / or uses of terms in this specification shall prevail.

[0188] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, by way of example and not limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. A management method for public landscape camera devices in a smart city, which is implemented based on the Internet of Things system for the management of public landscape camera devices in a smart city. The Internet of Things system for the management of public landscape camera devices in a smart city includes a user platform, a service platform, a management platform, and an object platform; among them, The number of the user platforms is multiple, and different user platforms respectively play the landscape pictures collected by different camera devices; The number of the object platforms is multiple, and different object platforms respectively correspond to the camera devices arranged at different camera positions; The method includes: Based on the service platform, counting the number of viewers of the landscape pictures corresponding to different user platforms within a preset future time period; and sending the number of viewers to the management platform; wherein, the number of viewers within the preset future time period is determined based on a third prediction model for the number of viewers of the landscape pictures in a preset historical time period, and the third prediction model is a machine learning model; Through the user platform, obtaining at least one user intention based on an intention acquisition strategy and determining the overall user intention corresponding to the user platform; the at least one user intention includes adjustment opinions on the landscape pictures; the intention acquisition strategy includes an intention acquisition frequency and a user intention weighting rule; the user intention weighting rule includes a viewing duration rule, an execution ratio rule, and an anti-community weighting rule, and the anti-community weighting rule refers to a weighting rule for avoiding the influence of user groups on the result; Through the user platform, determining the camera parameters of the object platform corresponding to the user platform according to the overall user intention; Based on the object platform, obtaining new landscape pictures according to the camera parameters; Based on the management platform, determining the camera devices to be cancelled, where the camera devices to be cancelled are the camera devices corresponding to the landscape pictures whose number of viewers does not meet the preset conditions.

2. The method according to claim 1, the method further includes: Through the management platform, based on the number of viewers of the landscape pictures collected at the existing camera positions, predicting the number of viewers of the landscape pictures collected at the candidate positions through a second prediction model; the second prediction model is a machine learning model; Determining the candidate positions where the number of viewers meets the preset conditions as new camera positions; Setting corresponding camera devices based on the new camera positions.

3. The method according to claim 1, The step of obtaining at least one user intention based on an intention acquisition strategy and determining the overall user intention corresponding to the user platform through the user platform includes: Through the user platform, obtaining the at least one user intention based on the intention acquisition frequency; Determining the intention weight value of each user intention of the at least one user intention corresponding to the user based on the user intention weighting rule; wherein, the intention weight value in the viewing duration rule is related to the viewing duration of the landscape pictures, the intention weight value in the execution ratio rule is related to the ratio of successful execution of the user intention, and the intention weight value in the anti-community weighting rule is an anti-community weighting value obtained based on a contact map; the contact map includes edges between nodes, the nodes correspond to the users who currently propose user intentions, the node features include the user's address, workplace, social media interaction situation, and viewing habits, there is an interaction between the two nodes connected by the edge, and the edge feature includes the tightness; Determine the overall user intention corresponding to the user platform based on the at least one user intention and the intention weight value.

4. An Internet of Things system for managing public landscape camera devices in a smart city, the system comprising a user platform, a service platform, a management platform, and an object platform; wherein, There are multiple user platforms, and different user platforms play landscape images collected by different camera devices; there are multiple object platforms, and different object platforms correspond to the camera devices arranged at different camera positions; The object platform is used for: Collect the landscape image based on the camera device; Obtain a new landscape image according to the camera parameters; The user platform is used for: Play the landscape image, where the landscape image is collected by its corresponding object platform; Obtain at least one user intention based on the intention acquisition strategy and determine the overall user intention corresponding to the user platform; the at least one user intention includes adjustment opinions on the landscape image; the intention acquisition strategy includes an intention acquisition frequency and a user intention weighting rule; the user intention weighting rule includes a viewing duration rule, an execution ratio rule, and an anti-community weighting rule, and the anti-community weighting rule refers to a weighting rule that avoids the influence of user groups on the result; Determine the camera parameters of the object platform corresponding to the user platform according to the overall user intention; The service platform is used for: Count the number of viewers of the landscape images corresponding to different user platforms within a preset future time period; and send the number of viewers to the management platform; where the number of viewers within the preset future time period is determined based on the number of viewers of the landscape images in a preset historical time period by a third prediction model, and the third prediction model is a machine learning model; The management platform is used for: Determine the camera devices to be cancelled, where the camera devices to be cancelled are the camera devices corresponding to the landscape images whose number of viewers does not meet the preset conditions.

5. The system according to claim 4, the management platform is further used for: Based on the number of viewers corresponding to the landscape images collected at the existing camera positions by the management platform, predict the number of viewers of the landscape images collected at the candidate positions through a second prediction model; the second prediction model is a machine learning model; Determine the candidate positions where the number of viewers meets the preset conditions as new camera positions; Set corresponding camera devices based on the new camera positions.

6. The system according to claim 4, The user platform is further used for: Obtain the at least one user intention based on the intention acquisition frequency; Determine the willingness weight value of the user corresponding to each user willingness in the at least one user willingness based on the user willingness weighting rule; wherein, The intention weight value in the viewing duration rule is related to the duration of viewing the landscape image, the intention weight value in the execution ratio rule is related to the ratio of successful execution of the user intention, and the intention weight value in the anti-community weighting rule is an anti-community weighting value obtained based on the contact map; the contact map includes edges between nodes, the nodes correspond to the users who currently propose user intentions, the node features include the user's address, unit, social media interaction situation, viewing habits, and there is an interaction between the two nodes connected by the edge, and the edge features include the degree of closeness; Based on the at least one user preference and the preference weight value, determine the overall user preference corresponding to the user platform.

7. A management device for a public landscape camera device in a smart city, the device includes at least one processor and at least one memory; the at least one memory is used to store computer instructions; the at least one processor is used to execute at least some of the computer instructions to implement the method according to any one of claims 1-3.

8. A computer-readable storage medium, characterized in that, The storage medium stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1-3 is implemented.

Images