294 results about "ControlNet" patented technology

The invention relates to the technical field of plasma control systems, in particular to a method for real-time synchronization and online switching of a redundancy control system. According to the technical scheme, the method comprises the following steps: system interaction: realizing interaction between a redundancy control system and an external system through a control network, a clock synchronization network, a central interlocking network and a real-time communication network; the main and standby systems synchronously run under the triggering of a unified clock; the main and standby systems share key data in real time, and time, data and operation states are synchronized regularly. Synchronous operation of the main control system and the standby control system is achieved, the operation states of key devices and processes of the control systems are monitored in real time, the interlocking unit is used for judging the effectiveness of the main redundant system and the auxiliary redundant system, online switching and abnormal alarm of the systems are completed when the devices or the processes are abnormal, and real-time release and sharing of key operation information of the control systems are achieved. State synchronization after the offline system is accessed and operation recovery of a redundant system are realized, so that the reliability of long-time steady-state operation of a real-time control system is improved.

The invention belongs to the technical field of computers, particularly relates to an AIGC paper-cut generation system and method based on culture gene multi-modal constraints, and solves the dual problems of'culture distortion 'and'physically unachievable' of a traditional AIGC generation result through mathematical fusion (such as a mixed integer programming model) of semantic, topology and process three-level constraints. The improved ControlNet architecture realizes directional fusion of culture gene coding vectors through a culture attention mechanism, and dynamically adjusts feature weights in a generation process, so that the generation precision of pattern details is improved to a submillimeter level; based on the combination of an interactive correction tool and an inverse diffusion inversion algorithm, the modification of a local pattern by a user can be quickly fed back to a system, and a'creativity-generation-optimization 'closed loop is realized.

The invention discloses a control gateway for multi-protocol conversion and an equipment access and data transmission method thereof, and belongs to the technical field of protocol conversion, and the control gateway comprises a protocol conversion module which is used for identifying a plurality of communication protocols and carrying out bidirectional conversion on different communication protocols; the data processing module is used for performing error detection and correction on the data and preprocessing the data; the security management module is used for providing multi-level security protection, including encryption, identity authentication, authorization and network security protection, for data transmission and equipment access; the flow control module is used for dynamically adjusting the transmission rate, the priority and the resource allocation of the data according to the network state and the user configuration; the equipment management module is used for managing equipment connected to the gateway; and the user interface module is used for providing an interactive interface and an API (Application Program Interface) of a user and the gateway. Through architecture design and function arrangement, an efficient, safe and flexible solution is provided for communication between networks and devices of different protocols.

The invention relates to a commodity picture generation method and device, computer equipment and a storage medium. The method comprises the following steps: receiving a to-be-processed commodity graph; determining a target commodity from the to-be-processed commodity graph, and extracting a commodity main body image of the target commodity; obtaining first description information of the commodity main body image, wherein the first description information is used for describing the target commodity; the commodity main body image and the first description information are input into a diffusion model to generate a target commodity graph containing the commodity main body image, the diffusion model comprises a pre-trained LoRA model and a ControlNet model, the LoRA model is used for generating a scene in the target commodity graph, and the ControlNet model is used for generating the scene in the target commodity graph. The ControlNet model is used for controlling the output quality of the diffusion model. By adopting the method, the high-quality commodity display image can be accurately generated, and the high consistency of the commodity main body in the generated image and the original image is ensured.

The invention discloses a control method and system for improving a transient stability area of a grid-connected inverter. It includes performing park coordinate transformation on a three-phase voltage and a three-phase output current of a connection point of a grid-connected inverter to obtain an actual voltage input value and an actual current input value required for current control; obtaining the grid-connected voltage phase through phase-locked loop processing; generating, by the power control module, the current command value required by the current control module; using the sliding mode current control method, by the current control module, to generate a dq voltage signal, and obtaining PWM modulated voltage control signals through inverse park transformation; using a LCL filter to filter out high-order harmonics to achieve grid-connected control. The present invention effectively improves the transient stability of the grid-connected inverter by using the sliding mode control method in the current controller.

A communication method is described that receives, by a policy control network element, a quality of service (QoS) request message from an application function network element. The QoS request message comprises a QoS requirement of a service of the application function network element. The policy control network element based on the QoS requirement, determines M policy and charging control (PCC) rules corresponding to the service, wherein N of the M PCC rules comprise first indication information, the first indication information indicates data flows of the service corresponding to the PCC rules comprising the first indication information to share a bit rate of the service. The N PCC rules are mapped to at least two QoS flows used to carry the data flows of the service, M and N are integers greater than 1, and N is less than or equal to M. The policy control network element sends the M PCC rules to a session management network element.

The invention discloses a dynamic split mirror generation system and method based on a controllable diffusion model, and belongs to the technical field of film and television production. The implementation method comprises the following steps of: 1, setting a keyword text by a user, and inputting the keyword text into ChatGPT to generate a script; 2, converting the script into a scene category, a camera motion mode, a character role position and action and image description in a shot language by utilizing ChatGPT; 3, using a CLIP model to carry out contrast training on the image encoder and the text encoder; 4, performing an OpenPose model on the action reference image to obtain skeleton key points of the character role, and converting the skeleton key points into image character actions; 5, performing action fine-grained control on the character action of the image by using a Stable Diffusion model and a ControlNet model, and generating a split image; 6, utilizing a Pika Labs model to generate a dynamic video from the split image; compared with the prior art, the accuracy of user role action matching under the scene based on multi-text and high-complexity actions is improved.

The invention relates to an energy storage power station control strategy updating method based on a cloud-side cooperation technology, and the method comprises the steps: starting to call a data processing program when an edge calculation gateway collects data or receives alarm information or a task processing request actively transmitted by an energy storage power station; an edge computing gateway data processing program judges the type of a currently executed task, if the currently executed task is alarm information, an alarm information analysis processing program is executed, if the currently executed task is a computing task actively requested by the energy storage power station, a computing task analysis processing program is executed, and if the currently executed task is a high-complexity computing task, a cloud server is requested to assist computing; the two-way updating method of the control strategy is designed, both the edge control gateway and the cloud server can actively initiate strategy updating, and compared with an edge-side or cloud-side-based one-way strategy updating method in the prior art, the method provided by the invention has more applicability.

The invention discloses a monocular 3D target detection method and device, and relates to the technical field of computer vision, and the method comprises the steps: obtaining a to-be-detected target detection image; inputting the target detection image into a pre-trained target detection model to obtain target detection information of the target detection image; wherein the training process of the target detection model comprises the following steps of: adding a multi-modal feature fusion module in a ControlNet model architecture to establish a Fusion-ControlNet model architecture; introducing the Fusion-ControlNet model architecture as a guidance module for condition generation into the Stable Diffusion model architecture to obtain a data set expansion model; and performing expansion processing on the KITTI data set by using the data set expansion model to obtain a KITTI expanded data set. According to the invention, the problem of low accuracy during monocular 3D target identification in the prior art is solved.

The invention provides a security protection method, a cloud access method, a network architecture, a device and equipment, and relates to the technical field of cloud computing. The method comprises the steps that under the condition that a first condition is met, a port unblocking request is sent to a second controller, and the port unblocking request is used for requesting to unblock a port of a target virtual public network IP address corresponding to a user instance by a public network gateway in a first available area; the first available area and the second available area adopt different SDN structures, and share one exit router and one NAT gateway with the second available area; the first controller is used for controlling the second controller and the third controller, the second controller is used for controlling the public network gateway, the third controller is used for controlling the NAT gateway, and the public network gateway is provided with a port blocking strategy aiming at a virtual public network IP address corresponding to a public network IP address bound with the virtual machine in the first available area. According to the scheme, the problem that the public network access strategy in the heterogeneous available area is invalid is solved.

The invention provides an addressing method for a strategy control network element, an application side network element and a control plane network element, and the method comprises the steps: obtaining a device address of a strategy binding network element retrieval database network element based on the local static data configuration information of the application side network element; retrieving a device address of a database network element according to the policy binding network element, and sending the control plane network element query message to a corresponding target policy binding network element retrieval database network element, so as to query an address of a target control plane network element in the target policy binding network element retrieval database network element according to the control plane network element query information; according to the address of the target control plane network element, sending the policy control network element address query request to the target control plane network element; and receiving a query result corresponding to the policy control network element address query request sent by the target control plane network element. According to the invention, the problem that the application side network element cannot acquire the IP address of the policy control network element according to the user IP address information is avoided.

The invention belongs to the technical field of image processing, and discloses a facial expression migration method and system based on a diffusion model and facial key points, and the method comprises the steps: obtaining a training sample pair comprising a source face key point, a target face image, and an expression migration result; a text encoder of the original stable diffusion model is replaced with a Swin Transform module, and an improved stable diffusion model is obtained; copying to obtain a copy network block according to a pre-training neural network block of the improved stable diffusion model, and adding a zero convolutional layer into the copy network block to obtain a ControlNet architecture; building an expression migration model according to the improved stable diffusion model and a ControlNet framework; and based on the training sample pair, training an expression migration model by using a control loss function and an identity maintenance loss function depending on a time step length to generate an expression migration result with identity features reserved. According to the method, the facial expressions of the source face image can be migrated to the target faces with different expressions, and meanwhile, the identity information of the target faces is reserved.

The invention provides a policy control network element addressing method, an application side network element and a control plane network element, and the method comprises the steps: obtaining the addresses of a plurality of control plane network elements based on the local static data configuration information of the application side network element according to a user terminal address and a mobile communication network access external network identifier; determining a first target control plane network element, and sending a policy control network element address query request to the first target control plane network element; if the session association information of the user terminal is not queried in the first target control plane network element, acquiring an address of a second target control plane network element sent by the first target control plane network element; sending a policy control network element address query request to the second target control plane network element according to the address of the second target control plane network element; and receiving a query result sent by the second target control plane network element. According to the invention, the problem that the application side network element cannot acquire the IP address of the policy control network element according to the user IP address information is avoided.

The invention relates to a method and a device for generating a sewing processing graphic file by text description, mainly based on a ComfyUI server, a Stable Diffusion model and a ControlNet model are combined for use, a descriptor is converted into a picture, and a picture contour is extracted to generate a stitch path. And the stitch server optimizes a needle point path through an algorithm, and generates a binary data packet containing coordinates, a connection sequence and a null shift instruction. And after receiving the data packet, the sewing machine equipment decodes the data packet, reconstructs a path and performs processing operation, and finally sews a pattern corresponding to the descriptor on the material. According to the method, full-process automation from characters to file processing is achieved, the design efficiency is remarkably improved, the use threshold is lowered, and the method is suitable for the fields of personalized design, intelligent manufacturing and the like and has wide application value and innovative significance.

Provided are a service processing method, apparatus and device, a storage medium and a program product, applied to the technical field of communications, the method comprising: an artificial intelligence control network element receiving an artificial intelligence (AI) service session creation request of a terminal sent by an access and mobility management function (AMF), the AI service session creation request carries terminal position information and AI service information; based on the terminal position information and the AI service information, distributing a corresponding quality of service (QoS) strategy for the AI service, and determining an artificial intelligence service network element meeting the AI service requirement; sending an AI service session establishment request to an artificial intelligence service network element; receiving an AI service session establishment response sent by the artificial intelligence service network element, wherein the AI service session establishment response carries an uplink tunnel address corresponding to the AI service; and sending an AI service session creation result to the AMF, wherein the AI service session creation result carries the uplink tunnel address. The method can provide effective QoS guarantee for the AI service of the terminal.

The invention relates to a sectional type positioning installation method and system for a steel structure frame, and relates to the technical field of building structure construction. The method comprises the steps that three-dimensional coordinate measurement is conducted on the embedded bolt through a total station, a deviation value recording table is obtained, and a control network and a marking system are established; the method comprises the following steps: performing ultrasonic flaw detection and pre-tension reinspection on a steel structural member, performing ground pre-assembly to obtain a single-section steel frame model, and performing optimization treatment to form a number marking system and an assembly sequence diagram; a steel structure is divided into a plurality of independent installation units, and preliminary positioning is achieved through a temporary fixing technology; secondary measurement is carried out, fine adjustment is carried out through an ejector and a wire drawing device, accumulated deviation is dispersed through an overall balance method, and the pre-camber is set to obtain a stable frame system; the high-strength bolts are finally tightened, and welding construction is conducted; the roof system is installed in a batched concentrated hoisting mode. The problem that accumulated errors are difficult to control in traditional steel structure installation is solved, and the installation precision and efficiency are improved.

The invention discloses an artificial intelligence image generation system suitable for updating a historical block facade, and the system employs a Flux model loader module, an image input module, a DeepSeek language interaction module, a LoRA module, a sampler module, a Controlnet processing module, a graph generation module, and a mobile terminal applet module to form a ComfyUI workflow, and completes the generation of a street reconstruction image through the workflow. Wherein the language model can convert natural language input of a common user into a professional street design drawing, more convenient and low-threshold image generation is realized, the ControlNet module can ensure spatial rationality of a generation result, and the LoRA model performs special training on local features to generate a street reconstruction image conforming to local culture and architectural styles. According to the method, the street reconstruction scheme conforming to the local characteristics can be conveniently and quickly obtained, and the design efficiency and interactivity are greatly improved.

A precision measurement and tamping operation method for an operating conventional-speed railway without a control network. The method comprises the following steps: S100, laying a temporary reference station (1) and performing initial precision track measurement; S200, processing office-processed data; and S300, performing tamping operation and precision measurement. The temporary reference station (1) is laid offline along an operation route by means of precise single-point positioning; when precision track measurement is performed by using a track inspector (2), the temporary reference station (1) may serve as a reference station and cooperate with an onboard moving station (3) to perform dynamic post-processing calculation, such that accurate coordinates of the center line of the route are obtained; and when precision tamping is performed by using a tamping vehicle (4) equipped with a tamping vehicle locator (5), the temporary reference station (1) may serve as a main reference station to provide real-time dynamic positioning for the tamping vehicle (4), thereby realizing centimeter-level mileage positioning accuracy, and realizing autonomous synchronization with field mileage.

The invention discloses an improved plant disease data enhancement method, which comprises the following steps: S1, introducing a plant disease data set and a potential diffusion model, taking each plant picture as the input of the potential diffusion model, and taking a plurality of groups of preset prompt texts as the text prompt conditions of the potential diffusion model; s2, extracting image features in the plant picture by using a ControlNet algorithm, and taking the image features as guidance to adjust the output of the potential diffusion model so as to preliminarily optimize the potential diffusion model; s3, maintaining an original weight matrix of the potential diffusion model unchanged by using an LORA optimization algorithm, and updating low-rank decomposition of the potential diffusion model so as to optimize the potential diffusion model again; and S4, extracting text features in each prompt text by using the text embedding guidance model, and embedding the text features into the denoising network of the potential diffusion model to further optimize the potential diffusion model. The quality of the generated plant disease data is improved.

The invention discloses an AI-based text cartoon device and method, and the method comprises the steps: extracting image parameters through a script analysis module, combining with the training of a LoRA model and a Stable Diffusion frame, achieving the multi-dimensional control of the standing position, action and scene style of a character through a Regional Prompter plug-in, a ControlNet plug-in and an Adtailer plug-in, and generating a key frame picture with high consistency; and the local redrawing module repairs details through segmentation and Inpaining technologies, and finally synthesizes a cartoon or a video. According to the method, the problems of inconsistent roles and incoherent scenes of pictures generated by an existing AI tool are solved, the controllability and efficiency of the generated content are remarkably improved, the manual adjustment cost is reduced, and the method is suitable for automatic generation of cartoon creation and story explanation videos.

The invention provides a metadata processing acceleration method, electronic equipment and a storage medium, and relates to the technical field of distributed storage. According to the technical scheme, the acceleration card obtains the metadata request, the metadata is cached to the cache module of the hybrid storage architecture based on the dynamic cache management strategy, and the acceleration card directly queries the cache and controls the network card to directly construct the response packet to return when the cache is hit, so that the data processing efficiency is improved. According to the method and the device, the problems of high delay and CPU resource scrambling and the like caused by the fact that a metadata request depends on a CPU to analyze a network protocol and inquire software cache in the prior art can be solved, so that the technical effects of remarkably reducing metadata processing delay, improving metadata throughput, reducing CPU occupancy rate, optimizing hardware resource utilization rate and reducing cross-node synchronous bandwidth occupancy can be achieved.

Provided in the present disclosure are a terminal communication method, a network element device, a terminal, an electronic device, a computer-readable storage medium and a computer program product. The method comprises: receiving a registration request message initiated by a terminal when communicating with a non-terrestrial network, the registration request message carrying available duration; according to the available duration, configuring an available duration timer; and, when the available duration timer overruns, releasing the connection between a network access control network element and the terminal. The present invention prevents a large number of terminals from simultaneously initiating signaling interactions to a network access control network element when being about to disconnect from a non-geostationary transmission device, thus solving the problem of signaling storms. In addition, the present disclosure controls the release of connections to the terminals by means of the available duration timer, so as to further improve the ability of a network side to control the availability of the terminals, thus avoiding the loss of terminal context due to discontinuous coverage.

The invention relates to a stability analysis method for control points in a tunnel in the technical field of tunnel measurement, which comprises the following steps of: S1, establishing a wire control network and a leveling network in the tunnel, and remeasuring the control points; s2, selecting a traverse point as an analysis point, comparing the two groups of horizontal angle data obtained after retest, and judging whether the two groups of horizontal angle data meet the standard or not; s3, establishing a deviation limiting system of two adjacent control points, wherein the deviation limiting system comprises a side length deviation standard, an elevation deviation standard and a difference relative precision standard of coordinate differences; s4, taking two control points adjacent to the analysis point as two target points, comparing the deviation limiting systems of the three control points pairwise, and integrating the judgment condition that the horizontal angle is relatively poor to obtain the stability condition of the analysis point; according to the method, comprehensive analysis can be carried out from multiple aspects, and the accuracy of stability judgment is improved.

The invention provides a data flow control device and method, a storage medium and a computer program product. The device comprises a strategy control network element; the strategy control network element is used for determining a data use strategy corresponding to the first data according to the data demand information of the first data; and deploying the data use strategy on the data flow network element corresponding to the first data, so that the data flow network element uses the first data based on the data use strategy.

The invention relates to the technical field of image restoration, in particular to a diffusion model image restoration method based on regional mask and dynamic exit, and mainly solves the technical problems of high calculation overhead and iteration step number redundancy of an existing diffusion model image restoration method. According to the method, on the basis of an extension architecture of encoding and decoding, a UNet denoising network and a ControlNet conditional constraint module, on the basis of retaining the core functions of an existing encoder, an existing decoder and an existing Unet denoising network, a space region mask generation sub-module, a back diffusion dynamic exit module and the ControlNet conditional constraint module are creatively integrated; and a triple optimization architecture of'constrain structure protection, space redundancy reduction and time step number reduction 'is formed, so that the structure consistency is met and efficient acceleration is realized during image restoration.

The invention provides a control network constraint adjustment method and system based on an RANSAC data selection algorithm, and the method comprises the steps: obtaining known starting data of a control network after determining that free network adjustment is completed; performing constraint adjustment pre-operation based on the known starting data by using an RANSAC (Random Sample Consensus) algorithm, and selecting target starting data in combination with an operation result; and the actual constraint adjustment operation of the control network is implemented based on the selected target starting data. By adopting the scheme, the known starting data is utilized to carry out constraint adjustment on the GNSS control network, the starting data is obtained through calculation and is compared with the known data, gross error values can be efficiently and quickly eliminated, and the starting data is quickly and accurately determined preferentially while engineering requirements are met.

The present invention discloses a policy control method, apparatus and system, relating to the field of mobile communications. The policy control method includes: in response to a terminal switching from a first system to a second system, an access control network element of the second system determines whether policy retention ends according to policy retention parameters, where the first system is a 4G system and the second system is a 5G system, or the first system is a 5G system and the second system is a 4G system; in the case where policy retention does not end, the access control network element of the second system does not send updated policy control parameters to the terminal and the radio base station to which the terminal is connected. Embodiments of the present invention enable the terminal to maintain the original policy control parameters for a period of time by using policy retention parameters after the terminal switches systems, thereby improving the stability of terminal access.

The invention discloses a cloud environment VPN service disaster recovery method, a product, equipment and a medium, and relates to the technical field of cloud computing, and the method comprises the steps: periodically synchronizing the control plane data of a cloud management main environment in a cloud management platform to a cloud management standby environment; when the cloud management platform currently meets a preset cloud environment switching condition, starting a disaster recovery switching process of the cloud management platform; destroying the cloud management service of the cloud management main environment, and starting the cloud management service of the cloud management standby environment to switch the cloud management standby environment into a new main environment; drifting a mapping relation between a logic address of a message queue in the cloud management platform and an original physical address based on the physical address of the new main environment, and updating an OVN-SB physical address set recorded in a VPN gateway node based on a physical address set of each management node in the new main environment; therefore, the VPN gateway node is controlled to communicate with the new main environment and perform VPN service. When the cloud management environment is switched, the VPN gateway node can adapt to the change of the address of the cloud management environment so as to carry out VPN service.

The invention provides an embedded part group precision control method for wind tunnel construction, and belongs to the technical field of wind tunnel construction, and the method comprises the steps: building an initial control network, dividing embedded part groups in a segmented manner, and determining the position of an optimal control point through an iteration experiment method and a graph theory network model; establishing an error propagation mathematical model to describe a measurement error propagation rule, performing control point position optimization calculation by applying a double-layer game optimization algorithm and a cross subwolf optimization algorithm, calculating three-dimensional coordinates of the embedded part by combining a resection method and a forward intersection method, and implementing real-time monitoring and dynamic adjustment of the laser tracker; and finally, overall precision verification is carried out through the total station and the laser scanner, and the technical problem that accumulated error propagation is difficult to effectively control in the installation process of the wind tunnel embedded part group is solved.

The invention discloses a corn leaf disease image generation method and device based on a ControlNet and LoRA enhanced Stable Diffusion model, and relates to a corn leaf disease image generation method and a corn leaf disease image generation device based on a ControlNet and LoRA enhanced Stable Diffusion model. Belongs to the technical field of image processing. According to the method, a data set covering seven types of corn leaf diseases is constructed, and a Stable Diffusion model integrating ControlNet and LoRA is trained on the basis of the data set so as to generate an image of a corresponding disease type. Afterwards, the generated image is combined with a real image to be used for training a Swin Transform model and generating a vivid corn leaf disease image, the problems of scarcity of field data and insufficient diversity are solved, and the effectiveness of the image in a disease recognition task is verified through quantitative evaluation.