59results about How to "Reduce bandwidth usage" patented technology

The invention discloses an industrial equipment real-time monitoring system based on edge computing, and the system comprises the following steps: collecting, processing and analyzing the operation data of industrial equipment through distributed edge computing nodes, constructing a node cooperation mechanism, and completing the processing of the data; the dynamic data scheduling module is used for dynamically optimizing a data transmission path and task allocation according to node performance; the intelligent model dynamic migration module is used for adjusting lightweight model deployment in combination with node resource conditions; the multi-modal fault sensing module is used for extracting and analyzing multi-modal features by using a cross-node collaborative algorithm; the edge node self-adaptive management module is used for adjusting node operation parameters in real time and repairing an abnormal state; and the equipment relevance reasoning module is used for constructing an equipment relevance model through a graph calculation method and generating a real-time relevance graph. According to the method, efficient monitoring and anomaly detection of industrial equipment are realized through resource optimization and intelligent analysis of collaborative edge computing nodes.

The invention provides a multi-path embedded video coding method and system based on generative artificial intelligence. The multi-path embedded video coding method comprises the following steps: firstly, coding input cue words and demand parameters, and outputting cue word vectors and demand parameter code streams; then decoding the cue word vector and the demand parameter code stream, and outputting a feature vector and a feature map; extracting the feature vector, the feature map and the representation guide vector to obtain multi-embedded feature information; reconstructing a picture by using the obtained multi-embedded feature information through a generative AI and obtaining a video to be iterated; performing video quality evaluation on the video to be iterated; and generating a binary code stream for transmission through the current cue word vector, the demand parameter code stream and the multi-embedded feature information input. According to the method, important information characteristics are transmitted in a highly abstracted manner, and a high-definition video is reconstructed at a receiving end, so that the storage space required by video coding and the size of a transmission information stream are greatly reduced, and higher-definition transmission under a narrower bandwidth is realized or video coding under an extremely low bandwidth becomes possible.

The invention provides a video monitoring platform adaptive play support optimization method, which comprises the following steps that: a server acquires equipment information and user habits of a terminal, and correspondingly configures a plurality of types of video stream preprocessing strategies; and the server selects a video stream preprocessing strategy based on the prediction operation, and pulls the preprocessed video stream to the corresponding terminal in advance. Wherein when the video picture proportion of the stream pushed by the server is not matched with the picture proportion of the terminal player, the server carries out cutting and scaling on the pushed video stream based on a black edge removing algorithm. According to the method, efficient self-adaptive stream pushing is realized by predicting user habits and combining the preset strategy, on one hand, adaptive coding video streams and code rates can be pushed according to requirements, so that bandwidth occupation is reduced, playing delay is reduced, and terminal types are expanded; and on the other hand, the video stream matched with the picture proportion is pushed by using the black edge removing algorithm in a feedback mode on the server side, so that the effective information of the picture is reserved to the maximum extent, and the playing pressure of the terminal can be reduced.

The invention relates to an image transmission method and device based on layered coding and a storage medium, and the method comprises the steps: a coding unit receives a YUV444 format image, carries out the transcoding and chroma extraction, and obtains a basic coding layer and a chroma enhancement layer; the transmission unit uses the basic coding layer and the chroma enhancement layer as independent coding streams and sends the coding streams to a decoding unit; a decoding unit receives and decodes the basic coding layer coding stream and the chroma enhancement layer coding stream to obtain a basic image and a color enhancement image; the decoding unit reconstructs the base image and the color enhanced image into a YUV444 format image. According to the image transmission method based on layered coding, the problem of YUV420 color distortion is solved by independently transmitting chroma high-frequency information.

The invention discloses a page picture resource progressive loading method, device and system and a storage medium, and relates to the technical field of data processing. The method comprises the steps that a timer event is set, and when a page is loaded, picture resources are searched in advance through the timer event; monitoring the page, and when the page changes, searching for picture resources needing to be asynchronously loaded; when the picture resources needing to be asynchronously loaded are found, the picture resources needing to be asynchronously loaded are replaced with thumbnails, and asynchronous loading is prepared; and monitoring the intersection condition of the picture resource and the visual area of the page, and when the picture resource is intersected with the visual area, asynchronously loading the original resource of the picture. By adopting the method provided by the embodiment of the invention, the data volume when the page is loaded for the first time is reduced, so that the loading speed is increased; repeated loading of the same picture resource is avoided, and network requests and bandwidth occupation are reduced; therefore, the best loading experience can be obtained no matter how the user interacts with the page.

The invention discloses an image transmission method, device and equipment of a remote operation system and a medium, the method is applied to a local operation end, the remote operation system comprises the local operation end and a remote operation end, and the method comprises the following steps: acquiring an initial ultrasonic image corresponding to a target object at the local operation end, performing feature extraction on the initial ultrasonic image to obtain a target ultrasonic image; determining an ultrasonic sub-image corresponding to the target ultrasonic image based on the knowledge graph, and determining an image value of the ultrasonic sub-image according to the ultrasonic image evaluation dimension; and transmitting the ultrasonic sub-image to a remote operation end according to the image value. Based on the technical scheme, the image is divided into the sub-images of different areas, the image values of the sub-images are determined, and the transmission methods corresponding to the sub-images are determined according to the image values, so that the preferential transmission of the high-value image is ensured, the overall bandwidth occupation and the system load are reduced, and the transmission efficiency is improved. And a key technical support is provided for the accuracy and safety of a remote operation.

The invention relates to the technical field of control management systems, in particular to an AI-based sports examination room system and a sports examination scoring method.The AI-based sports examination room system comprises a plurality of cameras and further comprises an area dividing unit, a track is divided into a plurality of unit recognition areas, and a starting area and a terminal area are divided by combining the distribution positions of a starting line and a terminal line of the track; the preparation identification unit is used for acquiring image data in different starting areas, analyzing and calculating to obtain a target starting line, and marking the starting area corresponding to the target starting line as a preparation area; the project confirmation unit is used for determining a plurality of to-be-determined terminal lines and to-be-determined end point areas corresponding to the preparation areas, analyzing the average speed of the testers passing through the to-be-determined end point areas and the average speed of the testers passing through the two unit identification areas before and after the to-be-determined end point areas, comparing and analyzing the speeds of the testers before and after passing through the to-be-determined terminal lines, and determining a target terminal line; and determining a test item based on the target finishing line and the target starting line.

The invention relates to image processing and video monitoring, in particular to a real-time target tracking method and system based on double-camera linkage, and the method comprises the steps: initializing the system and defining a coordinate system; when the target enters the high-definition picture of the high-definition zoom camera, judging whether the target enters the field of view of the wide-angle fixed-focus camera or not in real time; if the target enters the view field of the wide-angle fixed-focus camera, guiding the high-definition zoom camera to turn to the target by using the large view field advantage of the wide angle, and carrying out picture fusion and asymmetric weight ROI video coding; if the target does not enter the field of view of the wide-angle fixed-focus camera, starting an independent closed-loop PID tracking mode for the high-definition zoom camera, and gradually switching the high-definition zoom camera from the wide-angle end to the long-focus end to realize stable locking of the target in a full-focus range; according to the technical scheme provided by the invention, the defect that the moving target is difficult to accurately track in the prior art can be effectively overcome.

The invention relates to the technical field of Internet of Things communication, discloses an Internet of Things multi-Android client state synchronization method, and solves the problems of poor state consistency, high back-end dependence, large delay and insufficient compatibility in the prior art. According to the scheme, the method comprises the steps that after a client completes initialization, MQTT connection with a rear end is established, an MQTT state message pushed by the rear end is monitored, after the message is received, equipment state information is obtained through analysis, local state cache and a UI interface are updated, meanwhile, the client serves as a distribution node, a UDP synchronization message containing the equipment state and the IP of the distribution node is established, and the UDP synchronization message is sent to the client. The UDP synchronization message is broadcasted to the local area network, other clients in the local area network receive the UDP synchronization message and then verify whether the distribution node IP is the client in the local area network, if the verification is passed, the message is analyzed, the state updating timestamp and the local cache timestamp are compared, and the local state cache and the UI interface are updated only when the state updating timestamp is larger.

The invention discloses a data generation method of a photoetching pattern, an electron beam exposure system, a medium and a product, which are applied to the technical field of photoetching. The method is applied to an electron beam exposure system comprising an upper computer and a lower computer. The method comprises the following steps: receiving first original coordinate data and second original coordinate data of a layout transmitted by the upper computer; performing linear interpolation based on the first original coordinate data and the second original coordinate data to obtain a plurality of pieces of original intermediate coordinate data; respectively mapping the first original coordinate data, the second original coordinate data and each piece of original intermediate coordinate data into first mapping coordinate data, second mapping coordinate data and each piece of mapping intermediate coordinate data of a lower computer; and performing exposure operation based on the first mapping coordinate data, the second mapping coordinate data and the mapping intermediate coordinate data. The data transmission quantity between the upper computer and the lower computer can be reduced, the bandwidth occupation and delay of data transmission are reduced, and the exposure efficiency is improved while the exposure precision is guaranteed.

The invention belongs to the technical field of Internet of Vehicles communication, and discloses a heavy diesel vehicle remote monitoring method and system based on edge computing, and the method comprises the steps: constructing a rule engine according to a predefined transmission rule, and deploying the rule engine to an edge computing node; the data stream of the heavy diesel vehicle is transmitted to an edge computing node, a rule engine is used for evaluating each data field in the data stream, a corresponding transmission level is distributed, and the transmission level comprises a first-level field, a second-level field and a third-level field; and matching a preset transmission strategy according to the data field with the allocated transmission level, executing differentiated data processing, and transmitting a processing result to an enterprise cloud platform through a network transmission layer. According to the method, low-value data is filtered in real time through edge calculation, the invalid data uploading amount is reduced, and data transmission is reduced; in addition, through a high compression ratio and a delay transmission technology, the bandwidth occupation of low-value data is reduced, and the utilization efficiency of network resources is optimized.

The invention discloses a global nerve drawing method and system based on a programmable rasterization engine, and belongs to the technical field of computer graphics, and the method comprises the steps: at the programmable rasterization engine, analyzing a rasterization descriptor according to a rasterization instruction, and extracting vector microoperation and control parameters; maintaining a task state machine according to the parameters and distributing a control signal, selecting an execution entry from a vector kernel table according to the control signal, and instantiating an operation into a parallel vector thread; in a vector thread execution process, tracking data dependence of a vector register and a synchronization state of a direct memory access unit, executing vector loading / storage operation so as to carry data between the register and an on-chip shared memory according to the data dependence and the synchronization state, and dynamically scheduling vector micro-operation to an execution component so as to complete rasterization calculation; and outputting a result to the neural rendering network to complete global neural rendering. According to the method, the multi-representation neural rendering load can be uniformly and efficiently supported on the AI accelerator, the memory access overhead is remarkably reduced, and the calculation efficiency is improved.

The application relates to the technical field of Internet of Things data processing, and discloses a local reasoning and abnormality detection method for a wireless sensor network, which comprises the following steps: constructing a hypergraph model according to a sensor node set in the sensor network and corresponding association relations, and defining an energy function; applying a sensitivity constraint to the potential function, so that the potential function satisfies a bounded difference condition; determining a constraint range of coupling strength and a truncation order m; performing local deterministic approximate reasoning to obtain a conditional edge probability of a target sensor node; calculating an abnormality score of the target sensor node, and performing abnormality detection. In the wireless sensor network, a hypergraph Gibbs model with a sensitivity constraint is constructed, and based on strong spatial mixing ensured by theory, the application realizes the probability reasoning of a deterministic error bound only by relying on a limited-step local communication, so that the explainable abnormality detection and positioning are completed on a resource-limited edge device.

The present application belongs to the technical field of remote control of robots, and particularly relates to a 5G cross-domain low-latency remote control system and method for an explosion-proof humanoid robot. The system comprises a master end teleoperation intelligent cockpit and a slave end remote explosion-proof humanoid robot, and the two are connected through a 5G network for information interaction. The system transmits video streams through H.264 encoding and WebRTC protocol, packs sensor data through JSON, and independently processes information for humanoid control and chassis control. The present application can realize accurate remote force feedback and control, with a humanoid control latency within 100 ms and a chassis control latency within 150 ms, effectively solving the stability and latency problems of remote cross-domain control and ensuring the personal safety of operators in dangerous operating environments.

This invention discloses an on-demand telemetry path planning method for mega-constellation satellite networks. First, an on-demand telemetry mechanism is established based on a programmable data plane, realizing a logical shift from "blind telemetry across the entire network" to "precise target point perception." Second, a projection dimensionality reduction strategy is used to eliminate non-target nodes, significantly reducing the computational complexity in large-scale networks. Next, a Hamiltonian cycle structure is used to unify the start and end points of each telemetry path, ensuring that a single closed path only requires initiating a telemetry process once at the source node, significantly reducing system process maintenance overhead. Finally, based on the capacity limit of a single telemetry message, the initial global Hamiltonian cycle is truncated to obtain a multi-loop planning scheme that satisfies the telemetry message capacity constraint, completing the on-demand telemetry path planning and thus constructing an efficient and low-overhead telemetry path planning scheme.

The application provides a network test transmission optimization method and a test system, and belongs to the technical field of network testing.The network test transmission optimization method of the application transmits only the 4-layer header and load description information, compared with transmitting complete data packets, the data transmission amount of the PCIe bus is greatly reduced, and the bandwidth occupation of the PCIe bus is significantly reduced;the FPGA intelligent network card locally completes load generation, assembly and detection, does not need to rely on the CPU processing of the mainboard, reduces the data interaction delay, and improves the concurrent test capability of the test instrument.

This application relates to the field of natural language processing, and more particularly to a method for ultra-long context tensor parallel incremental attention and differential compression transmission. It employs a multi-level precision sliding window attention mechanism, dynamically allocating precision levels such as fp16 / int8 / int4 based on the distance between the window and the current computation position, and the information entropy of the window content. A tensor parallel incremental attention engine enables cross-GPU / node incremental attention computation and asynchronous All-Reduce synchronization, improving computational efficiency. Semantic differential compression transmission technology is used, transmitting only the differential data based on similarity detection, significantly reducing cross-node transmission overhead. Combined with an extreme performance optimization and security control strategy, a balance between performance and security is achieved through hardware monitoring and dynamic power consumption / frequency adjustment. This invention can significantly reduce memory usage and transmission overhead in ultra-long context processing, improve computational efficiency and hardware utilization, enhance the understanding ability of large models for long texts, adapt to heterogeneous hardware environments, and is suitable for large model inference and training scenarios.

The application discloses a satellite image on-orbit change detection method and device, a storage medium and an electronic device, and relates to the technical field of information. The method comprises the following steps: searching for reference images matched with the to-be-detected images, and performing block processing on the to-be-detected images; extracting multi-scale semantic feature maps of each block image in the plurality of block images and the reference images; determining semantic probability distribution vectors corresponding to each block image and the reference images according to the multi-scale semantic feature maps of each block image and the reference images; determining a change mask corresponding to each block image based on the multi-scale semantic feature maps and the semantic probability distribution vectors corresponding to each block image and the reference images, and simultaneously evaluating the change detection uncertainty of each block image; and based on the change mask and the change detection uncertainty, transmitting part of data under a sparsification back transmission strategy to complete on-orbit change detection. The application can meet the real-time response requirement of satellite image change detection.

The invention discloses a method for realizing a centerless multicast conference, which comprises the following steps: S1, conference members adopt a multicast centerless mode to communicate and join in a centerless multicast conference group; s2, collecting voice data of conference members, performing energy detection, and sending the voice data when a speech is detected; and S3, after receiving the voice data, other conference members decode, cache, mix and play the voice data. According to the invention, the problem of user offline does not exist, and the service is available if the network is available: no conference server exists, so that the conference does not have a control center, and the fault of any member does not affect the group conference; the members can speak and receive voice at any time when the network is connected when entering the conference, the use of other users is not influenced when the network is not connected, and the users can listen to and speak in the conference immediately once the network communication is recovered.

The invention discloses a data real-time compression and transmission system and method based on a thinning algorithm, and relates to the technical field of data processing. The system comprises a data acquisition module, a data preprocessing module, a thinning algorithm module, a data compression module, a real-time display module and a data transmission module. The system uses 3D laser scanning to obtain three-dimensional point cloud data, and the data are preprocessed, denoised and partitioned. The thinning algorithm module recognizes and retains feature points, carries out thinning on non-feature points, and keeps data integrity and details. The compression module reduces the data volume, and the real-time display module dynamically loads the data blocks by using a graphics library, so that the display efficiency is improved. And finally, the data transmission module transmits the processed data to a client or a server through a network, and supports efficient data transmission and reception. The whole system integrates all links of data processing, and efficient and real-time data compression and transmission are achieved.

Embodiments of the present application provide a request processing method and device, electronic equipment and computer readable storage medium, and relate to the technical field of computers. The method comprises: determining a response end for responding to the display request according to the number of target electronic labels requested in the display request and / or the estimated response time of the central cloud server responding to the display request; and responding to the display request based on the response end. In this way, the display requirement of the electronic labels in the actual scene and the data processing capability of the central cloud server can be flexibly determined to respond to the display request by the central cloud server and / or the edge cloud server, thereby relieving the processing pressure of the central cloud server, reducing the network bandwidth occupation, and reducing the calculation delay.

The application discloses a kind of multi-modal event primitive extraction, coding and collaborative generation method and system for smart engineering construction scene, the system includes multi-modal data acquisition module, edge preprocessing module, event candidate extraction module, time-space alignment module, subject association module, event primitive coding module, multi-source fusion generation module and primitive output module. Through edge pre-processing, candidate event extraction, time alignment, space mapping, subject association, unified coding and multi-source fusion to video, environmental parameters, device status, positioning information, wearable terminal data and job log data and other multi-modal raw data, generate standardized event primitives, and output to risk identification, knowledge graph and early warning module. The present application can reduce the transmission pressure of raw data, improve the unified semantic expression ability of multi-source heterogeneous data, the fusion accuracy and the subsequent risk analysis efficiency, suitable for hoisting, high altitude, closed space and warehousing logistics and other smart engineering construction scenes.

The invention relates to the technical field of time frequency transmission, in particular to a PTP and SDH cross-network time service method and device, an SDH overhead channel is used for transmitting a PTP time signal, compared with an E1 service channel, the overhead channel is fixed in position in an SDH frame structure, is not influenced by branch pointer adjustment and is small in transmission delay jitter, meanwhile, PTP messages are not directly transmitted, and the transmission efficiency is improved. The PTP protocol is converted into a simplified time service protocol and then is transmitted by using an SDH overhead channel, so that the bandwidth occupation is reduced, the time synchronization frequency can be improved, the time service precision is improved, in addition, a time signal transmission path is more flattened, a multi-layer communication protocol nesting and a communication protocol conversion buffer area required by directly transmitting the PTP protocol are avoided, and the transmission efficiency is improved. According to the invention, the transmission delay jitter is reduced, the time service precision is improved, and the technical problem that the time service precision is seriously degraded when the PTP time signal crosses the SDH network for long-distance transmission in the existing scheme is solved.

The invention provides a federated learning-driven cross-mine intelligent risk identification method. Comprising the steps of mine multi-source operation image collection and mine risk event data set construction, mine intelligent risk perception and event recognition model construction, federated learning-based distributed mine risk model training and parameter optimization, model deployment and mine intelligent risk recognition. According to the method, through multi-layer feature coding and cross-scale collaborative modeling, the risk events of the key working area of the mine are accurately positioned and classified, and comprehensive identification of the risk events of the mine can be realized; federated learning and an adaptive sparse parameter communication method are adopted, cross-mine model training can be completed without concentrating original image data, and mine data privacy and distributed training efficiency are guaranteed; meanwhile, the comprehensive risk level of the whole image is generated in real time in combination with multi-event features and a global risk comprehensive evaluation mechanism, a direct quantitative decision basis is provided for mine safety management and instant early warning, and accuracy and operability are improved.

The invention relates to the field of acquisition and transmission data analysis, in particular to an acquisition and transmission data comparative analysis method based on parameter real-time monitoring. The technical problems that in the prior art, most of collected and transmitted data comparative analysis methods depend on manual data analysis, lack intelligent data processing and exception recognition capabilities, lack advanced algorithm support, incomplete data models and limitation of system architecture, and consequently complex data analysis and processing tasks are difficult to complete automatically are solved; according to the technical scheme, the acquisition and transmission data comparative analysis method based on parameter real-time monitoring comprises the steps that a parameter threshold value and a standard range are set, comparative analysis is conducted on real-time data and historical data, and abnormal points and trend changes in the data are recognized; according to the method, the data is sorted and sorted according to the time sequence, the real-time data and the historical data are compared and analyzed, the influence of interference factors on the data is eliminated through an algorithm model, and the identified abnormal data and trend are analyzed to realize automatic analysis of the data.

The invention discloses an unmanned aerial vehicle airborne ground multi-target online positioning method and system and a storage medium, an unmanned aerial vehicle is in communication connection with a vehicle end, the unmanned aerial vehicle carries out preliminary positioning on multiple ground targets and sends a preliminary geographic positioning result of the multiple ground targets to the vehicle end, and the method is executed by the vehicle end. Comprising the following steps: carrying out secondary correction on a ground multi-target preliminary geographic positioning result sent by an unmanned aerial vehicle, and projecting the corrected ground multi-target positioning result to a vehicle BEV space; target risk values of the multiple ground targets are calculated respectively, the multiple ground targets in the first data set are sorted according to the risk values from high to low, and a second data set is obtained; and the data set 2 is used for the vehicle end to carry out multi-target online positioning on the ground. According to the method, the preliminary positioning of the unmanned aerial vehicle is combined with the vehicle end correction, so that the multi-target high-precision online positioning of the road surface is realized, the vehicle end positioning calculation complexity is reduced, and the method has the characteristics of high positioning precision, strong real-time performance and the like.

This invention discloses a data transmission method, system, device, and storage medium for a unidirectional optical gate, relating to the field of data transmission technology. The method includes: a transmitting end acquiring the raw data to be transmitted, extracting the corresponding first integrity check value, performing deep binding and splicing, and generating a combined data stream; the transmitting end performing redundant encoding on the combined data stream and sending it to the receiving end through a physical unidirectional isolation channel; the receiving end using symmetric decoding logic to perform packet loss self-healing calculation, restoring the combined data stream, and disassembling the restored combined data stream to separate the raw data and the corresponding first integrity check value; the receiving end performing a second integrity check on the disassembled raw data locally, comparing the check result with the disassembled first integrity check value, and outputting the raw data when the checks match. The method described in this invention improves data transmission integrity and reliability, enhances system flexibility and ease of use, and optimizes the real-time performance of data transmission.

This invention discloses a multi-level on-board processing and transmission architecture for space debris cataloging, belonging to the fields of satellite overall design and space debris monitoring technology. The architecture includes: a storage and data exchange unit, a traditional field-programmable gate array (FPGA) processing module, an artificial intelligence (AI) computing processing module, an integrated electronic computer, a high-speed data transmission terminal, and a BeiDou short message terminal. The storage and data exchange unit serves as the information hub for business data interaction; the FPGA processing module performs high-reliability backup processing to generate target detection data; the AI ​​computing processing module performs high-performance intelligent processing to generate astronomical positioning data and anomaly monitoring data; the integrated electronic computer schedules the high-speed data transmission terminal, BeiDou short message terminal, or telemetry and control link to complete data downlink according to the priority of the anomaly monitoring data. This invention achieves on-board simplification and low-latency transmission of massive amounts of data through heterogeneous processing and multi-level data backhaul, meeting the high-timeliness requirements of space debris cataloging.

The present application relates to the technical field of address translation request scheduling, and particularly relates to an address translation request aggregation scheduling method and system. The present application maintains a pending request filter table, each table entry corresponding to a 4KB address range, recording address translation requests that have been sent but have not received responses. When a new request is received, the 4KB base address to which the virtual address belongs is calculated and the filter table is queried: if there is a corresponding table entry, the request is added to the associated waiting queue, and the new request is not sent; if there is no corresponding table entry, the new request is sent and a new table entry is created. Thus, during the waiting response period, subsequent requests in the same address range are aggregated, the number of repeated requests is reduced from N times to 1 time, the root complex load is significantly reduced, redundant transactions are reduced, and the hardware overhead is extremely low.

A video playback method, device and system. The method includes: a VR device receives a VR video stream of a first perspective with a first resolution sent by a server and a full-perspective VR video stream of a second resolution continuously sent by the server, wherein the first resolution is greater than the second resolution; the first perspective VR video stream is used to play the picture of the first perspective; when it is determined that the perspective has changed, the first or both of the full-perspective VR video stream and the first perspective VR video stream are used to play the picture of the current perspective. The solution provided by the embodiment of the present application overcomes the problems of large bandwidth occupation and high latency of network VR services, and can reduce the storage overhead of the server and avoid traffic bursts.