120 results about "Star network" patented technology

The invention discloses a lightweight mobile phone screen defect detection method based on improved YOLOv10, and the method comprises the following steps: S1, collecting mobile phone screen image data, carrying out the enhancement processing of a data set, marking the data set, and constructing a mobile phone screen defect training data set; s2, adopting a Starnet module, a C2f-ODConv module and a DetectLSCD module to improve a YOLOv10 network model structure, and constructing a mobile phone screen defect detection model of a YOLO-STAR network structure; s3, using the mobile phone screen defect training data set constructed in the S1 as input, and training the improved YOLOv10 network structure model; and S4, deploying the trained YOLOv10 network structure model to a mobile device, and carrying out real-time detection on a mobile phone screen image. According to the invention, by improving the architecture of the YOLOv10 model, the complexity and the calculation amount of the model are reduced, the training data set and the detection process are optimized, and the deployment efficiency and the real-time detection performance of the model on equipment are improved.

The invention discloses an insulator defect detection method and system based on RT-DETR improvement. The method comprises the following steps: collecting a disclosed insulator defect data set and obtaining an image as a data set through a video interception frame; the method comprises the following steps: carrying out improvement based on a target detection model RT-DETR, changing a backbone network into a StarNet star network to carry out feature extraction, carrying out further improvement, removing a P5 detection head, adding a P2 detection head, and improving a star convolution module in the StarNet star network; a multi-head attention mechanism is changed into a partial self-attention mechanism PSA, and further improvement is carried out to obtain an S-PSA attention mechanism module; a RepC3 convolution module is improved into a RepMob convolution module; combining the loss function by using Inner-IoU and CIoU to obtain an Inner-CIoU loss function, and carrying out loss calculation; and training the improved RT-DETR, verifying the improved RT-DETR to obtain an optimal weight, and loading the weight into the improved RT-DETR to carry out insulator defect detection. According to the invention, the defect detection is carried out on the insulator by using the improvement based on RT-DETR, so that the detection is more accurate and efficient.

The invention discloses a wireless cooperative control system and method for a multi-node brushless motor, and the system employs a layered architecture design, and comprises a main control layer, a drive layer, a communication layer, and a software layer. The master control layer is composed of a master node and serves as a task instruction scheduling center of the system, the driving layer is composed of N slave nodes, and N is larger than or equal to 3; the driving layer drives the brushless motor by executing an FOC vector control algorithm and monitors the state of the brushless motor; the communication layer constructs a star network topology based on an ESP-NOW protocol, a direct communication link is established between a master node and N slave nodes, and multi-node cooperative control is realized by adopting a dynamic channel switching mechanism and a timestamp alignment mechanism; and the software layer establishes communication with the main control layer through a USART serial port, receives the state data of the brushless motor of each slave node concentrated by the main control layer, and sends an instruction to the main control layer at the same time. Not only are wireless multi-node synchronization and anti-interference problems solved, but also high-precision motor control is realized.

The invention relates to the technical field of wireless communication, in particular to a hospital district wireless network system supporting an internet of things protocol, which is characterized in that a multi-protocol dynamic scheduling engine intelligently selects a protocol stack by utilizing a lightweight reinforcement learning model through a three-layer dynamic decision-making architecture; the time-sensitive hybrid topology network adopts dual-mode topology fusion to provide a star network and low-delay time slot distribution for emergency equipment; the cross-protocol security aggregation gateway realizes cross-protocol data encapsulation and field-level mapping based on a unified data head of the medical Internet of Things; the intelligent spectrum anti-interference unit monitors and analyzes a hospital frequency band electromagnetic environment in real time to obtain a spectrum sensing information result, and feeds back the spectrum sensing information result to the multi-protocol dynamic scheduling engine and time-sensitive hybrid topology network. According to the invention, the multi-protocol dynamic scheduling engine is combined with the time-sensitive hybrid topology network, so that the communication problems of equipment types and complex environments can be solved.

The invention discloses a quantum security multi-party computing method of a hybrid topological structure, which is applied to a quantum security multi-party computing system, the quantum security multi-party computing system comprises m sub-networks of star network topological structures, and each sub-network corresponds to a quantum center and a plurality of users; quantum communication is completed between every two quantum centers based on a ring network topology structure; the quantum security multi-party computing method comprises the following steps: each quantum center completes modular addition operation of a corresponding user secret value in a corresponding subnet to obtain a modular addition result of all user secret values in the corresponding subnet, and the modular addition result is recorded as a subnet secret value modular addition result; and on the basis of a subnet secret value modular addition result, each quantum center continuously participates in the secure multi-party computing protocol among the quantum centers to obtain the sum of all user secret values in the quantum secure multi-party computing system. According to the method, on the premise that private secrets of users are not leaked, safer and more efficient quantum security multi-party calculation can be realized among the users with limited quantum capabilities.

The invention provides an equipment networking method and system based on a distributed soft bus, and the method comprises the steps: starting a distributed soft bus module and carrying out the discovery of adjacent equipment, and obtaining a trusted equipment set after identity verification; a point-to-point communication link between the devices is established according to the trusted device set, and a distributed mesh interconnection topology is formed; performing routing calculation and data forwarding on data transmission in the mesh topology to realize end-to-end communication; meanwhile, dynamic changes of equipment in the network are monitored, topology updating is carried out, and the network stability is maintained. The scheme does not depend on the dependence of the traditional star network on the central node, realizes direct interconnection between devices through the distributed soft bus technology, improves the fault-tolerant capability and transmission efficiency of the network, and solves the problems of poor network reliability and limited expansibility in the prior art.

The invention relates to the technical field of image processing, and discloses an improved StarNet-YOLOv13-based unmanned aerial vehicle field tobacco virus disease lightweight detection method, which comprises the following steps: acquiring a to-be-detected image of a field tobacco plant, and then inputting the to-be-detected image into a trained tobacco virus detection model to obtain a detection result of the tobacco plant; wherein the tobacco virus detection model is obtained by replacing a backbone network with a star network, replacing a C3K2 module in a neck network with a feature fusion module and replacing a detection head with a DetectMBConv detection head in a YOLOv13 model; the feature fusion module is constructed by introducing a partial convolution mechanism into the C3K2 module; thus, through the star network (StarNet), the feature fusion module (DSC3k2PConv module) and the DetectMBConv detection head, the small target scab detection capability and the complex scene robustness can be improved while the parameter and calculation overhead can be reduced, so that the tobacco virus detection model can meet the deployment requirement of the computing power limited equipment side, such as an unmanned aerial vehicle, and the detection precision is ensured at the same time.

The invention discloses a Bluetooth-based ad hoc network voice communication method, which comprises the following steps of: firstly establishing a link between communication devices by using a Bluetooth ACL asynchronous link technology, and then establishing a link between the communication devices by using a Bluetooth eSCO or CIS synchronous isochronous link technology, after the connection, forming a linear or star network topology structure between the communication devices serving as connection points, and in the networking, connecting the communication devices by using the Bluetooth ACL asynchronous link technology and the Bluetooth eSCO or CIS synchronous isochronous link technology. And the communication devices transmit data to each other and filter or forward data sent by the neighbor devices to the far-end device, so that mutual voice communication among the group members is achieved. By adopting the technical scheme of the invention, the wireless ad hoc network communication in a local area range can be met, and the wireless ad hoc network can be interconnected and intercommunicated with other Bluetooth products in a relatively universal manner.

The invention relates to the technical field of communication, in particular to a broadband access and data transmission device supporting a star network or a MESH network, and the device comprises a network node configuration module which is used for configuring a star network or MESH network mode according to an access scene, inputting a network number, a connection authority and a role attribute of a node, calling a distance value between the nodes, calling a corresponding connection rule, and transmitting the connection rule to the network node configuration module; setting threshold values are compared, the connection relation between the nodes is determined, and a node connection table is generated. According to the invention, by flexibly configuring different network structures, efficient operation in diversified network environments can be realized, and especially under the condition of dense nodes or complex network topology, the serial numbers, permissions and roles of the nodes can be dynamically adjusted, and the accuracy and stability of network connection can be ensured. By accurately inputting the node information and optimizing the node connection relation according to the distance and the rule, the expansion capability and the adaptability of the network are effectively improved.

The invention relates to a low-orbit star network-oriented random target tracking estimation method, which comprises the following steps of: establishing a dynamic filter AUKF of multiple targets in a geocentric inertial coordinate system by taking position, speed and thrust acceleration vectors of the targets j (j = 1, 2,..., Nt) in the geocentric inertial coordinate system as state vectors; establishing a maneuvering target space-based observation model based on the measurement equation of the target under the geocentric inertial coordinate system; based on the calculation result of the maneuvering target space-based observation model at the current moment, constructing a maneuvering starting and ending detector, when it is judged that the target starts maneuvering, using a maneuvering dynamics filter AUKF to track and estimate the position, speed and maneuvering acceleration information of the target, and when it is judged that the maneuvering of the target is ended, using the maneuvering dynamics filter AUKF to track and estimate the maneuvering speed of the target. And a standard Kalman filter UKF is switched to carry out state updating on the position and the speed of the target. According to the method, the relay observation and whole-course tracking performance estimation problems of the low-orbit star network on the uncertain maneuvering target are solved, and the method can be used for verification and evaluation of a large-scale star network system detection simulation test.

A coordinated control system designed with a small form factor that can be located within a robot or an automated inline manufacturing system, comprising: an on board advanced distributed control hardware configured with artificial intelligence enabled processors incorporated within a System on Module (SOM) board, each SOM comprising four processors or nodes; interfaced with a cameras, motors, general purpose I / O through I2C expander, audio interface, Internet, wireless communication such as WiFi and Bluetooth to send and receive commands and any other data; electrically connected to a plurality of sensors for perceiving the environment and collecting data from infrared, tactile, proximity and other types of sensors; an external host flashing computer dedicated for uploading / downloading firmware, cloning and configuration setup; a non-volatile memory for storing control algorithms, configuration files and other essential data; a HDMI based user interface for robot operation, training and setup; USB C or Ethernet based internal star network for high-speed communication bypassing the standard PCI bus interface bus; an Ethernet switch board enabling multiple boards to access the Ethernet for both internal communication within the star network as well as external Internet access.

The invention relates to the technical field of power distribution networks, in particular to a park micro-grid cluster multi-source collaborative hierarchical power control system and control method, and the control system comprises a micro-grid cluster energy management system which is used for managing a micro-grid cluster; the micro-grid cluster energy management system comprises a plurality of sub-micro-grid energy management systems for managing sub-micro-grids; the micro-grid cluster is of a star-shaped network topological structure comprising a plurality of levels and comprises a plurality of sub-micro-grids, and each sub-micro-grid comprises a photovoltaic cluster, a fan cluster, an energy storage cluster and a flexible load cluster; the sub-microgrid energy management system obtains operation data of each sub-microgrid; the micro-grid cluster energy management system formulates a partition power distribution strategy based on the operation data of each sub-micro-grid; and the sub-microgrid energy management system carries out power distribution on each level of the corresponding sub-microgrid based on a partition power distribution strategy. Dependence on the power grid is reduced, and safe and stable operation of the frequency voltage of the park microgrid cluster is guaranteed.

The invention discloses a foreign language word card, which is characterized in that a picture is arranged on the front side of the card and expresses characteristic elements or related attributes of a word; the core method of five-star networked words (FSNW: Five Star Networked Word) introduced in the invention is arranged on the back side of the card. The FSNW comprises the word itself and five most meaningful contexts thereof, namely definition, synonym, antonym, hyponym and affix. The display form of the networked words included in the FSNW can be changed according to the selected difficulty level, for example, an English word beader, and when the difficulty level is low, the display form is beader; and when the difficulty level is middle, the display form is b.... r, namely, each middle letter is replaced by a point (.). The FSNW card can be used like a word guessing game, that is, after the card is read, a learner needs to spell the networked word; the FSNW helps a learner to get rid of learning words in an isolated context, helps the learner to fully understand the words, and can easily retrieve the words when the words are forgotten.

The invention discloses a wireless sensor data acquisition system for an electromagnetic shielding scene. The wireless sensor data acquisition system comprises an industrial personal computer, a wireless instruction transceiving unit, a wireless signal relay unit and a data acquisition unit. An instruction link and a data link are established in a ZigBee mode and a Wi-Fi mode respectively, operation of the data link of the system is controlled by the instruction link, and therefore system nodes have high data transmission bandwidth under the condition that low power consumption is kept; a system instruction link is constructed by connecting two stages of ZigBee star networks in series through hardware, stable communication of acquisition nodes in an object which is relatively closed and has an electromagnetic shielding effect can be realized, and the application scene of a wireless sensor network is widened. The control operation of the whole system is completed through matched upper computer software, and the software mainly provides the state of each unit of the current system for an operator and provides automatic processing operation, so that the operation is simplified.

The invention discloses an unmanned aerial vehicle unmanned recovery operation method and system based on an automatic recovery vehicle, and belongs to the technical field of unmanned aerial vehicle automation. The system comprises a nest box body serving as a hub, a signal tower connected with the nest box body, an external take-off and landing site, an automatic recovery vehicle and a cloud management platform. The core of the method is that a cloud platform issues an instruction, and a recovery vehicle takes out an unmanned aerial vehicle from a machine nest and transfers the unmanned aerial vehicle to a take-off and landing site for flying; and after the unmanned aerial vehicle returns and lands, the recovery vehicle actively goes forward to capture and carry back to the nest, and charging and inspection are completed. The mobile recovery vehicle with a self-adaptive interface conversion module is introduced as an execution medium, a high-precision docking demand is transferred from the air to the ground, and a star network architecture with a nest as a core is adopted, so that the recovery risk of the unmanned aerial vehicle and the dependence of the system on a specific model are remarkably reduced, and the recovery efficiency is improved. And the reliability, the compatibility and the full-automatic operation capability in a complex environment are improved.

A method for avoiding high-frequency communication conflicts in LoRa multi-node communication includes the following steps: Step 1: Constructing a star network topology for LoRa wireless nodes; Step 2: Setting up a LoRa network for data terminal nodes; Step 3: Setting up relay nodes and agreeing on a data upload time period; Step 4: Dividing the uplink transmission period of terminal nodes into multiple time slots. Each node is assigned a corresponding transmission time slot; Step 5: Assigning priorities to each terminal node, creating a table of given node sequence lists and initial state of transmitted nodes, preprocessing the data, and uploading the data from the terminal node to the relay node. This invention achieves non-interference in uplink data transmission among terminal nodes by dividing terminal node priorities into two bits and uplink data transmission time slots. Terminal nodes listen to priority data packets and create tables, ensuring that uplink data transmission among terminal nodes does not interfere with each other, urgent data is uploaded first, and the two priority data bits guarantee the integrity of node data transmission.

A coordinated control system designed to have a small form factor can be located within a robotic or automated online manufacturing system, the coordinated control system comprising: on-board advanced distributed control hardware configured with artificial intelligence enabled processors incorporated within a System-on-Module (SOM) board, each SOM comprising four processors or nodes; wireless communication with cameras, motors, general purpose I / O via I2C extenders, audio interfaces, the Internet, such as Wifi and Bluetooth connections, for transmitting and receiving commands and any other data; a sensor electrically connected to a plurality of sensors for perceiving an environment and collecting data from infrared, haptic, proximity and other types of sensors; the external host flash memory computer is specially used for uploading / downloading firmware, cloning and configuring; the nonvolatile memory is used for storing a control algorithm, a configuration file and other necessary data; the HDMI-based user interface is used for robot operation, training and setting; the internal star network based on USB C or Ethernet is used for bypassing the standard PCI bus interface BUS to carry out high-speed communication; an Ethernet switch enables a plurality of boards to access the Ethernet for internal communication within the star network as well as external Internet access.

The invention provides an information transmission method and device, a terminal and network equipment, and relates to the technical field of communication. The information transmission method is executed by network equipment and comprises the following steps: judging whether satellite storage and forwarding equipment can meet storage and forwarding requirements of a terminal according to first information, the first information comprises at least one of the following items: the storage and forwarding capacity requested by the terminal, indication information about whether the terminal supports segmented transmission or not, and the minimum value of the storage and forwarding capacity which can be accepted and distributed by the terminal; the satellite store-and-forward equipment can distribute the maximum store-and-forward capacity for providing store-and-forward service for the terminal; according to the judgment result, second information is sent to the terminal, and the second information is used for indicating the time point when the terminal sends the store-and-forward request next time. According to the scheme, a negotiation mechanism that the satellite network element distributes all the storage and forwarding capacity requested by the satellite network element to the terminal is provided, and the storage and forwarding request of the terminal can be met to the greatest extent as far as possible.

The invention relates to the technical field of palm vein recognition, in particular to an efficient palm vein representation attack detection method and system based on a lightweight neural network. The method comprises the following steps: acquiring an image to be identified; inputting a to-be-recognized image into the Xception entry stream, and performing preliminary feature extraction on the to-be-recognized image to obtain an initial feature; inputting the initial features into Block modules of a star network, and further mining high-order nonlinear features in the image by each Block module through convolution and star operation; and performing classification identification on the high-order nonlinear features to obtain classification result data. Real and forged palm vein images are distinguished by capturing vein texture features in the images, a single-channel near-infrared palm vein image is input, and a binary result of true and false classification is output.

The embodiment of the invention discloses a method for establishing a publishing and subscribing relationship based on a reliable unicast protocol. The method comprises the following steps: starting a node management component, a node management backup component and each node communication component to complete memory and network monitoring; a star network topology structure is established by using each node communication component and the node management component, and registration operation of each node communication component and information sharing of the node management component and the node management backup component are completed; establishing a publishing and subscribing relationship when a publisher in each node communication component operates preferentially; establishing a publishing and subscribing relationship when subscribers in each node communication component operate preferentially; releasing the interest by utilizing a publisher in each node communication component; the subscribers in the node communication components are utilized to cancel interest release; exiting the star network topology structure by using each description node communication component; and closing the node management component, the node management backup component and the node communication component, and stopping memory and network monitoring.

The present invention belongs to the field of marine observation communication, and specifically discloses a marine multi-channel intelligent communication device and method based on a handshake mechanism. The marine multi-channel intelligent communication device of the present invention is a high-security communication device based on a handshake mechanism and suitable for poor marine observation network environments, designed between the server and the receiving end through a hardware expansion communication interface on the basis of the traditional multi-channel communication technology. The marine multi-channel intelligent communication method of the present invention aims to achieve high-speed and high-security data transmission. For the real-time observation data obtained by the existing commonly used hydrological and meteorological collectors in the marine observation system, it implements the comparison and immediate recall and reissue of delayed, missed, lost, and missing data based on the handshake mechanism, and uniformly deploys multiple communication means such as digital radio stations, dedicated line networks, VSAT, 4G, Beidou, and low-orbit satellite networks (Starnet) to achieve redundant communication guarantees.

The utility model discloses a DCS (Distributed Control System) network system of a solvent recovery device. The DCS network system comprises an upper computer, the upper computer is connected with an A-zone control room, the A-zone control room is connected with a B-zone control room, an A-zone control chamber is arranged outside the A-zone control room, and a B-zone control chamber is arranged outside the B-zone control room; a first switch is arranged in the A-area control room, is connected with an upper computer, and is connected with a control cabinet, a first frequency conversion cabinet and a power supply cabinet; and a second switch is arranged in the B-area control room, is connected with the first switch, and is connected with a second frequency conversion cabinet. According to the application, the switch is arranged, so that the communication problem when the distance between the solvent recovery areas is relatively long is solved; the star network is adopted, so that faults in a control room with centralized nodes are easy to remove, and the ring network is adopted, so that communication of other nodes is not affected when any node in dispersed nodes outside the control room fails, and investment is low.

The present application discloses a performance test system and method for an intersatellite network optical terminal, which relates to the field of satellite optical communication technology. The test system is applied to the aiming, capturing and tracking test of the network optical terminal to be tested. The test system includes: an optical path switching module, which is used to switch the optical path according to the optical path switching instruction; an intersatellite beam dynamic simulation module is used to receive the test beam emitted by the network optical terminal to be tested through the switched optical path, and simulate and monitor the intersatellite transmission of the test beam; the ground inspection optical terminal is used to generate an auxiliary beam according to the test needs, and send it to the optical terminal through the intersatellite beam dynamic simulation module; the host computer is used to test the network optical terminal to be tested through the intersatellite beam dynamic simulation module, and generate the test results of the network optical terminal to be tested based on the optical axis calibration results, optical path trajectory, and time data. The above test system is used to realize the performance test of the network optical terminal in different scenarios, reduce the difficulty of optical path installation and adjustment of the test system, and improve the test efficiency.

The invention provides an annealing kiln temperature control system and method and a storage medium, relates to the field of annealing kiln control, and solves the technical problem that stable and accurate control of the temperature of an annealing kiln is difficult to realize in the prior art. The system specifically comprises a sensor detection unit which comprises infrared thermometers located in each transverse sub-zone of the annealing kiln, a flow thermometer located in a transverse air duct of the annealing kiln, and an extended compensation wire thermocouple; the execution control unit comprises an electric butterfly valve arranged in a cooling air duct of the annealing kiln and a fan used for adjusting the air temperature; the star network topology structure comprises an Ethernet switch and an upper computer, the Ethernet switch is used for establishing communication connection with the upper computer by taking equipment of the annealing kiln as slave station equipment; the upper computer is used for acquiring annealing kiln control data; adjusting the temperature of the annealing kiln; and an annealing kiln comprehensive database is established, and operation data of the annealing kilns are recorded, summarized and uploaded in real time. The method is used for controlling the temperature of the annealing kiln.

The invention belongs to the technical field of automotive electronics, and relates to a high-anti-interference controller PCB layout structure for an electric power steering system, a PCB is a multilayer board, and the PCB layout structure is characterized in that a power ground loop of a high-frequency switching circuit is directly connected to a grounding end of a common mode inductor to realize single-point discharge of high-frequency noise; the analog signal circuit is far away from the high-frequency digital signal circuit in layout; the MCU power supply decoupling capacitors are arranged around the MCU in a distributed star network; the PCB layout structure is provided with a power ground plane and a signal ground plane which are completely isolated, and the power ground plane and the signal ground plane are connected only through a single common mode inductor; in the PCB layout structure, a driving signal, a feedback signal, a sampling signal and a power signal of the inverter bridge part are respectively arranged in different PCB layers; the current sampling precision and the system reliability are remarkably improved, the steering hand feeling is improved, the electromagnetic compatibility is excellent, and the engineering application value is very high.

The invention provides a multi-channel star network chain channel simulator testing device and method, and relates to the technical field of satellite communication. Multiple test tasks of a multi-channel star network chain channel simulator can be used as target test tasks one by one according to a test time sequence through a display control module; calling at least one target communication interface of the target test task from a plurality of communication interfaces through a signal transceiving module; measuring a target measurement parameter of the target test task based on the at least one target communication interface through a measurement module; receiving a measurement value of the target measurement parameter through a calculation module, and calculating a test result of the target test task according to the measurement value; storing a test result of the target test task through a storage module; and after the multiple test tasks are tested, obtaining and outputting test results of the test tasks from the storage module through the display control module. According to the invention, the measurement result accuracy and the measurement efficiency of the multichannel star network chain channel simulator can be improved.

The application belongs to the technical field of launch vehicle electronic information system, and particularly relates to a launch vehicle communication network system based on Powerlink. The system comprises a two-stage star network of a launch vehicle, an inter-stage star network, a first-stage tail star network and a main bus adopting a Powerlink bus; a plurality of terminal slave nodes in the two-stage section access a two-stage network repeater through a star topology communication link, a master node and the two-stage network repeater are connected in series to the main bus through a two-stage interface; a plurality of terminal slave nodes in the inter-stage section access an inter-stage network repeater through a star topology communication link, a redundant master node and the inter-stage network repeater are connected in series to the main bus through an inter-stage interface; a plurality of terminal slave nodes in the first-stage tail section access a first-stage tail network repeater through a star topology communication link, and are connected to the main bus through a first-stage tail interface. The system has high real-time performance and high reliability.

The invention provides a clustering networking method and device for a pure directional self-organizing network, and relates to the technical field of networks, and the method comprises the steps: initializing nodes, classifying the nodes into each cluster through a clustering and cluster head consensus mechanism, and selecting a cluster head node of each cluster; constructing a star network by taking each cluster head node as a central node through a rapid networking mechanism combining wide and narrow beams; and in response to the situation that the cluster head node cannot continuously serve as the cluster head, reselecting the cluster head node and reconstructing the network. By determining the cluster head node and constructing the star network, the robustness, survivability and communication efficiency of the network are improved. By adopting the directional antenna, the network capacity and the communication distance are improved, and certain robustness and survivability are ensured while the concealment, the safety and the anti-interference capability are enhanced.

This disclosure presents an intelligent star map recognition method, apparatus, device, and medium based on multi-network mapping. One specific implementation of the method includes: extracting features from a star image using a pre-constructed star feature extraction network and an extracted set of star information to generate feature information for each node; constructing a graph network from the star information set based on star-sensor camera parameters to generate a star network structure; performing star map recognition on the star network structure using a pre-constructed star map recognition network model, the feature information for each node, and prior attitude information to generate a star recognition result; caching the star recognition result; and outputting the star recognition result to the spacecraft's attitude control system in real time. This implementation can ensure a high star recognition rate while significantly reducing the storage complexity of the feature library and the computational complexity of the algorithm, and exhibits stronger robustness to changes in the number of stars and noise in star magnitude information.

The present invention discloses an insulator defect detection method and system improved based on RT-DETR. A publicly available insulator defect data set is collected, and images obtained by intercepting frames from a video are used as the data set. Based on the object detection model RT-DETR, improvements are made. The backbone network is changed to the StarNet star network for feature extraction, and further improvements are made. The P5 detection head is removed, a P2 detection head is added, and the star convolution module therein is improved. The multi-head attention mechanism is changed to the partial self-attention mechanism PSA, and further improvements are made to obtain the S-PSA attention mechanism module. The RepC3 convolution module is improved to the RepMob convolution module. The loss function is combined with Inner-IoU and CIoU to obtain the Inner-CIoU loss function for loss calculation. The improved RT-DETR is trained and verified to obtain the optimal weights, and these weights are loaded into the improved RT-DETR for insulator defect detection. The present invention uses the improvement based on RT-DETR to detect insulator defects, making the detection more accurate and efficient.