16results about How to "Improve expression accuracy" patented technology

The invention provides a radiation source individual identification method, device, equipment, medium and product, and the method comprises the steps: obtaining a physical signal feature and a deep learning signal feature corresponding to an intermediate frequency signal of a to-be-identified radiation source individual, the physical signal feature being obtained by extracting the physical feature of the intermediate frequency signal, and the deep learning signal feature being obtained by extracting the physical feature of the intermediate frequency signal; the deep learning signal features are obtained by performing deep learning-based feature extraction on the intermediate frequency signals; respectively performing convolution processing on the physical signal features and the deep learning signal features to obtain convolution physical signal features and convolution deep learning signal features; and fusing the convolutional physical signal feature and the convolutional deep learning signal feature to obtain a fused signal feature, and obtaining an identification result of the radiation source individual based on the fused signal feature. According to the invention, the radiation source individual identification accuracy can be improved.

The invention discloses a day-ahead optimization method and system for a PET-containing AC / DC hybrid power distribution network based on a cloud model, and belongs to the field of day-ahead optimization scheduling of a power system. The method comprises the following steps: calculating three characteristic quantities of expectation, entropy and hyper-entropy of a cloud model to obtain the expectation, entropy and hyper-entropy of the cloud model; selecting a kernel function, and calculating probability density distribution of wind power and photovoltaic power prediction errors by adopting nonparametric kernel density estimation; setting confidence, finding the shortest interval to enable the probability that the prediction error value falls into the interval to be equal to the confidence, and obtaining upper and lower limits of wind power and photovoltaic power; constructing a day-ahead optimal scheduling model of the AC / DC hybrid power distribution network containing the PET considering the uncertainty of wind and light output prediction; and solving the model through a commercial solver to obtain a day-ahead optimal scheduling scheme of the AC / DC hybrid power distribution network containing the PET. According to the method, the uncertainty of wind and light output is considered, the power distribution network is allowed to accept more wind and light output, the total cost is reduced, and the power coordinating range of each port is expanded.

The application discloses a textile yarn intelligent production control system and method based on digital twinning, which comprises the following steps: S1, obtaining the running data of a spinning frame; S2, establishing a virtual running body of the spinning frame to form a virtual-real working condition state; S3, inputting the virtual-real working condition state into an improved Reformer model, introducing a yarn state neighborhood resonance regularization mechanism in an LSH attention layer, executing neighborhood rearrangement and cross-bar neighborhood splicing, and generating continuous state quantities; S4, adopting a Hilbert-Huang transformation algorithm on the continuous state quantities, introducing a critical precursor stationary phase separation mechanism, and forming multi-scale working condition components; S5, forming a spinning working condition trajectory; S6, executing linkage matching to generate linkage control quantities; and S7, applying the linkage control quantities to the spinning frame to update the virtual-real working condition state. The application can realize continuous modeling and multi-parameter collaborative control of the spinning working condition, and improve the stability of the running state of the spinning frame and the production control precision.

The application belongs to the technical field of automobiles, and particularly relates to a four-wheel side stiffness estimation method for a ground vehicle considering off-road environment disturbance. First, a vehicle dynamics model and a kinematics model considering coupled slope are established through coordinate transformation and dynamics analysis, so as to ensure the accuracy of the unmanned ground vehicle model in a complex off-road environment. Second, an unmanned ground vehicle path tracking system is constructed and state space equation discretization is carried out, so that the system can be used for four-wheel independent side stiffness estimation. Finally, a four-wheel side stiffness estimation method based on adaptive forgetting least squares is provided, which can ensure accurate and reliable estimation of tire side stiffness in an off-road environment.

The embodiment of the present application provides a method, device and equipment for generating a terrain grid of a wind farm. The method includes: determining a two-dimensional calculation area for generating a grid according to a terrain geometry model; dividing the two-dimensional calculation area into layers of grids according to an N-ary tree algorithm, where N is a positive integer greater than 1; in the process of dividing each layer of grids, after dividing N grid units of the current layer, when it is determined that the terrain surface curvature change value between a grid unit of the current layer and an adjacent grid unit is greater than a preset threshold, continue to divide the grid unit of the current layer into the next layer of grids until the terrain surface curvature change value is less than or equal to the preset threshold, and obtain a plane grid. The embodiment of the present application is used to solve the technical problem that the existing grid generation speed is very slow.

The invention discloses a map information interaction method and device, equipment, a medium and a product, and relates to the technical field of computers. The method comprises the following steps: displaying a virtual map corresponding to a virtual scene; in response to a received first drawing operation on the virtual map, displaying a first graph drawing result, the first graph drawing result being used for expressing an action strategy of at least one of the first virtual object and the second virtual object in the virtual scene through the virtual map; and sending the first graph drawing result to the second account, wherein the first graph drawing result is used for being displayed on a terminal interface logged in by the second account. According to the method, the accuracy of action strategy information transmission is improved, and the accuracy of communication between users is improved.

The application discloses a fault identification method, device and equipment of a flexible alternating current (AC) feeder, and a medium, and belongs to the field of power system protection. The method is as follows: equivalent circuits of a flexible direct current (DC) system in different modes are topologically connected according to boundary conditions of single-phase ground faults to construct a multi-mode equivalent network model; a fault excitation is applied at a preset fault point of the multi-mode equivalent network model, a theoretical current response of a preset protection installation under the fault excitation is calculated, and theoretical transient characteristic values are extracted from the theoretical current response; electrical signals at the preset protection installation are collected in real time, when it is detected that the electrical signals meet preset starting conditions, an actual current response of the preset protection installation is acquired, and actual transient characteristic values are extracted from the actual current response; the actual transient characteristic values are compared with the theoretical transient characteristic values, and the fault of the flexible AC feeder is identified according to a comparison result. The application can improve the accuracy of fault identification of the flexible AC feeder.

The embodiment of the invention provides a state publishing method and device, equipment and a storage medium, and relates to the technical field of computers, and the method comprises the steps: responding to a scene viewing operation triggered by a target object, and presenting a public scene page, wherein the virtual character information at least comprises respective published virtual characters and state description information of each contact object having an association relationship with the target object; presenting an editing page in response to the triggered switching operation, wherein the editing page comprises a plurality of candidate virtual roles associated with the target object; in response to a triggered editing related operation, presenting a target virtual character selected from the target pair in the editing page, and presenting target state description information set by the target object for the target virtual character; and in response to the triggered release operation, presenting the target virtual character and the target state description information in the public scene page. The method is used for improving information reading efficiency and information expression accuracy.

The invention discloses a full name information generation method and a related device, and relates to the technical field of natural language processing. According to the method, the pre-trained sequence generation model is utilized, and the intermediate vector after the vector is embedded based on the corresponding weight parameter fusion image is encoded layer by layer through the multi-layer encoder blocks in the encoder constructed based on the self-attention mechanism; decoding an intermediate vector after the vector is embedded on the basis of the weight parameter fusion graph layer by layer through a plurality of layers of decoder blocks in a decoder constructed on the basis of a self-attention mechanism to obtain a full-name embedding vector sequence formed by full-name embedding vectors of each target abbreviation sub-word; and generating full-name information of the target abbreviation information according to the full-name embedding vector sequence. Through the sequence generation model and based on graph embedded reinforcement learning, the purpose of improving the accuracy of the full name information of the target short name information is achieved.

The application discloses an indoor positioning method and system fusing WiFi fingerprints and space information, the method realizes online positioning through offline construction of a fingerprint map. Wherein, WISP features fusing WiFi features and space features are proposed, the space features are classified and coded according to whether reference points are in straight corridors or corners. In addition, the application also constructs a bag-of-words tree, that is, the bag-of-words tree is used to represent the correlation between reference points, so that in the online positioning stage, firstly, the WISP features of the current position and the bag-of-words tree are used to calculate the similarity between the WISP features of the current position and the WISP features under each cluster; then the product of the similarity and the corresponding TF-IDF weight value is calculated, finally, the cluster with the maximum product value is selected, and the position corresponding to the WISP features under the cluster is used to deduce the final positioning result, effectively reducing the influence degree of the perception confusion of WiFi signals.

The invention relates to the technical field of computer network security, provides a software backdoor detection method and device based on API (Application Program Interface) call sequence feature mining, and solves the technical problems of low detection accuracy and insufficient generalization ability of a backdoor program with high concealment and multi-modal features. The method comprises the steps of obtaining a target API calling sequence and network flow data in a target software running process, and obtaining an arrival time interval sequence; based on a preset API semantic mapping table, converting the target API calling sequence into a feature vector sequence, and converting the arrival time interval sequence into a discrete time interval sequence; constructing a heterogeneous behavior graph by mapping each feature vector in the feature vector sequence into a behavior node; inputting a pre-trained graph isomorphic network model to obtain a node embedding vector; and the node embedding vectors are combined to determine that the backdoor behavior exists in the target software, so that accurate detection of the software backdoor behavior with strong concealment and high camouflage degree is realized.

This invention relates to the field of knowledge graph generation technology, specifically to a knowledge graph generation method based on a vector database. It includes the following steps: S1, collecting knowledge data and preprocessing the data; S2, identifying and quantizing entities in the text data using the BERT-Base language model and GloVe word embedding technology, and storing them in a vector database; S3, updating and adjusting the entity vectors stored in the vector database in real time using adaptive entity embedding adjustment technology; S4, analyzing and learning the vector representations of entities and their relationships using a graph attention-based relational reasoning mechanism to generate a knowledge graph; S5, storing the knowledge graph in the vector database. This knowledge graph generation method based on a vector database, through adaptive entity embedding adjustment technology and a graph attention-based relational reasoning mechanism, achieves knowledge graph generation and real-time updating of entity vectors, improving the accuracy of the knowledge graph.

This application relates to the field of underwater robot technology and discloses a multimodal fusion intelligent detection method for bridge defects using underwater robots. The method includes: acquiring a set of underwater environmental parameters; dynamically adjusting imaging parameters by optimizing the light propagation path to overcome problems such as light attenuation and scattering in the underwater environment; utilizing multi-sensor data fusion technology, combining sonar and optical imaging data to generate a high-quality sequence of bridge surface images; extracting, classifying, and weighting defects in the images to generate an intermediate image set containing defect features; and performing multi-dimensional verification and local correction to ensure the accuracy of the defect features; finally, by evaluating the quality of the intermediate image set, standardized image data is generated to provide reliable input for defect extraction. This application can provide efficient and accurate technical support for bridge maintenance and safety assessment.

This application provides a security assessment method, device, storage medium, and program product for financial applications, relating to the fintech field. The method includes: quantifying the raw security data of the financial application according to preset data quantification rules to obtain multiple security dimension indicators; obtaining a security profile score based on the business scenario type and the multiple security dimension indicators; performing code structure parsing processing on the code data in the raw security data to obtain a code structure representation, and obtaining vulnerability attribution relationships based on the code structure representation, vulnerability detection results, and business metadata; obtaining the business system type to which the financial application belongs, and determining the vulnerability remediation priority based on the vulnerability detection results and the business system type using preset priority determination rules; and generating and outputting a security assessment result based on the security profile score, vulnerability attribution relationships, and vulnerability remediation priority. This application improves the accuracy of financial application security assessment results.

The present disclosure provides a facial image correction method and device, a computer-readable medium and an electronic device, and relates to the field of image processing technology. The method includes: in response to detecting a distorted face image, the distorted face image includes a face area, determining a non-distorted area in the distorted face image; if the non-distorted area is detected to include a partial area of ​​the face area, determining the adaptive correction strength corresponding to the face area; performing distortion correction on the face area according to the adaptive correction strength to obtain a face image after distortion correction. The present disclosure can effectively improve the distortion correction effect of larger faces in selfie images, improve the restoration of face content, ensure the accuracy of expression of face content after distortion correction, make the face content in the image more consistent with the real face, and improve the user's photo-taking experience.