5 results about "Effective algorithm" patented technology

The application discloses a perimeter security disturbance recognition algorithm based on GAF-ConvNeXt-TF, baseline removal and denoising pretreatment are respectively performed on one-dimensional time sequence signals generated by various disturbance behaviors; the pretreated one-dimensional time sequence signals are converted into two-dimensional time sequence diagrams, and a data total set is made according to the types of the disturbances; a ConvNeXt model is established, the training set is imported into the ConvNeXt model for training, so that a training weight file is acquired; and the test set is used to test the accuracy rate of model classification recognition. The application can enable the model to better learn the information of the time dimension; by using transfer learning, loading the pre-training weight can greatly reduce the training parameters, accelerate the model training convergence speed, and improve the accuracy of the disturbance recognition result, and provide an effective algorithm for the disturbance recognition field of the distributed optical fiber sensing system in perimeter security.

This invention belongs to the field of inductance parameter extraction technology, and discloses an efficient method and system for extracting and optimizing the layout of parasitic inductance parameters of power semiconductor modules. Based on the powerful simulation capabilities and co-simulation interfaces of commercial software ANSYS and MATLAB, it achieves efficient interaction of simulation data between software, facilitating the extraction of module parasitic inductance parameters and current sharing characteristics. Utilizing effective algorithms for analyzing parasitic inductance parameters and current sharing characteristics, it provides scientific guidance for optimizing module layout. This invention has the following significant advantages: Compared to the method of "human experience + hardware debugging," this invention adopts efficient simulation development methods, avoiding frequent hardware iterations in the manufacturing and verification stages, reducing development costs, thereby improving the R&D efficiency of multi-chip silicon carbide power modules and better supporting their higher-quality and more flexible development.

The invention relates to the technical field of high-risk operation monitoring, in particular to a high-risk operation intelligent monitoring management system based on multi-element linkage. According to the method, mutual linkage of key elements of high-risk operation is constructed, so that the technical standard of risk perception and management and control in the operation process is realized on the security level; meanwhile, through an electronic fence automatic drawing algorithm and an electronic fence automatic effective algorithm, the operation risk state is judged in a limited space; through micro-grid data acquisition, deep learning and continuous training, an operation risk high-accuracy AI model is obtained, and potential safety hazards of operation of operators on the working face of the fan platform and the working face of the fan tower drum are eliminated to the maximum extent; the working state of the crane is determined according to the proportion of the time nodes when the crane is located outside the electronic fence within the preset duration, and corresponding adjustment is performed when the working state is unqualified, so that the working accuracy of the crane is improved. According to the invention, the operation risk is effectively monitored, and the working efficiency is improved.

Disclosed are a multispectral image denoising method and device based on a structured tensor sparse model. The structured tensor sparse model is based on tensor sparse representation and integrates non-local centralized tensor sparse constraints. The model can sparsely represent multidimensional signals in a group domain, simultaneously realizes local intrinsic tensor sparse regularization and non-local similarity of multidimensional signals in a unified framework, learns an orthogonal dictionary from a group of tensors to generate a basis of a space to which the group of tensors belongs, instead of decomposing each tensor, and can reduce the number of dictionary atoms. A compromise parameter is adaptively determined by using statistical characteristics of data, and an effective algorithm is designed to solve an optimization problem of the proposed model. Therefore, structural similarity between a group of full-band image blocks can be utilized, non-local centralized tensor sparse constraints can be integrated, multidimensional signals in a group domain can be sparsely represented, and local intrinsic tensor sparse regularization and non-local similarity of multidimensional signals can be simultaneously realized in a unified framework.

The application belongs to the technical field of network services, and discloses a high-availability big data stream processing request placement method in a serverless edge network. The purpose is to meet the reliability requirements of users while minimizing processing delay. An effective algorithm is designed to find a suitable number of instances for the function of each data stream processing request, and to place the request while minimizing the average delay experienced by each user while meeting its reliability requirements. After the request is placed, the input data rate may change and is uncertain. An online learning algorithm is designed to predict the method to predict the data rate and dynamically adjust the standby instances to absorb the uncertain data rate. Based on the experiment of a real data set, it is shown that the placement problem of big data stream processing in the edge serverless network is effectively solved.