120results about How to "Avoid redundancy" patented technology

The invention provides a rainfall monitoring method and system based on unmanned aerial vehicle multi-mode noise reduction in the technical field of environment monitoring and unmanned aerial vehicle application, and the method comprises the steps: S1, collecting a visible light image and a long-wave infrared image through a multispectral vision module after an unmanned aerial vehicle takes off, and collecting a rainfall audio from an acoustic cabin through a microphone; s2, carrying out noise reduction processing on the visible light image, the long-wave infrared image and the rainfall audio; s3, extracting a raindrop size distribution histogram and a spatial density thermodynamic diagram based on the visible light image and the long-wave infrared image, and extracting acoustic MFCC features based on rainfall audio; s4, performing feature fusion operation based on the raindrop size distribution histogram, the visual density thermodynamic diagram and the acoustic MFCC features to obtain rainfall intensity; and S5, carrying out cumulant dynamic compensation based on the rainfall intensity to obtain the cumulative rainfall. The rainfall monitoring method has the advantages that the precision, robustness and practical level of rainfall monitoring are greatly improved.

The invention discloses a two-stage multi-mode bearing fault diagnosis method based on a pre-training large model, and belongs to the technical field of bearing fault diagnosis. The method aims at solving the problems that a traditional method is poor in generalization and poor in robustness under multiple working conditions and small sample conditions. The method comprises the following steps: firstly, constructing a learnable multi-modal Tokens which comprises a multi-scale patch Token, a feature Token and a fault Token, and realizing efficient extraction and fusion of multi-modal features; a time-frequency semantic fusion module is introduced, and comprehensive time-frequency features are output through adaptive frequency coding, time coding and multi-modal fusion; and inputting the multi-modal feature sequence into a pre-training BERT model, and adopting a two-stage training strategy, in the first stage, performing self-supervised pre-training by taking mask signal reconstruction as a target, and in the second stage, performing parameter fine tuning by taking fault classification as a target. According to the method, the diagnosis accuracy and the cross-working-condition generalization ability under the small sample condition can be remarkably improved.

The present invention provides surgical instruments configured to engage with bone anchor assemblies to provide modular platforms for performing various steps of surgical procedures, such as spinal rod reduction and counterrotation. For example, in one embodiment, the surgical instrument may include an outer sleeve terminating in a pair of extensions, an inner sleeve having a proximal threaded portion and a distal translational portion and configured to pass through the outer sleeve, and a pair of pivoting arms received in the extensions of the outer sleeve. These pivoting arms may be configured to extend into a channel in the outer sleeve to engage the bone anchor to the outer sleeve, and the proximal end portion of the outer sleeve may include one or more flat portions configured to engage with another instrument.

The application belongs to the technical field of positioning and measurement, and particularly relates to a target positioning method and device combining CSI coarse positioning and Y-type CCD fine correction, which comprises the following steps: S1, CSI acquisition equipment is arranged in a target area, and CSI data between a sending end and a receiving end is acquired; S2, a Y-type CCD sensor is geometrically calibrated and a coordinate system is corrected, and coordinate reconstruction is completed; S3, the CSI data is preprocessed, noise is removed, and input is normalized; S4, time domain features of the CSI are extracted by using a time domain branch; S5, frequency domain features of the CSI are extracted by using a frequency domain branch; S6, a gating fusion module is used to fuse the time domain and frequency domain features, and the approximate position coordinates of a target are output; S7, the Y-type CCD sensor is controlled to optically acquire a target area, and point light source coordinate information is acquired; and S8, error correction is performed on the data acquired by the Y-type CCD, and the accurate position coordinates of the target are output.

The embodiment of the invention provides a method and equipment for determining a power system configuration scheme and a medium. The method comprises the following steps: determining the initial inertia of a system; if the initial inertia of the system is smaller than the inertia demand value, the reactive response capability of each new energy node is calculated, and a first node is determined; if the first node is different from the previous configuration scheme, taking a new energy node of which the short-circuit ratio is smaller than a short-circuit ratio threshold value as a second node after simulating to carry out network construction type transformation on the first node; determining a current configuration scheme and a corresponding configuration cost; after the second node is simulated to be subjected to synchronous phase modifier configuration, the initial inertia of the system is determined again, and the step of judging whether the initial inertia of the system is larger than or equal to the inertia demand value or not is executed again; and if the first node is the same as the previous configuration scheme, selecting the scheme with the lowest configuration cost. According to the scheme, various configuration schemes of the new energy power system can be determined, the scheme with the lowest cost is selected as the target configuration scheme, and the economical efficiency of system transformation is guaranteed.

The application discloses a hyperspectral image target intelligent detection method, and belongs to the field of remote sensing image processing. The method comprises the following steps: selecting background training samples based on density clustering and spectral angle mapping; generating sufficient target training samples by using a random spectral band zero masking method; inputting the background training samples, a prior target and the target training samples into upper, middle and lower branches of a spectral aggregation and separation network respectively for training, and using a triple soft loss function to promote the network training to separate the background samples, the prior target and the target samples; inputting a pixel to be detected and a prior target into the upper and lower branches and the middle branch of the trained network for target detection, and outputting the network to form a final detection result. By using the application, the difference and similarity of the background and target pixels in the hyperspectral image and a prior target can be simultaneously discriminated, so that the detection result of highlighting the target and suppressing the background can be obtained.

This application relates to the field of battery technology, and more particularly to a battery management component and battery pack. The battery management component includes a battery management module, which comprises: a BDU component, the BDU component including a housing and a relay module, the relay module being located within the housing; and a BMS slave board assembly, located above the BDU component along the height direction of the battery pack. The BMS slave board assembly includes a BMS cover and a slave circuit board, the BMS cover being disposed on the housing, and a cavity being formed between the BMS cover and the housing, with the slave circuit board located within the cavity. This application is beneficial for improving the energy density of the battery pack.

This invention relates to the field of medical image processing technology, specifically to an automatic segmentation and recognition method and system for cancer images based on the fusion of multimodal images. The method includes: acquiring a multimodal medical image set containing computed tomography (CT) images, magnetic resonance imaging (MRI) images, and positron emission tomography (PET) images; generating spatially aligned data blocks through three-dimensional spatial registration and voxel alignment; extracting fusion features using an improved encoder-decoder network optimized based on the characteristics and spatial dependencies of different modal channels; obtaining enhanced semantic feature maps through multi-level context-aware processing; identifying suspected cancer regions through a region candidate generation network; and simultaneously outputting pixel-level segmentation masks and cancer type probabilities through a segmentation network, ultimately generating a three-dimensional segmentation and recognition result. This method can accurately capture multimodal-specific features, clearly distinguish the boundaries between lesions and normal tissue, simplify the processing flow, reduce errors, and achieve accurate and efficient segmentation and recognition of cancer lesions.

This invention discloses an infrared small target detection method based on a frequency-space collaborative gating attention network, belonging to the field of infrared small target detection. The method includes: inputting an initial feature map into a pre-trained frequency-space collaborative gating attention network model, the model including an encoder module, a wavelet prior module, and a decoder module; the shallow coding layer of the encoder module extracts low-level spatial features, and the deep coding layer combines frequency-space collaborative gating enhancement with frequency-domain prior information to obtain an encoder-enhanced feature map; the wavelet prior module obtains frequency-domain prior information based on the low-level spatial features and inputs it into the deep coding layer; the shallow decoding layer of the decoder module performs frequency-space collaborative gating enhancement, and the deep decoding layer extracts spatial detail features; the infrared small target detection result is obtained based on the output of the decoder module. This invention can improve the target detection rate and reduce the false alarm rate under conditions of complex backgrounds, strong noise, low signal-to-noise ratio, and insignificant target features, enhancing stability and robustness in practical engineering environments.

The invention provides a test case generation method and system based on multi-modal machine learning and dynamic optimization, and the method comprises the steps: carrying out the cross-modal feature association of a demand semantic text, a source code instruction stream and a historical defect event record in a software test scene, generating a multi-modal association feature space, inputting a dynamic attention alignment network, and carrying out the multi-modal feature association. Weight distribution is carried out on the business rule semantic dependency relationship, the code execution path dependency relationship and the defect propagation link association relationship, a joint representation vector is generated, double-task cooperative training of test case generation and defect positioning prediction is executed on the joint representation vector, feature extraction weights of double tasks are adjusted to generate a test case generation parameter set, and the test case generation parameter set is subjected to defect positioning prediction. Inputting the boundary test scene into a generator of the generative adversarial network, strengthening feature expression of the boundary test scene, generating a boundary test case set, performing priority ranking on the boundary test case set to calculate case coverage weight, and generating a target test case set. According to the invention, the effectiveness and comprehensiveness of the test case and the utilization efficiency of test resources are improved.

Embodiments of the present specification disclose a multi-granularity graph contrast learning method, device, equipment, medium and program product. The method comprises: using a granule enhancement method to adaptively refine an original graph to generate a plurality of target granules; the original graph is composed of a plurality of nodes and edges representing the interaction relationship between the nodes; the target granule represents a multi-granularity homogeneous region composed of at least Z nodes in the original graph; Z is a positive integer greater than 1; a positive sample pair and a negative sample pair are constructed based on a plurality of target granules; the positive sample pair is composed of nodes in the same target granule in the original graph; the negative sample pair is composed of nodes in different target granules in the original graph; a contrast loss is determined based on the positive sample pair and the negative sample pair; and an initial vector of each node in the original graph is optimized based on the contrast loss to obtain a target vector corresponding to each of the plurality of nodes.

The utility model discloses a kind of tool for bundled steel strand unpowered take-up, including support, power cylinder, respectively fixed mounting base and first clamping component installation part at both ends of support, between base and first clamping component installation part and the second clamping component installation part of support sliding connection;Power cylinder is used to drive second clamping component installation part reciprocating sliding between base and first clamping component installation part;First clamping component and second clamping component parallel layout, including the connecting anchor plate with the corresponding taper hole of steel strand, the lifting clamping piece compatible with taper hole, the baffle fixedly connected with connecting anchor plate at the big mouth end of taper hole, the elastic member with baffle and lifting clamping piece respectively at both ends abut;The big mouth end of taper hole is located at one end of connecting anchor plate close to base.The utility model discloses a kind of tool for bundled steel strand unpowered take-up, can simplify control oil circuit, reduce tool weight, improve take-up speed and be convenient for installation.

The application provides a road surface state recognition method and device, and solves the problems of single road surface state recognition category, coarse result granularity and inability to distinguish specific road surface abnormal types caused by dependence on simple threshold judgment and shallow features in the prior art. The method comprises the following steps: acquiring a vertical vibration acceleration data segment of a vehicle in operation; extracting a characteristic value according to a preset optimal feature subset, wherein the subset comprises a frequency domain peak value and a spectrum energy; inputting the characteristic value into a classification model to obtain a recognition result, wherein the model is obtained by training based on data comprising at least three road surface state labels. Through fixed feature optimization, the application realizes fine recognition of multiple states such as well frame difference, expansion joint and deceleration strip, and significantly improves the calculation efficiency while ensuring the accuracy.

This invention provides a method for predicting the TCM functions of a single herbal drug and for drug screening. The method includes: treating each atom in the molecule of the herbal drug to be predicted as a node in a graph; constructing feature vectors for each node based on the physical and chemical properties of each atom; constructing a feature vector matrix based on the feature vectors of all atoms; treating the chemical bonds between atoms as edges in the graph; constructing an adjacency matrix based on the bonding types between atoms; constructing a graph structure matrix based on the feature vector matrix and the adjacency matrix; inputting the graph structure matrix into a graph convolutional neural network model to obtain a function score vector, the length of which is equal to the total number of preset TCM functions, and each score in the function score vector representing the probability that the herbal drug to be predicted belongs to a corresponding preset TCM function; comparing each score in the function score vector with a preset threshold, and determining the predicted TCM function value of the herbal drug to be predicted based on the comparison result. This invention improves the accuracy of function prediction.

The application discloses a kind of knowledge embedding and retrieval method and system of multi-modal PDF document, specifically relates to knowledge embedding and retrieval field, the present application uses multi-modal big model identification picture, table and other non-text elements in PDF, generates its position coordinate, converts picture into structured literal description and stores source URL;Second, the original text, picture description and URL are integrated into Markdown format, then, adopt the strategy of gradual blocking: each picture information is individually used as text block, and the remaining pure text is blocked according to character length;Subsequently, all text blocks are vector embedded, and the type is stored as metadata into vector database;Finally, in the retrieval stage, relevant text blocks are returned based on vector similarity;In the generation stage, single mode or multi-modal model generates accurate answer using text content and picture URL.

This application provides a multi-scale island and reef remote sensing image mosaic mapping method, which can be applied to the fields of geographic information science and remote sensing big data processing technology. The method includes: acquiring island and reef image files corresponding to each island and reef in a target area; the island and reef image files are high-resolution remote sensing images that have undergone cloud removal processing and are cropped according to a buffer distance, with the geometric center of the island and reef as the reference; based on the spatial display scale corresponding to the buffer distance, stitching together the island and reef image files of all islands and reefs in the target area to generate an image mosaic map; generating a vector index map that corresponds one-to-one with the geometric position of the image mosaic map, the vector index map containing island and reef identification information of the corresponding geometric position of the island and reef; and superimposing the vector index map onto a predetermined area of ​​the image mosaic map to generate a standardized island and reef thematic map of the target area. This application also provides a multi-scale island and reef remote sensing image mosaic mapping device, equipment, and storage medium.

The utility model discloses a wire coil binding structure for a small animal exercise physiological parameter monitor, which comprises a storage box, the surface of the storage box is fixedly connected with the monitor, the inner wall of the storage box is rotatably connected with a rotating tube, the inner wall of the storage box is provided with a conductive slip ring and a power connection end, and the power connection end is connected with the rotating tube. The front end of the power connection end is electrically connected with a power supply circuit of the monitor, a power line is wound around the surface of the rotating pipe, and a sliding rail is fixedly connected to the top of the storage box through bolts. The L-shaped block drives the rack to move, so that the gear drives the first bevel gear to rotate through the shaft rod, then the first bevel gear drives the rotating pipe to rotate through the second bevel gear and winds the power line, the power line of the parameter monitor can be properly wound, redundant power lines can be collected into the storage box, and the power line can be conveniently stored. Therefore, power line redundancy of the parameter monitor in actual use is avoided.

The image non-uniformity correction method and device, the infrared image device and the computer readable storage medium provided by the embodiments of the present application, the method comprises: acquiring a same group of in-focus images and out-of-focus images collected for a target scene; respectively decomposing the in-focus images and the out-of-focus images through wavelet transform to obtain corresponding in-focus sub-band images and out-of-focus sub-band images; determining a fixed image noise component contained in a high-frequency component in the out-of-focus sub-band images based on the high-frequency component; determining a corresponding image component estimation value according to a sub-band coefficient of the in-focus sub-band images and the corresponding fixed image noise component; reconstructing the image component estimation value through inverse wavelet transform to obtain a non-uniformity corrected image of the target scene, thus avoiding information loss and redundancy of the image signal in the decomposition process, so that the non-uniformity corrected image of the target scene is obtained more quickly and more clearly.

This invention discloses an experimental water tank device for investigating changes in sand ripples on a near-bottom high-concentration sediment bed. The device includes a water storage tank, a straight-channel experimental water tank, and a sedimentation tank. The two sides of the straight-channel experimental water tank are connected to the water storage tank and the sedimentation tank, respectively. A transparent glass panel is installed on the surface of the straight-channel experimental water tank. A trolley track is fixedly installed at the upper end of the straight-channel experimental water tank, and a trolley is mounted on the track. A measuring mechanism is installed on the trolley, and a rangefinder is installed in the middle of the trolley. A flow meter is installed on one side of the trolley, and a 3D scanner is installed on the other side. The 3D scanner is installed on the lower end of the trolley's platform, with its scanning probe facing outwards and tilted downwards. A high-definition camera is positioned on the outside of the transparent glass panel. The measuring mechanism establishes a data transmission connection with a data terminal via a wiring harness. This invention features simple and quick assembly, high automation and efficiency, and the ability to continuously collect data on changes in sand ripples on the sediment bed surface in multiple dimensions.

The application discloses a data acquisition method and device, electronic equipment and storage medium. The method comprises: acquiring audio to be processed collected through a microphone array, the audio to be processed comprising multiple sentences of multi-channel audio; for each sentence of multi-channel audio, acquiring a phase difference feature between at least one group of specified channel audio as target data, the specified channel audio comprising at least two position-associated channel audios. For each sentence of multi-channel audio, the method performs phase difference calculation by incorporating the audio of at least two position-associated microphone channels, which can generate higher-dimensional phase information coverage on one hand, so that the DOA feature obtained after phase difference is more intensive, thereby helping to improve the accuracy of the DOA encoding result; on the other hand, since the position-associated microphone channel audio is used for phase difference calculation, data redundancy can be avoided, thereby helping to improve the efficiency of DOA encoding.

This utility model discloses a thermal management system suitable for containerized energy storage systems, comprising an air conditioning refrigerant circuit and a battery thermal management circuit, both of which include plate heat exchangers. Through this structure, the integration of the air conditioning refrigerant circuit and the battery thermal management circuit reduces system complexity and cost, while also reducing space requirements and maintenance costs. The battery thermal management system utilizes the plate heat exchanger to form a heat exchange with the air conditioning refrigerant circuit, effectively dissipating the heat absorbed from the energy storage battery module through the air conditioning system. The air conditioning refrigerant circuit dynamically adjusts the refrigerant distribution to regulate the heat exchange distribution between the air conditioning refrigerant circuit and the battery thermal management circuit. The battery thermal management circuit includes a phase change thermal storage unit connected in parallel with the plate heat exchanger, releasing stored heat / cold energy during charging and discharging intervals, reducing the frequent start-stop of the compressor and PTC heater.

The invention relates to a process equipment model selection method, device and equipment for ship block manufacturing and a medium. According to the method, production demand prediction, site current situation analysis, production takt extraction, process equipment parameter analysis and cost dynamic evaluation and score screening can be carried out, so that the whole process scientization and datamation of equipment model selection are realized, the process equipment demand of the shipyard segmented manufacturing process can be accurately matched, the production layout is optimized, and the production efficiency is improved. Equipment redundancy or insufficient performance is avoided, cutting and manufacturing precision and efficiency are improved, production capacity is improved fundamentally, energy consumption and maintenance cost are reduced, and product quality stability is enhanced. By accurately matching process equipment and production requirements and optimizing the production process, the segmented manufacturing production capacity and production efficiency can be improved, the production cost can be reduced, the stability of the product quality is improved, the process equipment with high intelligent and automatic degree is introduced, and upgrading and updating of the shipyard production technology are promoted.

The present application takes the position closed structure and the orthogonal component difference as the core, converts the information to be processed into the residual structure representation under the background reference state constraint, the background reference state is constructed by the receiving and transmitting parties according to the consistent generation rule, and can be used as the structural key or the key derivation reference in the safe scene, so that the transmission and storage load can be concentrated on the residual field and the necessary auxiliary field. The framework uniformly carries the encoding, compression, encryption, inversion, decoding and structured domain calculation and topological consistency reasoning, realizes the integration of compression and encryption, storage and calculation. The system improves the equivalent storage density and energy efficiency performance through structured aggregation, and provides consistent logical pointing and calculation basis for reconstruction and reasoning.

The application belongs to the technical field of underwater topographic survey, and particularly relates to a high-precision data acquisition method and system for underwater topography of an unmanned ship, which comprises the following steps: performing static alignment on an inertial measurement module, calibrating eccentricity vectors between GNSS and a multi-beam sounding module, and establishing a unified space-time reference system; dynamically generating a non-uniform survey line set according to digital elevation information, collecting echo data and synchronously acquiring position and attitude information; periodically acquiring layered sound velocity parameters, correcting sounding deviation to obtain corrected water depth values; performing multi-stage in-flight real-time filtering on beam data to remove abnormal values, fusing the filtered data with position and attitude information based on the unified space-time reference system to generate three-dimensional point coordinates; constructing an underwater digital elevation model, and calculating a mean error by using water depth differences in overlapping areas of adjacent survey lines, and outputting the model when the mean error is less than a threshold value. The application can output an underwater digital elevation model meeting the expectation by sensor alignment, layered sound velocity correction, three-stage filtering and mean error evaluation.

The application relates to the technical field of power system data analysis and artificial intelligence, and discloses a decision target and multi-source heterogeneous data auxiliary decision report generation method and system; the generation method realizes the automatic generation of the power auxiliary decision report through decision target driven information dimension analysis, multi-source text standardization processing, long text global semantic coding, key information screening, small sample driven optimization and structured rule mapping; the extraction result is highly matched with specific decision targets such as dispatching and fault disposal, and the decision adaptability is strong; the method can be migrated across scenes with few labeled samples, has strong generalization ability, and has low adaptation iteration cost; in combination with secondary error correction and semantic filtering, the method effectively resists the interference caused by oralization and non-standard format, is stable in extraction precision, retains the cross-paragraph correlation information in the long text through semantic blocking and cross-block attention fusion, and significantly improves the efficiency of power text analysis and the objectivity and timeliness of the decision report.

The application provides a 3D reconstruction system for mountain pine forest environment, and relates to the technical field of unmanned aerial vehicle (UAV) simultaneous localization and mapping (SLAM). The system comprises a hardware platform and a software system, which are composed of a UAV, a laser radar, an inertial measurement unit (IMU) and a global navigation satellite system (GNSS). In view of the problems of sparse features in forest areas, easy drift and high redundancy caused by repetition during mapping, the software system adopts three modules to build a closed loop: 1) a data alignment and system initialization module, which realizes optimal robust alignment of laser and GNSS trajectories by embedding weighted Procrustes analysis into a RANSAC framework; 2) a key point cloud frame selection module, which scores the current frame by comprehensively considering geometric, density and structural characteristics, and dynamically adjusts the space-time threshold to select key frames in combination with historical feedback; and 3) a graph optimization module with multi-dimensional factor constraints, which fuses odometry, GNSS and loop factors to build a factor graph, and globally optimizes the mapping algorithm by using incremental smoothing. The application effectively suppresses cumulative errors and improves the accuracy and efficiency of forest reconstruction.

The application relates to the technical field of testing, and discloses a test method for an application system, which comprises the following steps: acquiring a system configuration file, a local application program package and a corresponding local Java bytecode file of a local application system, and determining a target full-amount corresponding relationship of the local application system; acquiring different Java bytecode files of different versions of application program packages, and determining a program difference list; acquiring an outbound configuration file, and combining the target full-amount corresponding relationship and the program difference list to determine a first full-amount test range of the local application system and a plurality of second full-amount test ranges of a plurality of other application systems, and then testing the application system to obtain a target test result of the application system, which realizes bidirectional accurate analysis of a test range of a local program change on itself and other systems and of other system changes on the local system, comprehensively covers transaction interface and business function test requirements among multiple systems in an enterprise, and guarantees the overall operation stability of the application system.

The application belongs to the technical field of automobiles and discloses a front windshield lower decorative plate and an automobile, which comprise a front windshield lower decorative plate body and a side shielding plate which is detachably installed on the side of the front windshield lower decorative plate body and covers the exposed area of a sheet metal gutter assembly. The side shielding plate comprises a hard rubber part and a soft rubber part. The hard rubber part is shaped according to the contour of the exposed area of the sheet metal, and the soft rubber part is arranged on the circumferential edge of the hard rubber part and is attached to the adjacent automobile body part. The application aims to solve the problems of insufficient shielding caused by the exposed joint of the front windshield lower decorative plate and the sheet metal gutter assembly and poor sealing caused by the direct exposure of the joint gap between the two, improve the overall appearance of the front compartment area, and prevent dust and water vapor from the outside from entering the interior of the front compartment through the gap.

The invention provides a hybrid deep learning automobile sales prediction method fused with multi-source features and application, and belongs to the technical field of hybrid deep learning fused with multi-source features. The method comprises the following steps: 1, inputting subjective evaluation data into a semi-supervised feature generation layer, and extracting features in the subjective evaluation data through a method based on semi-supervised learning feature clustering; 2, inputting static non-time-series data and the subjective evaluation type data and the time series data in the step 1 into a parallel feature extraction layer, and performing feature extraction; 3, performing deep fusion on the output in the step 2, and inputting the merged feature vector into a fusion optimization layer; step 4; and using an Adam optimization algorithm to train the models formed in the steps 1-3. According to the method, combined feature extraction and fusion of non-time-sequence feature mixed data of time sequence features, static attributes and subjective evaluation classes are realized, so that the automobile sales volume prediction precision is optimized.