15results about How to "Guaranteed fidelity" patented technology

The present application provides a kind of night image halo removal method, including constructing prior knowledge, according to prior knowledge, construct initialization network and depth unfolding network, using initialization network to extract variable initial value set from night halo pollution image;Night halo pollution image and variable initial value set are input into depth unfolding network and are operated multiple iterations, and the final halo removal image is extracted from night halo pollution image;Wherein, depth unfolding network includes multiple proximal networks, and each proximal network participates in an iteration operation.The present application makes full use of prior knowledge, uses the depth unfolding network with multiple proximal networks to iterate mapping diagram, halo-free feature map, constraint variable and halo removal image, so as to extract the final halo removal image, to better preserve image texture and restore image details, under the premise of not increasing application cost, the halo and stripe artifact in image are preferably removed.

The invention relates to the field of data transmission, in particular to an internet of things data adaptive compression method and system based on edge computing, and the method comprises the steps: obtaining a historical data set and a to-be-compressed data set, calculating the correlation degree of a target dimension and a contrast dimension, segmenting the to-be-compressed data set, and clustering all segments; calculating the data anomaly probability of the target dimension in the target cluster; calculating a data adaptive compression deviation value of the target dimension in the target segment; calculating an upper door slope and a lower door slope of a target point in a target dimension based on a revolving door algorithm, calculating an upper door effective slope of the target point in the target dimension, similarly obtaining a lower door effective slope, obtaining a compression starting point of the target dimension based on the upper door effective slope and the lower door effective slope, and similarly obtaining a compression ending point; and compressing the data of the target dimension. According to the technical scheme, compression resource consumption can be reduced, and compression efficiency can be improved.

This invention discloses an attention acceleration method, apparatus, device, and medium based on a diffusion model. The method includes: acquiring the input feature set, current time step, and local window parameters corresponding to a target multi-head attention module in a target diffusion model; reshaping the shape of each input feature in the input feature set based on a preset tensor format to obtain a three-dimensional feature set; determining a feature recombination strategy corresponding to the three-dimensional feature set based on the current time step, and recombinating the three-dimensional feature set based on the feature recombination strategy to obtain a target feature result set corresponding to the target multi-head attention module; and performing local window attention calculation on the target feature result set based on the local window parameters and the feature recombination strategy to obtain the target attention result corresponding to the target multi-head attention module. Through the technical solution of this invention, efficient acceleration can be achieved while maintaining generation quality, improving the actual utilization rate of hardware.

The invention provides an image fusion method and device, and belongs to the technical field of image fusion, and the method comprises the steps: importing a plurality of original high-resolution hyperspectral images, and carrying out the down-sampling analysis of the original high-resolution hyperspectral images, and obtaining an original low-space hyperspectral image and an original high-space multispectral image; performing fusion analysis on the original low-space hyperspectral image and the original high-space multispectral image through a training model to obtain a target high-resolution hyperspectral image; and updating and analyzing the training model according to the target high-resolution hyperspectral image and the original high-resolution hyperspectral image to obtain an image fusion model. According to the method, the fidelity of the space and the spectrum of the fused image is kept, the calculation requirement is reduced while the spectrum and the space information are recovered, the calculation complexity is also reduced, better algorithm interpretability is achieved, and therefore a more accurate fusion result is obtained.

The invention discloses a video compression method based on motion estimation and affine transformation substitution. The method comprises the following steps of: 1, performing feature analysis and initial coding on an image to realize initial compression of an original video frame, and establishing spatial geometric association between adjacent frames; 2, mapping high-resolution homography to a low-resolution coding space by using a down-sampling matrix based on spatial geometric correlation to obtain a down-sampling homography matrix; step 3, according to the mask region of the moment coding data obtained by calculation in the step 2, separating to obtain a code of a cutting background, namely a code of an increment background, and realizing differential compression; and 4, according to the codes of the incremental background obtained in the step 3, the decoding end combines the stored codes at the T-1 moment to splice and fuse reconstructed coding data at the T moment, and then video frames are decoded. According to the invention, the adaptive compression coding capability for different regions in the video compression process is effectively improved.

The invention provides a rolling bearing multi-working-condition data expansion method based on digital twinning, and the method comprises the steps: collecting original vibration signals of a rolling bearing under various working conditions through a physical experiment table system, and recognizing an equivalent boundary condition; solving the six-degree-of-freedom coupling dynamic model to obtain an acceleration vibration signal for fault diagnosis; performing model fault verification on the six-degree-of-freedom coupling dynamic model through the five evaluation dimensions to obtain a high-fidelity model; based on the high-fidelity model, utilizing an AGA algorithm and a KAN network to construct a nonlinear mapping relation between a working condition and an optimal physical parameter vector, outputting an accurate model parameter suitable for a new working condition, and obtaining an adaptive digital twinborn model; and automatically generating a hybrid enhanced data set by utilizing a batch processing interface of the constructed virtual interaction platform. According to the method, data generation and expansion of the rolling bearing under multiple working conditions can be realized, and rich and controllable data support is provided for fault diagnosis, health monitoring and model training.

This invention provides a self-supervised arbitrary-scale point cloud upsampling method and system, including: automatically adjusting the point density of midpoint interpolation using a local geometric feature analysis module to construct an intermediate candidate point set; obtaining interpolation points by controlling the arbitrary-scale upsampling ratio of the intermediate candidate point set through farthest-point sampling technology; introducing an implicit surface projection optimization mechanism, using the interpolation points as seed points for two-stage optimization, and achieving surface projection by estimating the projection direction and projection distance through an implicit neural function; and optimizing the boundary feature representation of the surface projection by generating a two-dimensional projection map through dynamic scaling of the sparse input point cloud based on a pseudo-true value projection comparison mechanism based on point cloud scale transformation. Through adaptive dynamic neighborhood optimization of local geometric features, arbitrary-scale upsampling ratio control, combined with estimation of projection direction and projection distance, and pseudo-true value projection comparison mechanism, the method ensures geometric detail fidelity, point cloud uniformity, and arbitrary-scale generalization ability.

The application provides an intelligent pre-control system and method for engine cold-heat shock test, wherein the method comprises the following steps: receiving real-time state data through a prediction module, combining a working condition instruction generated by a main working condition control module, and online simulating and predicting a future thermal state of a measured engine under a current to-be-executed working condition instruction; an intelligent scheduling unit dynamically decides whether to insert a pre-adjustment step before execution of the current to-be-executed working condition instruction based on the future thermal state, and determines a starting time and a duration of pre-adjustment; if it is decided to insert, an adaptive control instruction is generated, otherwise the current to-be-executed working condition instruction is taken as a to-be-executed instruction; and an execution control module drives an execution mechanism group to operate a temperature control system according to the adaptive control instruction or the to-be-executed instruction. The application is suitable for a working condition switching process of engine cold-heat shock test, can inhibit temperature overshoot caused by system thermal inertia through active prediction and advanced intervention, and guarantees test data accuracy and improves test efficiency.

This disclosure relates to the field of geophysical exploration technology, and particularly to a bubble response suppression method, apparatus, medium, and product based on seismic oceanography. The method includes: acquiring raw marine seismic data, which exhibits a bubble response; simulating the raw marine seismic data using a far-field wavelet to obtain a first wavelet; extracting a second wavelet from the raw marine seismic data using statistical methods; performing high-order autocorrelation fusion processing on the low-frequency signal of the second wavelet and the high-frequency signal of the first wavelet to obtain a fused desired wavelet; and applying the fused desired wavelet to the raw marine seismic data to perform bubble response suppression processing, thereby obtaining target marine seismic data. Therefore, the bubble response suppression method based on seismic oceanography provided in the exemplary embodiment of this disclosure not only effectively suppresses the bubble response but also avoids the influence on high-frequency components in the data.

The application discloses a kind of high-efficiency human posture estimation methods for embodied intelligent interaction, including feature extraction module, including convolution, for carrying out down-sampling to input image, obtain image down-sampling feature;Multi-stage posture representation module, the multi-stage posture representation module is composed of three dynamic Token perception stages, respectively using first dynamic Token perception submodule, second dynamic Token perception submodule and third dynamic Token perception submodule are realized, three dynamic Token perception submodule network structure is same, and all includes Token aggregation unit and mixed scanning unit;Image down-sampling feature is serialized and input to multi-stage posture representation module, for from coarse to fine perception human key point position in the image down-sampling feature.The application enhances the modeling capacity of human key point local and global information, realizes accurate, efficient posture estimation under lower computational load, improves the execution effect of robot, embodied intelligent interaction and other related tasks.

A hyperspectral image reconstruction method based on spatial-spectral fusion comprises the following steps: extracting a deep feature map of Bayer data according to a Bayer format spatial position through an adaptive spatial convolution module, constructing a spatial mode mask and generating an adaptive kernel weight; constructing a parallel spatial network branch and a spectral network branch model, respectively performing local feature enhancement in the spatial dimension and the spectral dimension, and adding to obtain a spatial-spectral feature map; constructing a spatial-spectral fusion attention module, and adopting a multi-dimensional cross-spectral self-attention unit and a spatial rectangular self-attention unit to fuse the feature map; and finally outputting a hyperspectral image through channel mapping. The present application can directly use Bayer data to reconstruct a hyperspectral image, reduce errors introduced by demosaicing processing, effectively balance spatial details and spectral fidelity through global and local feature collaborative optimization, and improve the robustness and accuracy of the model.

The application discloses a cell type annotation method, device, equipment and medium, relates to the technical field of bioinformatics and medical image processing, and comprises the following steps: constructing a whole image feature matrix and a spatial neighborhood graph according to cell self characteristics and cell spatial positions of single cell samples of historical multiple tissue images; inputting the whole image feature matrix and the spatial neighborhood graph into a cell annotation model comprising a feature encoder, a spatial aggregator, a classification decoder and a prototype consistency constraint module; mapping the whole image feature matrix into an initial hidden vector by the encoder, determining attention coefficients of effective neighboring cells on the single cell samples based on the spatial neighborhood graph, weighting and summing the initial hidden vector according to the attention coefficients by the aggregator, outputting a predicted type probability distribution according to spatial perception characteristics by the decoder to generate a reconstructed feature; obtaining a prototype consistency loss value by the constraint module; updating and iterating the model based on the prototype consistency loss value to obtain a target cell annotation model; and performing cell type annotation by using the model.

This application relates to the technical field of intelligent battery monitoring, and in particular to an intelligent monitoring system for engineering vehicle batteries. The system includes an electrical parameter sensor, and further comprises: a data acquisition unit for acquiring the real-time vibration frequency and real-time driving speed of the engineering vehicle; an adaptive vibration suppression unit for dynamically adjusting the equivalent stiffness of the mounting structure of the electrical parameter sensor based on the real-time vibration frequency, causing the natural frequency of the mounting structure to deviate from the real-time vibration frequency; and a collaborative noise reduction unit for: determining the center frequency of signal processing based on the real-time vibration frequency, and simultaneously adjusting the signal processing bandwidth associated with the center frequency based on the real-time driving speed, and using the determined center frequency and the adjusted signal processing bandwidth to perform noise reduction processing on the signal output by the electrical parameter sensor. This application can solve the problem of inaccurate battery signal acquisition in road rollers under vibration and speed-changing coupled operating conditions, achieving high-precision monitoring.

The present application relates to power conversion circuit electric parameter measurement technical field, disclose a kind of based on the abnormal detection method of steering controller of multi-signal analysis, comprising: obtaining power loop current sampling sequence and pulse width modulation instruction signal;According to the high-frequency switching ripple in phase current extracted by the jump edge timing of pulse width modulation instruction signal;With pulse width modulation instruction signal as phase reference, construct quadrature reference sequence, and obtain quadrature current component from high-frequency switching ripple using phase-sensitive detection principle;Combined with DC bus voltage, the complex impedance of power loop is calculated, and the performance degradation state of component is determined according to the evolution trend of complex impedance, the present application uses the masking effect of endogenous high-frequency excitation penetration control algorithm, analyzes the electrical characteristics reflecting the physical properties of circuit, and improves the signal-to-noise ratio of electric parameter monitoring.

The invention discloses a method for calculating the water entry slamming pressure of an open cavity in waves, relates to the technical field of ocean engineering, is used for predicting the water entry process of the open cavity, and comprises the steps of constructing a geometric model of the open cavity, constructing a wave field, and simulating the movement process of a target cavity entering the wave field and the movement process of the target cavity in the wave field. Constructing a time domain motion equation of the target cavity, solving the time domain motion equation of the target cavity by adopting a numerical integration method, and calculating a hydrostatic pressure component, an entry water impact dynamic pressure component, an in-cavity air fluctuation pressure component and a wave induction dynamic pressure component of the target cavity; and performing decoupling calculation on the water entry slamming pressure of the target cavity based on the four sub-components. According to the method, the coupling dynamic model of the wave field and the water entry process of the open cavity is established by combining a multi-phase coupling action mechanism, and the calculation precision of slamming pressure distribution and time history characteristics of the open cylindrical cavity at the water entry moment and in the subsequent process under the wave condition is improved.