39 results about "Noise correlation" patented technology

The invention discloses an orchard unmanned aerial vehicle multispectral image noise restoration method based on a generative adversarial network, and relates to the technical field of multispectral image noise restoration, and the method comprises the steps: synthesizing a noisy sample containing stripes, photon particles, thermal noise, Rayleigh, shielding, compression artifacts, overexposure and waveband crosstalk on a noise-free reference image through a conditional generative adversarial network; a crosstalk matrix T obtained through optical calibration is introduced; based on the canopy / soil / sky semantic partition, main-cooperation-subdivision level superposition is implemented; generating hierarchical labels by taking energy conservation consistency and thermal noise correlation as thresholds; a branch and region self-adaptive repair module corresponding to classification is arranged in the image-to-image generative adversarial network, and joint loss training containing T regularization is adopted; and outputting the repaired image subjected to physical consistency checking. The method gives consideration to visual quality and spectral fidelity, inhibits cross-band pollution, improves the stability of indexes such as NDVI and NDRE, and is suitable for precision agricultural monitoring.

The invention belongs to the technical field of Internet of Things data processing, and discloses a multi-mode Internet of Things sensor data fusion processing method and system. Comprising the following steps: acquiring multi-modal sensor data in a monitoring environment; performing interference source positioning analysis on the data to generate an interference source distribution map; separating interference waveform characteristics of each modal data based on the distribution map; constructing an interference propagation network by using the features so as to quantify the coupling relationship of interference among different modes; performing noise correlation weakening on original data based on the network to obtain decoupled sensor data; performing dynamic fusion analysis on the decoupling data, and constructing a multi-modal fusion feature vector; and finally, accurately detecting a target event or state based on the feature vector. According to the method, the negative influence of noise correlation on fusion is eliminated, and the accuracy and robustness of target detection and the reliability of the system in a complex environment are improved.

The invention discloses an image denoising method based on multi-scale feature enhancement and local pixel disturbance. According to the method, features of different scales are extracted by introducing a multi-scale convolution module and cavity convolution, so that the perception capability of complex noise is enhanced. Meanwhile, flat areas are calculated through the local variance of the pseudo clean image, and pixel disturbance is carried out in the areas so as to weaken noise correlation. According to the method, channel attention and space attention mechanisms are combined, and learning of important features is enhanced. In the aspect of loss function design, pixel loss, multi-scale feature loss and local weighting loss are fused. In practical application, a cosine annealing learning rate scheduling and exponential moving average strategy is adopted, and the convergence speed and stability of the model are remarkably improved. According to the method, image details can be effectively reserved, and the method is suitable for denoising tasks of various complex scenes such as natural images and medical images.

The invention provides a signal processing method and device, electronic equipment, a chip and a storage medium, and relates to the field of communication, and the method comprises the steps: carrying out the noise estimation of a receiving signal received by at least one receiving antenna, so as to obtain an initial noise correlation matrix of the receiving signal; according to the initial noise correlation matrix, determining an interference scale factor used for representing the proportion of white noise and interference in the received signal; according to the interference scale factor, correcting the initial noise correlation matrix to obtain a target noise correlation matrix; and performing noise and interference suppression on the received signal based on the target noise correlation matrix to obtain a demodulation signal. Therefore, the difference between the estimated noise correlation matrix and the real noise correlation matrix can be reduced, the accuracy of noise correlation matrix estimation is improved, and on the basis, noise and interference suppression is performed on the received signal based on the accurate noise correlation matrix, so that the finally obtained demodulation signal is closer to the real sending signal.

The invention belongs to the technical field of mining dust collector noise control, and relates to a mining wet dust collector noise reduction method, which comprises the steps of S1, vibration sensor and electromagnetic damper installation, S2, vibration source monitoring and noise correlation analysis, and S3, adaptive damping adjustment and feedback optimization. A three-step closed loop of hardware installation, intelligent analysis and self-adaptive execution and feedback optimization is applied to the mining wet dust collector, real dynamic noise reduction is achieved, an electromagnetic damper and a gradual change damping material work cooperatively, an active and passive composite noise reduction mode is formed, and the mining wet dust collector can adapt to complex working conditions such as underground air quantity change and dust concentration sudden change in real time.

The invention discloses a noise correlation cross-modal matching loss adjustment method and device based on meta-learning. The method comprises the following steps: constructing a pairing sample set containing a multi-modal data pair; quantifying semantic correlation of each multi-modal data pair in the pairing sample set by using a cross-modal model, and selecting an initial loss function; constructing a loss adjustment network, and balancing the noise sensitive loss function and the noise insensitive loss function to obtain an adaptive loss function; optimizing the loss adjustment network by using a meta-learning algorithm; updating model parameters of the cross-modal model on the target noise data set by using the optimized loss adjustment network; and inputting the query target and each object in the matching database into the updated cross-modal model to obtain the similarity of the query object relative to each object in the matching database. The problem that the robustness and the matching accuracy of the cross-modal model are difficult to consider at the same time in the prior art is solved, and powerful support is provided for landing of the cross-modal model in an actual scene.

The invention relates to a beam gain constraint method for a phased array receiver, and the method comprises the steps: placing the phased array receiver in a far-field region as a receiving array, and obtaining the response of each port of the array to a reference signal and background noise through opening or closing a signal generator; a beam forming weighting coefficient is solved in combination with the steering vector and the noise correlation matrix, a corresponding beam forming network is built, a directional diagram and gain of each beam of the array after weighting are measured, a target gain is selected, a corresponding scaling factor is solved in combination with the relation between the target gain and the actually measured gain of each beam, and the beam forming network is built; and correcting the weighting coefficient of each beam by using the scaling factor, and finally achieving the purpose that the gain of each beam is constrained to the target gain. According to the method, an independent beam gain constraint unit does not need to be established, and correction of the weighting coefficient can be realized only by using an original beam calibration system, so that homogenization of each beam gain in a field range of the phased array receiver is met, and convenience is provided for the phased array receiver to realize high-resolution imaging in radio astronomical observation.

In accordance with a first aspect of the present disclosure, a radar system is provided, comprising: a plurality of receive channels; a first radar detector stage configured to detect, through said receive channels, at least one target using a detector threshold; a noise correlation determination unit configured to determine a noise correlation level indicative of noise correlation between the receive channels if the first radar detector stage has detected the target; a detector threshold adjustment unit configured to adjust the detector threshold in dependence on the noise correlation level determined by the noise correlation determination unit; a second radar detector stage configured to detect the target using the adjusted detector threshold. In accordance with a second aspect of the present disclosure, a corresponding method of operating a radar system is conceived.

The application provides a test method, device and equipment for the correlation between vehicle air tightness and in-vehicle noise. For a to-be-tested vehicle placed in a whole vehicle semi-anechoic laboratory, different air tightness states of the to-be-tested vehicle are set. A simulated sound source is played at a preset position outside the to-be-tested vehicle to simulate the noise in the whole vehicle semi-anechoic laboratory during the running of the to-be-tested vehicle, and air tightness data and in-vehicle noise information of the to-be-tested vehicle in different air tightness states are obtained. Based on the obtained air tightness data and in-vehicle noise information of the to-be-tested vehicle in different air tightness states, a mapping relationship between different air tightness data and in-vehicle noise information of the to-be-tested vehicle is determined. In this way, the efficiency of obtaining air tightness data and noise data and the accuracy of determining the correlation between air tightness data and noise data can be improved.

The invention discloses a semi-automatic test method for an energy storage prefabricated cabin, and belongs to the technical field of intelligent testing, and the method comprises the steps of test system construction, semi-automatic testing, multi-physics coupling test analysis and test report generation. According to the scheme, a semi-automatic test process is constructed, automatic acquisition and manual verification are combined, a load sequence and an acquisition instruction are preset in an automatic stage, three-dimensional temperature gradient distribution is generated, a noise spectrum and heat dissipation efficiency are weighted A, and fan rotating speed strategy adjustment, noise source positioning and abnormal data labeling are performed in a manual verification stage. A data foundation is laid for multi-physics field analysis; dynamically adjusting noise confidence and sample weight, generating an adversarial fusion threshold, adaptively adjusting the base classifier weight and the sample weight, establishing a multi-physics field coupling test model, outputting a heat dissipation performance level, a noise compliance judgment result, a coupling risk level and a heat dissipation-noise correlation index; and decision support is provided for optimization and safe operation of the prefabricated cabin.

The invention discloses a compressor low-noise operation rotating speed calibration method and device and a storage medium, and relates to the technical field of compressor control, the compressor low-noise operation rotating speed calibration method comprises the steps that rotating speed acceleration noise associated data is acquired, and a to-be-evaded rotating speed threshold value is determined according to the rotating speed acceleration noise associated data; whether the whole vehicle meets a static condition or a dynamic driving condition or not is determined, the vehicle meeting the static condition is in a static state, and the vehicle meeting the dynamic driving condition is in a driving state; when the whole vehicle meets the static condition, compressor noise data and compressor vibration data are determined; when the whole vehicle meets the dynamic driving condition, environment noise data and environment vibration data are determined; and according to the compressor noise data, the compressor vibration data, the environment noise data, the environment vibration data and the to-be-evaded rotating speed threshold value, data matching is conducted, and the target rotating speed of the compressor is obtained. The effect of accurately calibrating the low-noise operation rotating speed of the compressor is achieved.

The application provides a graph matching method for noise correlation, relates to the technical field of image matching, and comprises the following steps: initializing parameters of an online network and a momentum network, obtaining an input image and key points of the online network, inputting the online network to obtain key point representation of the online network, updating parameters of the momentum network, inputting the momentum network to obtain key point representation of the momentum network, calculating back propagation loss based on the key point representation of the online network and the key point representation of the momentum network to optimize the online network, repeatedly iterating until the online network converges to obtain node representation of the online network, obtaining a similarity matrix between nodes through the node representation of the online network, and calculating the similarity matrix through a Hungarian algorithm to obtain a key point matching matrix as a final matching result; the point alignment and edge alignment loss of the graph matching are calculated, the robustness of an image matching model to real noise correlation data is improved, and the accuracy of image matching is significantly improved.

The application discloses a computer room monitoring method and system based on a multi-source data fusion intelligent inspection robot, belongs to the technical field of automatic computer room monitoring, and comprises application of adversarial transfer learning on a four-dimensional fault semantic feature field, generation of a cross-modal causal graph atlas representing a fault evolution path through a graph neural network, and the like, wherein the application aligns the feature distribution of a standard computer room and a current computer room through a joint loss function, forcibly aligns the feature distribution of a source domain and a target domain through a gradient reversal layer, generates a cross-modal causal graph atlas in combination with an attention mechanism and a causal strength weight, the graph neural network further models the physical connection, functional dependence and time sequence correlation between nodes, encodes the causal relationship as an edge weight, filters noise correlation through a causal mask, and improves the stability of the causal graph atlas, and the cross-modal causal graph atlas can accurately capture a fault propagation path.

The invention provides an electric power marketing data cleaning and deduplication method and system, and the method comprises the steps: calculating the neighborhood density and local attribute fluctuation of a data point, and marking the data point as an initial noise point if the neighborhood density is lower than a threshold value and the fluctuation is higher than the threshold value; otherwise, dividing into core points or boundary points according to density; setting an asymmetric expansion neighborhood for the boundary point, calculating the sum of distance attenuation function values from an initial noise point to the boundary point in the neighborhood, and if the sum is higher than a first threshold value, re-classifying as a confirmation noise point; if the initial noise points are located in the overlapped neighborhood and the geometric mean value of the noise correlation indexes of the group of boundary points is lower than a second threshold value, the noise points are classified into secondary boundary points, and the secondary boundary points are classified into a nearest core point cluster; noise points are isolated and confirmed, and the stability is calculated in the cluster; and correcting the distance from the data point to the cluster centroid according to the stability, selecting the minimum corrected distance as a main data record, and combining redundant records.

This control device (30) for a glass diaphragm is provided with: an acquisition unit (62) that acquires noise-related information relating to noise flowing into the interior (12) of a vehicle (10); and a control unit (64) that controls the generation of superimposed sound from the glass diaphragm (1), said superimposed sound being superimposed on the noise, on the basis of the noise-related information.

The application discloses a gear fault diagnosis method based on sub-signal joint weighting envelope anti-noise correlation. Firstly, the original vibration signal sequence is converted into an envelope signal through a signal sequence element-by-element squaring-low-pass filtering-square root calculation process, then the envelope signal is reconstructed according to different time intervals to obtain a series of sub-signals, and the fault information representation of each sub-signal is calculated based on the L-matrix theory index, then the Sigmoid function is combined to assign a weight to each sub-signal, and then the joint weighting envelope anti-noise correlation function of the envelope signal sequence and the reconstructed sub-signals is calculated based on the envelope signal, the reconstructed sub-signals and the corresponding weights, finally, the characteristic frequency is determined according to the reciprocal of the time interval value corresponding to the characteristic peak in the change graph of the joint weighting envelope anti-noise correlation function with the time interval, and finally the gear fault is identified. The application can reliably identify the gear fault based on time domain analysis only under the condition of limited signal length and complex noise interference.

The present application belongs to the field of oil liquid abrasive particle signal detection, and particularly relates to a kind of abrasive particle signal extraction method based on noise correlation, including installing inductive abrasive particle detection sensor, collection detection signal and reference signal;Detection signal and reference signal are filtered to obtain detection low frequency harmonic and reference low frequency harmonic, and are divided into L detection wave and L reference wave, the waveform similarity of detection wave and its corresponding reference wave is calculated;Waveform similarity not greater than waveform similarity threshold is screened out, and its corresponding fragment position is marked as suspected abrasive particle signal fragment position;Detection signal and reference signal are subtracted to carry out low-pass filtering and obtain preliminary noise reduction signal;And according to suspected abrasive particle signal fragment position, preprocessing fragment is extracted;The numerical characteristics of preprocessing fragment are extracted, and ideal abrasive particle signal is established;The fragment matching degree is calculated and threshold is set, the fragment below threshold is excluded, and the extraction and counting of abrasive particle signal are completed;The present application can stably segment abrasive particle signal.

This disclosure relates to a method and related apparatus for determining time intervals. The method includes determining a first time interval for the current moment based on a timestamp recorded during communication with a sensor, and determining a second time interval for the current moment based on time intervals from historical moments. The method further includes determining the time interval for the current moment based on the first time interval, the second time interval, and noise correlation coefficients between the first time interval and the noise of the second time interval. According to embodiments of this disclosure, by adjusting the noise correlation coefficients of the measurement time interval and the estimated time interval, the weights of the measurement time interval and the estimated time interval relative to the final time interval can be changed, avoiding a fixed time interval and thereby improving the accuracy of the final time interval.

The invention relates to a low-noise time quantization circuit suitable for a laser radar focal plane detector, and belongs to the technical field of reading circuits of laser radar focal plane detectors. Compared with a traditional TAC architecture, the TAC architecture is based on a correlated double sampling principle, utilizes noise correlation between a reset signal and an effective signal in the same pixel readout circuit, caches a reset voltage and an effective signal voltage by using a pre-storage circuit in the pixel readout unit, and reads the reset voltage and the effective signal voltage to the correlated double sampling circuit through a column bus for subtraction; the influence of 1 / f noise, KCT noise, array fixed mode noise and other noise is greatly reduced, and the signal-to-noise ratio of focal plane imaging is improved.

The invention discloses a noise robustness auditory decoding method and system based on an adversarial domain adaptation network, and relates to the technical field of auditory decoding, and the method comprises the steps: obtaining a multi-channel electroencephalogram signal, and carrying out the parallel analysis and feature fusion; inputting the multi-dimensional signal integrity feature vector into a noise evaluation model, and outputting a noise complexity score; calling a feature decoupling module, and outputting a task-related feature vector and a noise-related feature vector; inputting the task related feature vectors into an attention decoder to obtain auditory attention distribution probability distribution; and an auditory attention distribution probability distribution and noise traceability and influence report is used as a noise robustness auditory decoding result of the target user. The technical problems that an existing auditory decoding method is low in decoding precision and insufficient in robustness in a noisy scene are solved.

This invention discloses a broadband array gain prediction method based on acoustic field spatial correlation correction. This method utilizes marine environmental data for acoustic field modeling, calculates the channel frequency response function from the target to each array element position, estimates the correlation coefficient of the received signal after channel distortion, and forms a signal correlation coefficient matrix. Furthermore, it uses an isotropic noise field model to estimate the inter-channel noise correlation coefficient after the processed frequency band deviates from the ideal frequency, forming a noise correlation coefficient matrix. Finally, it uses the two correlation coefficient matrices to calculate the array gain. This method can effectively reflect the coupling between the underwater acoustic array and the actual marine environment during operation, and can improve the accuracy of the array gain calculation results.

Embodiment of the disclosure relates to a method, apparatus, device and storage medium for determining a noise level. The method includes: determining, for each of a plurality of audio segments, feature values that each corresponds to a respective one of a plurality of predetermined audio features based on spectrum information and energy spectrum information corresponding to the plurality of audio segments in an audio; determining whether each of the plurality of audio segments is a speech segment based on whether feature values satisfy a first feature requirement related to speech; determining whether each of the plurality of audio segments is a noise segment comprising a predetermined noise based on whether feature values satisfy a second feature requirement related to the predetermined noise; and determining a noise level of the audio based on an audio segment other than the speech segment and the noise segment among the plurality of audio segments.

The invention discloses a multi-label chronic disease prediction method, device and system. The method comprises the following steps: acquiring electronic health record data of a plurality of to-be-predicted patients, and constructing corresponding medical history feature representations and patient feature representations; calculating the medical history similarity between the to-be-predicted objects according to the medical history feature representation, and constructing a patient similarity relation graph based on a similarity threshold value and a neighbor screening strategy; performing feature construction on edges in the patient similarity relation graph, and screening and filtering the edges through an edge classification model to obtain an optimized patient relation graph; and inputting the optimized patient relation graph and the corresponding patient features into a graph neural network model, and outputting a multi-label chronic disease prediction result corresponding to the to-be-predicted object. Through the patient relation modeling and edge filtering mechanism based on the medical history similarity, the influence of noise correlation on the prediction result is effectively reduced, and the accuracy and stability of multi-label chronic disease prediction are improved.

Image quality can be assessed in a patient-specific manner using deep learning. Deep learning is used to extract image quality metrics (e.g. noise metrics, resolution metrics, noise correlation metrics) directly from image-based features, thereby enabling patient-specific image quality assessment. A neural network, or other suitable machine learning algorithm, is trained and implemented to generate noise maps from patient CT data, where the noise maps quantify the pixelwise noise level in the patient CT data.

This invention discloses an aerial target recognition method based on bootstrapping multimodal learning, comprising the following steps: S1, constructing a training dataset containing multimodal aerial target samples and their category labels, and setting a model structure for feature extraction, target classification, and modality reliability evaluation for each modality; S2, performing bootstrapping enhancement on some samples during training, generating modal perturbations and combination schemes for the samples based on noise correlation constraints, forming an enhanced training set for revealing supervision; S3, extracting the representation of each modality and obtaining classification outputs, while characterizing the prediction differences between modalities and the uncertainties within modalities, generating a reliability score characterizing the credibility of the modality; S4, jointly optimizing the classification learning objective and the reliability constraint objective to suppress the interference of noise correlation on the learning process and update the model parameters; S5, extracting the features of each modality and obtaining the corresponding reliability scores during the testing phase, performing reliability-weighted fusion of the classification outputs of each modality, and outputting the final aerial target recognition result; This invention can adaptively evaluate the reliability of each modality and perform weighted fusion of the classification results even when there is noise correlation in the test data, thereby improving the robustness and recognition accuracy of aerial target recognition.

The embodiment of the invention provides a method, a device and equipment for establishing a noise correlation model of a wind generating set and a medium, and relates to the technical field of wind power generation. The method comprises the following steps: establishing a site noise attenuation model; based on complete machine noise tests in at least two different acoustic states, noise contributions corresponding to blades and a cabin in the wind generating set are separated; based on the noise contribution of the cabin, testing at least one noise source part in the unit and testing the acoustic characteristics of the cabin structure, and calculating the overall external radiation noise of the cabin; integrating the site noise attenuation model, the noise contribution of the blade and the overall external radiation noise of the cabin, and establishing a correlation model; the method provided by the invention effectively solves the problems that the noise source of the wind generating set is difficult to identify and the effective noise reduction processing is difficult to carry out due to the lack of a quantitative model for correlating the components and the overall noise in the wind generating set in the related technology.

The application discloses a differential amplifier noise optimization method and device, a storage medium and an electronic equipment. The differential amplifier noise optimization method comprises the following steps: acquiring circuit topological parameters, device process parameters and noise correlation performance indexes of a differential amplifier; based on the Lie algebra principle, a two-dimensional special unitary group generator matrix corresponding to a two-terminal noise mode of the differential amplifier is constructed by using the circuit topological parameters and the device process parameters; a noise correlation operator based on the two-dimensional special unitary group generator matrix is established; the noise correlation operator and the noise correlation performance indexes are input into a group theory optimization solver to determine a group of target optimization circuit parameters; and the differential amplifier is adjusted according to the target optimization circuit parameters, so that the actual noise correlation performance of the adjusted differential amplifier meets the noise correlation performance indexes. The application can effectively eliminate the correlated noise and improve the robustness to non-ideal factors such as process and temperature.

The invention relates to a horizontal line array deep sea multi-path target correlation and elimination method, which comprises the following steps: step 1, performing statistics to obtain a noise correlation coefficient by using pre-formed beam data under a horizontal line array pure background noise condition, and performing statistics to obtain a noise correlation coefficient by using multi-path signal pre-formed beam data under different signal-to-noise ratio conditions of the horizontal line array; performing statistics to obtain correlation coefficients under different signal-to-noise ratio conditions, and determining a multi-path correlation discrimination threshold through clustering analysis; and step 2, according to the sound velocity profile, the sea depth, the submarine topography, the receiving depth and the target distance depth interval related channel parameters and the propagation model, obtaining main path arrival pitch angles of different distances. The method is suitable for solving the problems of multi-path signal association and multi-path false target elimination of horizontal line array deep sea passive detection, and can provide technical support for subsequent deep sea passive warning tracking, passive distance measurement, target identification and the like of a horizontal line array.

The application discloses a noise robustness hearing decoding method and system based on an adversarial domain adaptation network, relates to the technical field of hearing decoding, and comprises the following steps: acquiring a multi-channel electroencephalogram signal, and performing parallel analysis and feature fusion; inputting a multi-dimensional signal integrity feature vector into a noise evaluation model to output a noise complexity score; calling a feature decoupling module to output a task-related feature vector and a noise-related feature vector; inputting the task-related feature vector into an attention decoder to obtain a hearing attention allocation probability distribution; and taking the hearing attention allocation probability distribution and a noise tracing and influence report as a noise robustness hearing decoding result of a target user. The technical problem of low decoding accuracy and insufficient robustness of existing hearing decoding methods in a noisy scene is solved.

The application provides a power marketing data cleaning and deduplication method and system, wherein the neighborhood density and local attribute fluctuation degree of data points are calculated, if the neighborhood density is lower than a threshold value and the fluctuation degree is higher than a threshold value, the initial noise point is marked; otherwise, the density is divided into core points or boundary points; the asymmetric extended neighborhood is set for the boundary points, the sum of distance decay function values from the initial noise point to the boundary points in the neighborhood is calculated, if the total sum is higher than a first threshold value, the confirmed noise point is reclassified; if the initial noise point is located in the overlapping neighborhood and the geometric mean value of noise correlation indexes of the group of boundary points is lower than a second threshold value, the noise point is reclassified as a secondary boundary point and is classified into the nearest core point cluster; the confirmed noise point is isolated, the stability in the cluster is calculated; the distance from the data point to the cluster center is corrected according to the stability, the main data record with the minimum corrected distance is selected, and the redundant records are combined.