14 results about "Wiener filter" patented technology

The application discloses a TDI-CCD image deartifact method and device, equipment, medium and product, and relates to the technical field of image processing. The method comprises the following steps: constructing a deep deblurring network; the deep deblurring network comprises a Wiener filter module and a generative adversarial network, and the Wiener filter module is arranged at the input end of a generator of the generative adversarial network; the Wiener filter module is used for filtering a frequency domain image; the deep deblurring network is trained, and the trained deep deblurring network is used as a deartifact model; during the training of the deep deblurring network, a point spread function and a regularization parameter of the Wiener filter module are used as learnable network parameters; and an image with row smear obtained from a TDI-CCD camera is processed by using the deartifact model to obtain a deartifact image. The application can effectively remove image artifacts and improve the definition of a TDI-CCD image.

The application relates to a cloud-edge collaborative computing framework and processing method for multi-modal data stream fusion processing, which comprises the following steps: S1: based on the original data stream collected by the edge computing node, an improved Wiener filtering algorithm is used for noise suppression, signal denoising is realized through adaptive threshold wavelet transform, and a time stamp alignment technology is used to solve the time delay difference problem of multi-modal data, and a space-time aligned and purified data stream is generated. The application has the effects that through the combination of the improved Wiener filtering algorithm and the adaptive threshold wavelet transform, the noise suppression efficiency of the original data stream is significantly improved, the time stamp alignment technology effectively solves the time delay difference of multi-modal data, the generation of the space-time aligned and purified data stream ensures that the subsequent processing has a unified time sequence reference, and the space-time attention fusion network adopts the collaborative architecture of a bidirectional gate recurrent unit and a lightweight 3D convolution network.

The application discloses a method for improving voice conference call quality, relates to the technical field of computer voice processing, and comprises the following steps: using an Escape-TDNN model, adopting diversity samples and double noise reduction technology; using a multi-index fusion algorithm, providing personalized audio services, using adaptive beam forming technology and Wiener filtering algorithm, adopting a human voice separation model, combining with a voiceprint recognition result for post-processing optimization, using natural language processing technology, combining with migration learning technology, using voice synthesis and post-processing technology, fusing microphone array space information and video visual features, combining with environmental prior information, introducing an attention mechanism, and supporting online update of the human voice separation model. The application separates background sound by using an algorithm, isolates noise, makes voice clear, introduces text-to-speech, copes with situations where voice cannot be emitted or a microphone fails, accurately identifies and separates human voice by means of voiceprint technology, improves recognition and purity, and enhances audio effect.

The application discloses a photoacoustic signal enhancement method and device based on adaptive multi-band spectral subtraction and Wiener filtering, and belongs to the technical field of speech signal enhancement. The application divides a signal into sub-bands according to Mel scale, determines a subtraction factor and a lower limit factor by a monotone decreasing function linkage according to real-time signal-to-noise ratio of each sub-band, and makes the two factors negatively correlated with the signal-to-noise ratio, so that the denoising strength and the spectral bottom filling depth are automatically adapted; the weighted moving average of the power spectrum of adjacent sub-bands is carried out to smooth isolated spectral peaks and spectral valleys left by spectral subtraction; a residual amplitude limiting is introduced before Wiener filtering, the maximum noise residual threshold is determined based on adjacent frame statistics and limiting, and Wiener filtering is carried out by taking the data after limiting as clean signal estimation; the noise spectrum is recursively updated by a forgetting factor which is dynamically adjusted according to the signal-to-noise ratio, and is only executed in the speech inactive section. The application has an output signal-to-noise ratio improvement of more than 50% when the input is 0dB, can stably enhance without damage under high signal-to-noise ratio, and is suitable for photoacoustic speech acquisition and enhancement scenes.

A method (1000) for calibration of an antenna array is provided. The method comprises deriving (1100) at least one Wiener filter coefficient for the antenna array based on at least one power delay profile characteristic related to the antenna array and at least one signal-to-noise ratio related to the antenna array. The method further comprises estimating (1200) at least one transmitter parameter and / or at least one receiver parameter using the at least one Wiener filter coefficient for the antenna array, and calibrating (1300) the antenna array based on the at least one transmitter parameter and / or the at least one receiver parameter.

This invention discloses a method and system for estimating the time of bloodstain traces in judicial practice, relating to the fields of forensic identification and physical evidence technology. The invention acquires spatial distribution AC magnetic susceptibility tensor data of bloodstains and pure substrate magnetic signal array data; constructs a background noise covariance matrix and inputs it into an adaptive Wiener filter model to perform deconvolution separation, generating pure bloodstain intrinsic magnetic susceptibility spectrum data stripped of background interference; performs nonlinear fitting processing on the spectrum data using the fractional-order Haveryak-Nicaraguan relaxation equation, analytically generating relaxation dynamics feature set data; maps this data to a high-dimensional space of a Riemannian manifold to perform geodesic distance optimization calculation, outputting the bloodstain retention time corresponding to the minimum distance. This invention achieves non-destructive, non-contact detection, completely eliminates magnetic field interference from complex substrates, overcomes the measurement distortion of high-dimensional biochemical features, and significantly improves the accuracy and robustness of time estimation.

The application relates to an anti-mainlobe interference adaptive monopulse angle measurement method for a uniform linear array, which comprises the following steps: obtaining array echo data and calculating a covariance matrix; estimating the number of signal sources and the mainlobe interference angle by using the covariance matrix; constructing a mainlobe interference blocking matrix by using the mainlobe interference angle, and preprocessing the array echo data to obtain preprocessed array data; projecting a corresponding static steering vector of a beam direction to a signal subspace to obtain a sum beam projection vector; performing Wiener filtering on the preprocessed array data by using the sum beam projection vector to solve sum beam adaptive weights; calculating a difference beam projection vector by using the sum beam projection vector, and solving difference beam adaptive weights by using the difference beam projection vector; and calculating a target deflection angle by using the sum beam adaptive weights and the difference beam adaptive weights and adopting an amplitude comparison monopulse method. The method has the advantages of high anti-mainlobe interference capability and high angle measurement precision of a target near the mainlobe interference.

This application belongs to the field of radar detection technology, specifically disclosing a radar anti-slicing interference method based on time-domain concealment and Wiener filtering recovery. The method includes: receiving a mixed signal determined based on the radar transmitted signal; determining a state judgment threshold based on the amplitude of the mixed signal; determining the echo state of the mixed signal at each time point based on the state judgment threshold to generate a state sequence; using weighting coefficients to perform weighted cancellation on the upper and lower branch signals to obtain the residual signal; based on the residual signal, using a sliding window method to determine a target range threshold to determine an estimated signal; and reconstructing the signal based on the estimated signal to obtain the recovered signal; determining the average amplitude of all non-zero points in the upper branch signal to determine a shutdown criterion coefficient; if the shutdown criterion coefficient does not exceed a preset energy discrimination threshold, then the recovered signal is used as the anti-interference output signal. This application can improve interference suppression efficiency and reduce costs.

The application discloses a kind of based on tri-state classification voice activity detection's self-adapting wiener filtering noise reduction method and system and storage medium, for solving the technical problem that cochlear implant internal speech noise reduction algorithm is not good under non-stationary noise environment The effect of reducing noise is not good.The application first by improved voice activity detection audio frame is divided into speech frame, silence frame and noise frame three categories, and minimum value tracking in time-frequency domain is carried out to realize the power spectrum estimation of noise.The system further introduces noise environment change detection mechanism, dynamically assesses noise stability, and adjusts wiener filter parameter accordingly.Aiming at different types of frame and frequency point, the system dynamically optimizes filter parameter, including smoothing coefficient, gain lower limit and frequency band weight.Through this multi-level adaptive processing mechanism, the method can realize higher speech signal-to-noise ratio and clarity under complex noise environment, suitable for cochlear implant single microphone noise reduction.

This invention discloses a method and system for denoising mobile communication data based on big data processing, relating to the field of big data processing technology. The method includes: S1, acquiring raw mobile communication signal data at the edge layer; S2, performing Wiener filtering denoising on low-noise data to obtain a denoising result; S3, uploading high-noise data to the edge computing layer and performing improved adaptive wavelet threshold denoising to obtain a first-level denoising result; S4, inputting the first-level denoising result from S3 into a lightweight convolutional recurrent neural network model based on knowledge distillation at the edge computing layer; and S5, uploading the denoising result data to the cloud-based big data computing and management layer, iteratively optimizing the lightweight denoising model of each edge node through a federated incremental learning mechanism. This method and system for denoising mobile communication data based on big data processing achieves efficient denoising processing of mobile communication big data through a three-layer collaborative computing architecture of edge-cloud.

The application provides a crack detection method based on automatic filling pseudo labels, a lightweight target detector is trained by using data samples to obtain network model weights, a new collected image is subjected to one-stage crack detection by using the trained network model weights, and suboptimal prediction samples are screened as candidate regions for two-stage detection, for the suboptimal prediction samples, the image recognition degree is improved by using a Wiener filter and a morphological method, a Canny edge detection operator is used to perform two-stage crack detection on the cracks, the prediction results of the two stages are comprehensively considered, and finally, the crack detection is realized; in addition, high confidence samples of the one-stage prediction results are screened as pseudo labels, the confidence of the suboptimal prediction samples is improved according to the two-stage optimized prediction results, the pseudo label samples are filled, and the network model is retrained by using label data and pseudo label data. The application has high robustness and strong generalization, and is more suitable for a target detection task.