34 results about "Linear filter" patented technology

The invention provides an OBD message analysis and attitude rendering method and system based on a vehicle dynamics model, and belongs to the field of vehicle electronics and control. The method comprises the steps that a message is analyzed through an OBD interface, and vehicle operation parameters are obtained; establishing a unified time base and carrying out resampling and synchronous processing on the data of each channel; constructing a fifteen-degree-of-freedom vehicle dynamics model comprising a vehicle body six-degree-of-freedom pose, a vehicle body angular velocity, a four-wheel rotation angular velocity and unsprung mass vertical displacement for state prediction; designing an observation function to map a model state to an OBD observation space, and adopting an observability enhancement strategy; predicting and observing data are fused by using a nonlinear filter, and abnormal observation gating is carried out based on a mahalanobis distance and an amplitude threshold value; compensating system time delay and adaptively adjusting observation noise; and performing three-dimensional real-time rendering on the attitude information. According to the method, high-precision and high-robustness vehicle attitude estimation can be realized only by depending on the OBD message, and the system cost is effectively reduced.

The application relates to an audio signal enhancement method and device, computer equipment and a storage medium. The method comprises the following steps: when a voice packet is received, sequentially decoding and filtering the voice packet to obtain an audio signal; when the audio signal is a forward error correction frame signal, extracting a feature parameter from the audio signal; converting the audio signal into a filter speech excitation signal based on a linear filtering parameter obtained by decoding the voice packet; performing speech enhancement processing on the filter speech excitation signal according to the feature parameter, a long-time filtering parameter and a linear filtering parameter obtained by decoding the voice packet, to obtain an enhanced speech excitation signal; and performing speech synthesis based on the enhanced speech excitation signal and the linear filtering parameter to obtain a speech enhancement signal. The method can improve the timeliness of audio signal enhancement.

The application provides a method and device for intra-channel and inter-channel nonlinear joint compensation of a WDM system based on an improved LDBP and electronic equipment, and belongs to the technical field of optical fiber communication. The method compensates dispersion by using a linear filter in the time domain, the corresponding network structure is designed as one-dimensional convolution, and the pulse broadening effect is combined in the overlap reservation method, the physical model of nonlinear compensation is improved by considering the interference of intra-channel dispersion pulse broadening and the walk-off effect between channels, and the purpose of intra-channel and inter-channel nonlinear joint compensation is achieved. The application fully utilizes the random interaction between inherent dispersion and nonlinearity in the optical fiber communication system to compensate the phase shift caused by the nonlinear effect, can effectively equalize the nonlinear distortion of the signal under the condition of controlling the calculation complexity and reducing the implementation cost, and is expected to play a better role in the long-distance transmission of the large-capacity WDM system.

The invention discloses a tobacco online near infrared spectrum detection device and method based on a linear optical filter, and relates to the technical field of spectrum processing. Near-infrared light is emitted to a target area, a plurality of near-infrared bands are obtained through a linear optical filter, and reflected light of each near-infrared band is collected to obtain an original near-infrared band set; preprocessing the target original near-infrared band to obtain a target near-infrared band; performing feature extraction on the target near-infrared band to obtain target sub-features; and constructing a target feature vector according to all the target sub-features, and searching a preset database according to the target feature vector to obtain a detection result. Multiple near-infrared bands and frequency-combining and frequency-doubling absorption information of substance molecules corresponding to different bands are obtained through the linear optical filter, and then the different near-infrared bands are preprocessed and then subjected to feature extraction, so that prediction of different chemical substance components in tobacco is improved, and the online detection efficiency of the tobacco is improved.

The invention particularly relates to a sea clutter pulse interference suppression method for cascaded Hammerstein robust filtering, and the method comprises the steps: receiving an echo signal, and carrying out the preprocessing of the echo signal, so as to obtain a direct signal and a sea clutter signal, which are separated from each other; a cascaded Hammerstein filter is used to carry out multi-stage filtering processing on the separation signal, and a non-linear mapping output signal is obtained; wherein the cascaded Hammerstein filter comprises a first nonlinear filtering module, a second linear filtering module and a third nonlinear filtering module which are arranged in a cascaded manner; and carrying out filtering processing on the nonlinear mapping output signal through a robust filter to obtain an output signal after interference suppression. According to the method, nonlinear distortion and non-Gaussian pulse interference can be effectively suppressed in a complex marine environment, and the anti-interference capability, the signal fidelity and the adaptivity of a radar and a communication system are improved.

Multiple successively spatially downsampled versions of a rendered image frame are generated for at least one two-dimensional signal component of one or more two-dimensional signal components of an im

The present application relates to a kind of back wave field synthesis processing method, device, computer equipment and storage medium, method includes the following steps: respectively obtaining source back wave data and source wavelet data;According to source back wave data and source wavelet data, determine sparse linear filter coefficient;Using the numerical value of non-zero sparse linear filter coefficient and delay, time delay weighted superposition of forward wave field, synthesis back wave field.The above-mentioned back wave field synthesis processing method, based on source back data is the convolution of source wavelet data and sparse linear filter, determine sparse linear filter coefficient by source back data and source wavelet data, then according to the principle of superposition, time delay weighted superposition of forward wave field, synthesis back wave field.No need to rely on back data to obtain back wave field by solving wave equation, power is greatly reduced, can improve processing efficiency, reduce processing time, can also reduce memory consumption, so as to reduce power cost.

This disclosure relates to the field of seismic imaging technology, and particularly to a method, apparatus, device, and medium for generating common imaging point gathers for wave equations. The method includes: identifying a source wavelet based on a selected source location; synthesizing the return wavefield of a single-channel seismic record corresponding to a geophone based on the forward propagation wavefield of the source wavelet and a pre-determined linear filter; synthesizing a single-channel seismic image corresponding to the single-channel seismic record using the return wavefield and the forward propagation wavefield; calculating the offset distance corresponding to each single-channel seismic image; classifying the single-channel seismic images according to the offset distance; and superimposing the classified single-channel seismic images to obtain an offset common imaging point gather; performing amplitude processing on the offset common imaging point gathers; and determining the amplitude-processed offset imaging point gathers as the common imaging point gathers for wave equations. This disclosure can improve the efficiency of common imaging point gather generation.

The application discloses a remote sensing image target detection method based on an RFTZN neural dynamics model, and belongs to the technical field of remote sensing image target detection, which comprises the following steps: based on a to-be-detected remote sensing image and a detection target feature, a target detection task of the remote sensing image is constructed as a CEM optimization problem, and the CEM optimization problem is converted into an equivalent linear matrix equation; an RFTZN neural dynamics model is constructed; based on the RFTZN neural dynamics model, a solver is constructed, and the linear matrix equation is solved to obtain a linear filter; and the linear filter is applied to the to-be-detected remote sensing image to extract a detection target. The RFTZN neural dynamics model is constructed by innovatively introducing a feedback regulation mechanism and a dynamic self-adaptive scale factor, so that the solving speed and the anti-noise interference capability of the RFTZN neural dynamics model are far superior to those of a traditional model, and the dynamic optimization problem in a target detection process can be efficiently solved.

The embodiment of the invention provides a satellite receiver signal processing method and device, and the method comprises the steps: carrying out the demodulation processing of a received signal, and obtaining a baseband signal; performing correlation calculation processing on the baseband signal and a local correlation sequence to obtain a correlation value; constructing a signal space based on the correlation value; performing linear filtering processing on the correlation value of the signal space to obtain a linearly filtered signal space; based on the signal space after linear filtering, the maximum correlation value under the same Doppler frequency and different code phase delays in the first dimension is searched, the sum of the distances between the maximum correlation value and the correlation values under the same Doppler frequency and other code phase delays is calculated, and the maximum distance is selected from the sum of the distances corresponding to the Doppler frequencies; performing nonlinear filtering processing on the correlation value of the Doppler frequency corresponding to the maximum distance to obtain a correlation value after nonlinear filtering; and determining whether a target signal exists based on the correlation value after nonlinear filtering. The performance of capturing the target signal can be improved.

The application discloses a multi-frame filtering amplitude-phase decoupling deep neural network stereo echo cancellation method and system based on human ear hearing. The method comprises the following steps: using a microphone to receive a signal and two-channel reference signal to estimate the real part and the imaginary part of a linear filter, and using a multi-frame filter structure to estimate a linear echo signal in real time to obtain a residual echo complex spectrum; estimating a near-end speech amplitude spectrum, training a neural network for estimating the near-end speech amplitude spectrum through a human ear hearing related cost function; performing secondary estimation on the near-end speech complex spectrum, and training the neural network through the human ear hearing related cost function; and performing stereo echo cancellation through the trained neural network. The application decouples the mapping of the near-end speech complex spectrum in the stereo echo cancellation system into amplitude spectrum mapping and phase mapping, and finely recovers the near-end speech complex spectrum through secondary estimation, thereby improving the quality and intelligibility of the near-end speech.

The invention provides an IMU gyroscope optimization method and system based on a neural network, and relates to the technical field of IMU gyroscope optimization. The method comprises the following steps: firstly, acquiring a time sequence signal of the gyroscope, and processing the time sequence signal by utilizing linear filters with different bandwidths to obtain a multi-resolution signal branch; then, inputting each signal branch into an adaptive expansion convolutional network, generating a feature representation set, and calculating a cross-resolution attention weight; next, fusing the branch features by using the weights to obtain a fused feature set, inputting the fused feature set into a neural network to be optimized, and outputting the denoised angular velocity, bias estimation and attitude increment; and finally, constructing a multiple loss function based on the optimization parameters to carry out network training, and obtaining optimized angular velocity and attitude estimation by using actual measurement signals. According to the method, the problem that an IMU output signal is unstable due to the fact that attitude errors caused by time-varying bias and noise are difficult to effectively suppress in the prior art is solved.

A coder (encoder or decoder) determines, for a current block (CB), a reference block (RB) based on a template matching (TM) cost associated with a first reference template of the RB and a first current template of the CB. The coder further determines, based on the first reference template matching the first current template being flipped in a direction, coefficients of a linear spatial filter based on filtered values, resulting from the linear spatial filter applied to first samples of a second reference template of the RB, being set to corresponding values of second samples of a second current template flipped in the direction. An adjusted RB is generated by applying the linear spatial filter with the determined coefficients to the RB. The coder codes the CB based on a residual of the CB determined based on the adjusted RB.

Devices, systems and methods for temporal prediction of parameters in non-linear adaptive loop filtering are described. In an exemplary aspect, a method for visual media processing includes configuring, for a current video block, one or more parameters of a clipping operation that is part of a non-linear filtering operation; and performing, based on the one or more parameters, a conversion between the current video block and a bitstream representation of the current video block, wherein the one or more parameters is coded in accordance with a rule.

This invention discloses a broadband linear graded filter for spectral imaging, possessing high transmittance and excellent stray light suppression performance. The filter employs a multilayer thin-film structure, including an interference cutoff film system above a substrate and a narrowband filter film system below. The narrowband filter film system achieves high transmittance in the 500-800 nm wavelength range through an optimized design using multiple alternating combinations of high and low refractive index materials. The interference cutoff film system effectively suppresses stray light through its long-pass and short-pass design.

The invention provides a method, a device and a system for audio signal processing. The present implementation more specifically relates to speech enhancement techniques that combine statistical signal processing with neural network inference. In some aspects, a speech enhancement system may include a linear filter, a deep neural network (DNN), and a non-linear post filter. The linear filter and the non-linear post filter are configured to suppress noise in an audio signal using statistical signal processing techniques. More specifically, the linear filter de-noising an input audio signal based on a temporal correlation between consecutive frames of the audio signal. The DNN inferes a speech signal and a noise signal (respectively representing speech and noise components of the audio signal) based on the de-noised audio signal. The non-linear post filter suppresses residual noise in the speech signal based on one or more Gaussian mixture models (GMMs) associated with the speech signal and the noise signal.

This disclosure provides methods, devices, and systems for audio signal processing. The present implementations more specifically relate to speech enhancement techniques that combine statistical signal processing with neural network inferencing. In some aspects, a speech enhancement system may include a linear filter, a deep neural network (DNN), and a nonlinear post-filter. The linear filter and the nonlinear post-filter are configured to suppress noise in audio signals using statistical signal processing techniques. More specifically, the linear filter denoises an input audio signal based on a temporal correlation between successive frames of the audio signal. The DNN infers a speech signal and a noise signal (representing a speech component and a noise component, respectively, of the audio signal) based on the denoised audio signal. The nonlinear post-filter suppresses residual noise in the speech signal based on one or more Gaussian mixture models (GMM) associated with the speech signal and the noise signal.

A technique for texture filtering. A transition is made from a first mipmap corresponding to a first texture resolution to a second mipmap corresponding to a second texture resolution. The first texture resolution is lower than the second texture resolution. As compared to standard trilinear filtering, initiation of the transition is delayed by an offset (or bias), which serves to delay the initial use of the second mipmap until it has been loaded. Following initiation of the transition, first and second weightings are selected with a nonlinear filter, and the system interpolates between the first mipmap and the second mipmap by applying the weightings. During an initial portion of the transition, the nonlinear filter has a slope that is higher than that of the standard trilinear filter.

The invention relates to an optical angle measurement and speed measurement fused earth-moon target orbit determination method. An orbit dynamic model is established under a CRTBP earth-moon centroid rotating coordinate system; determining a state vector, and establishing a state equation according to the state information of the orbit dynamics model; determining calculated values and measured values of an azimuth angle, an elevation angle, an azimuth angle speed and an elevation angle speed, and establishing a measurement equation; and estimating earth-moon space target orbit information by adopting an EKF algorithm based on parameters of the state equation and the measurement equation, and according to an optimal control theory, taking an estimation error of minimizing a nonlinear state quantity as an optimal control target, combining the state equation and the measurement equation, optimizing a filtering parameter linearization point, and obtaining accurate earth-moon space target orbit information. The problem that the traditional angle measurement information combined with a linear filter cannot accurately reflect strong nonlinear dynamic characteristics and cannot provide high-precision track information of a target is solved.

The invention discloses a memory-driven spacecraft inertia estimation method and system and a related product, and belongs to the technical field of spacecraft control. The method provided by the invention comprises the following steps: establishing an attitude motion model containing inertia uncertainty based on a spacecraft attitude motion dynamics relationship; tracking error variables are introduced and reconstructed into an attitude tracking error system motion model; processing the signal by using a fixed parameter linear filtering operator and eliminating a differential term to obtain a first algebraic equation; transforming a regression signal based on the equation to construct an expansion algebraic model; and a historical information accumulation mechanism is introduced in combination with the extended memory regression amount to complete inertia parameter online estimation. According to the method, extra continuous external excitation is not needed, estimation can be achieved only by utilizing the normal attitude motion inherent signals of the spacecraft, meanwhile, the requirement for high-precision speed differential information is avoided, high precision and online real-time performance of inertia estimation are taken into consideration, on-orbit inertia dynamic changes can be effectively tracked, and attitude control stability is guaranteed.

Systems and methods for suppressing noise and detecting voice input in a multi-channel audio signal captured by a plurality of microphones include (i) capturing a first audio signal via a first microphone and a second audio signal via a second microphone, wherein the first and second audio signals respectively comprises first and second noise content from a noise source; (ii) identifying the first noise content in the first audio signal; (iii) using the identified first noise content to determine an estimated noise content captured by the plurality of microphones; (iv) using the estimated noise content to suppress the first and second noise content in the first and second audio signals; (v) combining the suppressed first and second audio signals into a third audio signal; and (vi) determining that the third audio signal includes a voice input comprising a wake word.

The invention discloses an infrared astronomical image iterative denoising method based on collaborative filtering operator linearization, and the method comprises the steps: obtaining an original observation image of a ground-based infrared telescope, evaluating the background noise level, and then carrying out the variance stabilization transformation, thereby obtaining a transformation domain image; performing rough hard threshold filtering on the transform domain image to generate a guide graph; performing global block matching on the guide graph to obtain a locked index table; constructing a linear collaborative filtering operator based on the locked index table; in each iteration, firstly, signal enhancement and amplitude modulation are executed to obtain an enhanced input signal; calling a linear collaborative filtering operator to carry out single linear filtering operation, and calculating filtering output; updating the estimated value based on the residual error, and finally obtaining a converged image after K times of iteration; and correcting the converged image by adopting optimal algebraic inverse transformation, and outputting a restored image. By utilizing the method, the problem that dark and weak star points are mistakenly deleted by a traditional algorithm can be effectively solved, and the limit detection magnitude of an observation system is improved.

The invention discloses a remote sensing image target detection method based on an RFTZN neurodynamics model, and belongs to the technical field of remote sensing image target detection, and the method comprises the steps: constructing a target detection task of a remote sensing image as a CEM optimization problem based on a to-be-detected remote sensing image and a detection target feature, and converting the problem into an equivalent linear matrix equation; an RFTZN neurodynamic model is constructed; constructing a solver based on the RFTZN neurodynamics model, and solving the linear matrix equation to obtain a linear filter; and applying the linear filter to an image to be remotely sensed, and extracting a detection target. According to the method, the RFTZN neurodynamics model is constructed by innovatively introducing a feedback regulation mechanism and a dynamic adaptive scale factor, so that the solution speed and the anti-noise interference capability of the RFTZN neurodynamics model are far better than those of a traditional model, and the dynamic optimization problem in the target detection process is efficiently solved.

Systems, methods, and instrumentalities are disclosed for performing intra template matching prediction (TMP) and local illumination compensation (LIC) in combination. The LIC may be used in conjunction with an intra TMP. For example, the use of LIC (e.g., at the CU level) may be adaptively selected for a codec block for intra TMP prediction. LIC may be considered (e.g., used) in performing an intra TMP prediction block search. LICs may be used for some (e.g., not all) intra TMP candidates. Whether the LIC is used may be signaled. Performing LIC with Intra TMP may be mutually exclusive from performing Intra TMP with Linear Filter Model (FLM) mode.

The application discloses a method and device for training a neural network that comprises receiving an input tensor in an input layer of a neural network, the input tensor representing one or more ch

The invention relates to a computer system based on a specific calculation model, and discloses a single-chip microcomputer communication signal optimization method based on deep learning, and the method comprises the steps: obtaining a difference between a current sampling value and a historical sampling value of a communication interface, and mapping the difference into an integer index; according to the integer index and the binary state vector of the previous moment, performing cyclic shift on the state of the previous moment and performing XOR with the feature mask to update the state of the current moment; counting the number of logic 1 bits in the state, and if the number exceeds a preset threshold value, executing bitwise AND operation to perform sparsity regularization; and traversing a final state, accumulating a preset weight according to logic 1, and reconstructing a communication signal. According to the invention, by constructing a kinetic model in which bitwise evolution and Hamming density regularization are coordinated, high-dimensional nonlinear feature extraction and internal stability control are realized under the constraint of zero floating point computing power; the problems that a traditional model is poor in real-time performance on a single-chip microcomputer and poor in linear filtering anti-noise capacity are solved.

A method in a communication system comprising a transmitter and a receiver communicating through a communication channel comprises:Obtaining a first received sequence corresponding to the transmission of a first pilot sequence by the transmitter to the receiver;Estimating r responsive to the first pilot sequence and the first received sequence, r being equal to the convolution of a first linear filter and a second linear filter;Determining a plurality of candidates for the first linear filter responsive to the estimated {circumflex over (r)};Obtaining a second received sequence corresponding to the transmission of a second pilot sequence by the transmitter to the receiver; andDetermining the non-linear function and selecting one candidate for the first linear filter among the plurality of candidates responsive to the second pilot sequence and the second received sequence.

The invention belongs to the field of image processing circuits, and particularly provides a hyperspectral image target detection circuit based on a constraint energy minimization algorithm innovatively from the perspective of an analog circuit, so that one-step hyperspectral image pixel point target detection of a hardware circuit is realized, and the defects in the prior art are overcome. The built integral circuit model can be divided into two parts, one part is an optimal filter coefficient calculation module realized through an integral negative feedback circuit, and the other part is an FIR linear filter module realized through a multiply-accumulate circuit. According to the designed circuit model, the actual calculation effect of the pixels of the hyperspectral image data set can be further studied if the spectral information of the target is known in the target detection process. The circuit can be applied to scenes such as remote sensing monitoring and environment detection, hardware implementation of the circuit has important engineering application value and development potential, and core intellectual property right can be formed in the field.

The present application provides a method for processing an audio signal, an electronic device, and a computer-readable storage medium. The present application relates to the technical field of audio processing. The method for processing the audio signal includes: obtaining a current far-end signal and a microphone signal; the microphone signal includes a near-end signal and an echo signal generated by a speaker playing the far-end signal; performing linear echo cancellation processing on the microphone signal to obtain a linear filtered signal; inputting the linear filtered signal and the far-end signal into a pre-trained residual echo cancellation DNN model to output a gain signal corresponding to the near-end signal; and determining a target audio signal to be output to the far-end for playback based on the gain signal and the linear filtered signal.