42results about How to "To achieve the purpose of denoising" patented technology

The invention discloses a monocular vision-based lane line and front obstacle detection method, which relates to a lane line detection and front obstacle detection method and aims at solving defects of inaccurate lane line detection and low front obstacle detection precision in the prior art. The method particularly comprises steps: 1, a vehicle-mounted CCD camera is adopted to acquire an original image, gray processing is carried out on the original image, and a gray image is obtained; 2, the gray image is pre-processed, and a binary image after clutter removal is obtained; 3, the binary image after clutter removal is subjected to near half area lane line initial detection and line crossing alarming detection based on Hough transform; 4, according to point coordinates on the near half area lane line initial detection line obtained in the third step, parabolic fit is carried out on a near-view field lane line and a far-view field lane line, and a fit lane line is obtained; and 5, obstacle detection is carried out in the fit lane line obtained in the fourth step. The method is applied to the technical field of image processing.

The invention discloses an image processing method based on wavelet transform, belongs to the technical field of image processing methods, and solves problems that time domain information is discarded in image processing through Fourier transform. The image processing method based on wavelet transform comprises the steps that (1) an original image is acquired; (2) wavelet decomposition is performed on the original image so that a low-frequency sub-band LL and three high-frequency sub-bands LH, HL and HH are obtained, and non-linear enhancement is performed on high-frequency coefficients so that wavelet coefficients after image enhancement are obtained; and (3) wavelet inverse transform is performed on the two wavelet coefficients after processing so that the enhanced image is obtained. The image processing method based on wavelet transform is used for image processing.

The invention relates to a wireless multi-hop network technology, in particular to a receiving and transmitting method for realizing irrelevant receiving by a relevant network code. In order to improve the wireless multi-hop communication transmission efficiency and reduce the technical complexity and the communication error code rate, the method comprises the following steps of: carrying out relevant network PCNC coding in a source end physical layer; detecting and utilizing a unique mapping relation to map a detected signal into a de-noised mixed signal by a relay through the channel attenuation coefficient and carrying out differential modulation on the mixed signal to be transmitted to two mutually communicated source ends; during receiving at the source ends, realizing the irrelevant receiving under the condition of not learning any channel characteristic by a differential relation of signals before and after mapping by the mixed signal to recover a signal sum of the two continuous time slots and subtracting the signal at the local end to obtain a signal at the opposite end. The invention is mainly used for the transmission in wireless cooperative communication needing the relay and the transmission between a base station without a straight transmission line and a mobile terminal.

The invention relates to a character recognition method and system based on an attention mechanism, and relates to a deep learning and image processing technology. According to the method, a convolutional neural network and a linguistic module based on an attention mechanism are used as the backbone of a deep learning model, a customized loss function is used to reinforce feature map extraction, the model is guided to learn to distinguish a foreground and a background during training, and forward and reverse bidirectional decoders are introduced to perform bidirectional decoding on characters.The method and the system are high in anti-interference capability, attention drifting can be reduced, and meanwhile, the situation that final recognition fails due to the fact that the first character is difficult to recognize during the forward decoding of the model can be avoided.

The invention provides a point cloud noise removal method and system, computer equipment and a computer readable storage medium. The point cloud noise removal method comprises the steps: obtaining point cloud data; determining abnormal points in the point cloud, and deleting the abnormal points; and determining noise points in the point cloud after the abnormal points are deleted, and deleting thenoise points. Through two main steps of abnormal point deletion and noise point deletion, the purpose of point cloud noise removal is achieved, and a good point cloud noise removal effect is achieved, and meanwhile the processing speed of point cloud noise removal is increased.

The invention discloses a switching median filtering based encrypted image denoising method, which belongs to the field of multimedia information security and image processing. The method involves an image owning party, an image denoising party and an image receiving party, the image owning party has a noisy image, and hopes to denoise the image by the image denoising party and then transmit the image to the image receiving party, but because the image owning party distrusts the image denoising party, after the image is encrypted, the image owning party transmits the image with covered contents to the image denoising party to process. The method comprises the following steps: firstly, carrying out encryption on the image by using a particular image encryption method, and due to the particularity of the encrypted image, carrying out denoising processing on the image by using a non-traversing switching median filtering; firstly, dividing the image into blocks, and finding the mean value, standard deviation and mid-value of each block; then, obtaining a threshold according to the standard deviation, and determining a noise point; and finally, replacing the pixel value of the noise point with the mid-value. The method disclosed by the invention not only effectively realizes the image denoising, but also protects the information security of images.

The invention discloses a fractal-wavelet self-adaption image denoising method based on a multivariate statistical model. The fractal-wavelet self-adaption image denoising method comprises the following steps: an extended GGD model, namely an MGGD model, is selected to establish the multivariate statistical model, and image denoising is completed by fractal-wavelet transform; a vector (x) of P2 closest to a vector (x) of P1 is obtained by minimizing a residual R1, and a parameter alpha and a parameter beta are self-adaptively adjusted; fractal-wavelet noiseless image coding is predicted self-adaptively for noise images by quadtree-based segmentation to realize denoising. The fractal-wavelet self-adaption image denoising method can overcome defects that in the prior art, the operation process is complex, the consumed time is long, and the reliability is low, and has the advantages that the operation process is simple, the consumed time is short, and the reliability is high.

The invention provides a new threshold function seismic data de-noising method based on wavelet transform. The method addresses the problem that the quality of seismic data in a low signal-to-noiseratio region of oil and gas seismic exploration cannot meet the requirements of treatment and interpretation. By establishing a new threshold function overcoming the defects of hard and soft thresholds and using wavelet transform as a core idea, the de-noising goal of seismic data is gained, and a new threshold function is determined. Compared with the current similar technologies, the method hasthe advantages that the de-noising result is closer to original useful signals and the quality of the seismic data is greatly improved.

The invention provides an SD-OCT denoising method based on an unsupervised adversarial neural network, and the method achieves the generation of an image with EDI-OCT quality from SD-OCT through the unsupervised learning of the migration from an SD-OCT image domain to an EDI-OCT image domain, thereby achieving the purpose of denoising. The global structure loss and the local structure loss are added to the model, so that the structure information and the local details of the image can be effectively maintained. Compared with an existing image noise reduction algorithm, the method has the advantages that image noise and strip artifacts are effectively removed from the processed image of the model, local details of the image are well reserved, and the processed image is closer to the original image. Meanwhile, the parameter index of the model prediction image is higher than other algorithm processing results. Therefore, no matter from subjective visual effect or objective quality evaluation, the processing result of the model provided by the invention is superior to the processing results of other algorithms, which indicates that the model provided by the invention is feasible and effective in the aspect of SD-OCT image denoising.

The invention provides a multimode three-dimensional ultrasound detection and imaging device and method. The device comprises a box, an emission and reflection receiving spiral track, a transmission receiving spiral track, an emission and reflection receiving end, a transmission receiving end and a back-end processing system. When the device is used for detection, a detection object is arranged in the middle of the box and surrounded by the emission and reflection receiving spiral track and the transmission receiving spiral track. The emission and reflection receiving end can slide back and forth on the emission and reflection receiving spiral track and the transmission receiving end can slide back and forth on the transmission receiving spiral track for performing spiral three-dimensional scanning on the entire detection space. The multimode three-dimensional ultrasound detection and imaging device and method inverses a path (reflection channel) through which reflected ultrasonic waves pass and a path (transmission channel) through which transmitted ultrasonic waves pass through reflected wave information and transmitted wave information for further achieving three-dimensional imaging of characteristic distribution of the transmission channel and reflection channel. According to the multimode three-dimensional ultrasound detection and imaging device and method, the simple device is used for achieving fast and convenient detection, and the cost is low. The multimode three-dimensional ultrasound detection and imaging device and method can be expanded to non-medical application.

The invention provides a seismic data de-noising method adopting an adaptive wavelet threshold function. The method addresses the problem that the quality of seismic data in a low signal-to-noise ratio region of oil and gas seismic exploration cannot meet the requirements of treatment and interpretation. By establishing a new threshold function overcoming the defects of hard and soft thresholdsand using wavelet transform as a core idea, the de-noising goal of seismic data is gained and a new threshold function is determined. Compared with the current similar technologies, the method has theadvantages that the de-noising result is closer to original useful signals and the quality of the seismic data is greatly improved.

The invention discloses an SVM-based recognition method of building angular points in a high-resolution optical image. The process of the method comprises the steps that the high-resolution optical image is subjected to ortho-rectification, the influences of vegetation on a classification result are removed through an NDVI, a binary support vector machine (SVM) is used for marking out building blocks in the image in a classified mode, a candy operator is used for performing edge detection on the area where the building blocks are located, a structure chart of point sets of building nodes is obtained through hough transformation, and the SVM is used for classifying the point sets to obtain the building angular points. Through the method, the recognition precision of the building angular points in the optical image can be greatly improved, and the information extraction rate of buildings and other artificial objects in the image with a high spatial resolution is increased.

The invention discloses a nuclear magnetic resonance image denoising method based on image-block self-similarity prior. The method comprises the following steps of based on self-similarity among imageblocks, constructing a clustering regular term; based on the clustering regular term and a low rank decomposition denoising method, establishing an image denoising model; and optimizing the image denoising model, and based on the optimized image denoising model, carrying out denoising processing on an image. Disadvantages in an existing denoising method are overcome. Noises are removed and simultaneously detail information of the image can be kept to a greater degree.

The invention provides a denoising method and device for mobile terminal trajectory data. The method comprises the steps: obtaining trajectory data including a first position, a second position, a third position and corresponding time based on each piece of mobile terminal trajectory data; performing calculating to obtain a first distance, a second distance, a third distance, a first time and a second time; recording the first track data of which the first time and the second time are less than the set duration, temporarily confirming the noise data of which the first time and / or the second time are / is not less than the set duration, and recording the noise data as second track data; performing calculating to obtain a second position included angle and first, second and third speeds; temporarily determining the first track data as noise data and recording the noise data as third track data under the condition that the acquired data judges that the first track data meets the noise filtering condition; and judging whether the second position and the third position are equal or not for each piece of track data in the second track data and the third track data: if yes, determining thedata as noise data, and removing the corresponding mobile terminal track data. According to the scheme, the purpose of denoising the track of the mobile terminal can be achieved.

The invention relates to a cable porcelain shell terminal infrared image de-noising method considering the inter-scale correlation, which includes the following steps: (1) inputting a cable porcelain shell terminal infrared image to be de-noised; (2) carrying out two-dimensional wavelet decomposition on the infrared image to get a scale coefficient of the highest decomposition layer and wavelet coefficients in three directions of different decomposition layers; (3) calculating the absolute value of the inter-scale correlation coefficient of the wavelet coefficients; (4) using a means clustering method to divide the absolute value of the correlation coefficient obtained in step (3) into a correlation coefficient of effective wavelet coefficients and a correlation coefficient of non-effective wavelet coefficients; (5) directly setting the non-effective wavelet coefficients obtained in step (4) to zero, and retaining the effective wavelet coefficients obtained in step (4); and (6) using the scale coefficient of the highest decomposition layer obtained in step (2) and the processed wavelet coefficients obtained in step (5) to carry out two-dimensional wavelet reconstruction to get a de-noised infrared image. Through the de-noising method of the invention, noise can be removed, and the details of the image can be all retained.

The invention relates to a sensor performance on-line test device and method based on a multi-threshold wavelet under strong interference. First, a test point environmental parameter signal detected by a sensor is subjected to filtering and collection; then the collected signal is subjected to noise reduction processing through fuzzy multi-threshold wavelet transformation; wavelet decomposition is carried out, wavelet coefficients are obtained, the membership degree of each wavelet coefficient is obtained, the wavelet coefficients with the membership degrees exceeding the preset threshold are rejected, wavelet reconstruction is carried out by utilization of the wavelet coefficients with the membership degrees within the preset threshold, detection data after noise reduction is obtained, sensor characteristic indexes are calculated according to the detection data after noise reduction and environmental parameter values of the test point, and sensor performance on-line automatic test is finished. Noise reduction processing is carried out with combination of filtering and fuzzy multi-threshold wavelet transformation, the signal outline after noise reduction is obvious and clear, no detail signals are lost, fidelity with an original signal is kept, and the signal to noise ratio of the signal is raised obviously.

The invention relates to an image processing model training method and device and a medium. The method comprises: for each training scene in N training scenes, acquiring M frames of long-distance shotimages respectively, wherein N and M are positive integers larger than 1; for each training scene, obtaining a short-distance standard image of the training scene; for each training scene, cutting the M frames of long-distance shot images into images with the same field of view as the short-distance standard image, and aligning the cut M frames of long-distance shot images based on the short-distance standard image; forming a training sample pair of each training scene by the M frames of long-distance shot images and the short-distance standard images after alignment of each training scene, and forming a training sample set by N training sample pairs corresponding to the N training scenes respectively; and training the image processing model based on the training sample set. By the adoption of the method, the digital zoom effect is effectively improved, and meanwhile the purpose of denoising is achieved. By adopting the method, a user can see more clearly when shooting distant characters or other objects.

The invention relates to a satellite-borne laser altimeter data denoising method. The satellite-borne laser altimeter data denoising method is characterized by including the following steps that S1, satellite-borne laser altimeter data are obtained; S2, the satellite-borne laser altimeter data are decomposed into a plurality of intrinsic mode components and a monotonic remainder term based on empirical mode decomposition; S3, a Hurst exponential value of each intrinsic mode component is calculated; and S4, based on the Hurst exponential value of each intrinsic mode component, a pre-establish optimization strategy is adopted to obtain de-noised satellite-borne laser altimeter data. Compared with the prior art, the satellite-borne laser altimeter data denoising method can adaptively denoisethe satellite-borne laser altimeter data, and has the advantages of good stability and high accuracy.

The embodiment of the invention provides a denoising method and a mobile terminal, relates to the technical field of communication, and aims to solve the problem that motor noise doped in a recorded audio file cannot be eliminated. The denoising method comprises the following steps: under the condition that a first target file contains noise data, obtaining a target feature fingerprint set of thefirst target file; under the condition that a first feature fingerprint set matched with the target feature fingerprint set is obtained from a database, obtaining a second target file corresponding tothe first feature fingerprint set; and updating the first target file according to the second target file, wherein the database is used for storing files without noise data. The denoising method in the embodiment of the invention is applied to the mobile terminal.

The invention relates to a magnetotelluric denoising method and system. The method comprises the steps of carrying out the overlapping segmentation of a magnetotelluric pulsation signal, determining the approximate entropy of each sounding data segment, determining the sounding data segment with the approximate entropy greater than an approximate entropy threshold as an effective sounding data segment, and determining other sounding data segments as to-be-denoised sounding data segments; selecting approximate entropy to screen the data segments, so that 'over-processing' of the effective sounding data segments is prevented; performing singular value decomposition on the to-be-denoised sounding data segments, multiplying the decomposed singular value matrix, the left singular matrix and theright singular matrix to obtain a plurality of component matrixes, constructing corresponding component data segments, determining the approximate entropy of each component data segment, and summingthe screened component data segments of which the approximate entropy is greater than the approximate entropy of the sounding data segments; obtaining denoised sounding data segments, and achieving the purpose of denoising; and finally, reconstructing the denoised sounding data segments and the effective sounding data segments to obtain a reconstructed denoised magnetotelluric pulsation signal, thereby improving the accuracy of data denoising.

The invention discloses a monitoring method for a server fan rotating speed. The method includes the following steps: a fan rotating speed monitoring request is received; sampling in a preset number of times is performed on a fan rotating speed signal, and sampling signals in a preset number are obtained; and filtering processing is performed on the various sampling signals by using a median filtering algorithm, and fan rotating speed data after processing are obtained. By applying the technical scheme provided by the monitoring method for the server fan rotating speed, the probability of mistaken judgment on the fan rotating speed is greatly lowered, manpower and material resources are saved, the monitoring method is simple and easy to carry out, and cost is greatly lowered. The inventionfurther discloses a monitoring device and equipment for the server fan rotating speed and a storage medium. Corresponding technical effects are achieved.

The invention discloses a sound emission signal denoising method based on a denoising autoencoder (DAE), which is characterized in that the denoising autoencoder is trained through unsupervised learning to learn more stable invariance features, so that the error between a reconstructed signal and an original signal is converged to a minimum value, and the denoising purpose is achieved. A denoising experiment is carried out on the basis of processing 3011 corrosion sound emission signal samples; the experiment result shows that when the number of hidden layer neurons is 300, the denoising model has a good denoising effect, and the denoising model of the denoising autoencoder has better denoising performance and generalization compared with a wavelet threshold denoising method. The denoising autoencoder denoising model is applied to sound emission signal denoising, can effectively remove noise, and is of great significance to subsequent sound emission signal recognition processing.

The invention relates to a device and method for online automatic testing of sensor performance based on multi-threshold wavelets in a strong interference environment. First, the environmental parameter signals of the test points detected by the sensor are filtered and collected; The signal is subjected to noise reduction processing. First, wavelet decomposition is performed to obtain the wavelet coefficients, and the degree of membership of each wavelet coefficient is obtained. The wavelet coefficients whose membership degree exceeds the preset threshold are discarded, and the wavelet coefficients whose membership degree is within the threshold are used for wavelet analysis. Reconstruct to obtain the detection data after noise reduction, calculate the sensor characteristic index according to the detection data after noise reduction and the environmental parameter value of the test point, and complete the online automatic test of sensor performance. The present invention combines filtering and fuzzy multi-threshold wavelet transformation to perform noise reduction processing. After noise reduction, the signal outline is more obvious and clear without loss of detail signal, the fidelity with the original signal is maintained, and the signal-to-noise ratio of the signal increases significantly.

The invention discloses a webpage denoising system and method based on maximum similarity matching, belonging to the technical field of the internet. The system comprises a webpage acquisition module, a pretreatment module, a webpage DOM (document object model) generation feature tree module, a feature tree maximum similarity matching module and an aggregation evaluation module, wherein the webpage acquisition module is connected with the pretreatment module and transmits webpage code data; the pretreatment module is connected with the webpage acquisition module and transmits a pretreated target webpage; the pretreatment module is connected with the webpage DOM generation feature tree module and transmits pretreated webpage data; the webpage DOM generation feature tree module is connectedwith the feature tree maximum similarity matching module and transmits feature tree data; the feature tree maximum similarity matching module is connected with the aggregation evaluation module and transmits a webpage content block candidate set; and finally, the aggregation evaluation module outputs the webpage content block. The invention can be better suitable for majority of content-type websites.

The present invention provides a VVC loop filtering method and system based on multi-layer GAN, and constructs a filtering model. The filtering model includes a multi-layer GAN model, and each GAN model is responsible for capturing image distribution at different scales; importing other methods of filtering The final video stream file and the key frame group of the original video file are used as training data, as the input of each layer of GAN model, and the pixel difference between them is used to train the filtering model; the trained filtering model is embedded in the VVC loop, and the video The sequence is divided into several small image groups, and an embedded filtering model is used to filter the divided video sequence; the invention reduces the overhead caused by loop filtering and improves the compression efficiency.

The invention discloses a method for recognizing an image scale invariant pattern under the noise condition. The method comprises two separated stages: a training stage of training the image of a specific object, namely a reference image, to obtain a correlation filter, and a correlation detection stage of recognizing the object to be detected by using the obtained correlation filter. In the method, noise estimation can change adaptively with the change of the signal to noise ratio, and high noise immunity is achieved. The main principle of the method is that: the image can be reconstructed by an intrinsic mode function correlated to the important structure of the image, thus the correlation filter can be built by some specific layers (the intrinsic mode function) to realize the scale invariant pattern recognition; moreover, as the noise distribution on the intrinsic mode function of each layer has a certain rule, the aim of eliminating noise can be fulfilled by the noise distributionrule.

The invention discloses a method for recognizing an image scale invariant pattern under the noise condition. The method comprises two separated stages: a training stage of training the image of a specific object, namely a reference image, to obtain a correlation filter, and a correlation detection stage of recognizing the object to be detected by using the obtained correlation filter. In the method, noise estimation can change adaptively with the change of the signal to noise ratio, and high noise immunity is achieved. The main principle of the method is that: the image can be reconstructed by an intrinsic mode function correlated to the important structure of the image, thus the correlation filter can be built by some specific layers (the intrinsic mode function) to realize the scale invariant pattern recognition; moreover, as the noise distribution on the intrinsic mode function of each layer has a certain rule, the aim of eliminating noise can be fulfilled by the noise distribution rule.

The invention discloses a hyperspectral remote sensing image classification method based on a combination of six-layer convolutional neural network and spectral-spatial information, which selects hyperspectral remote sensing image data of a certain number of bands, and performs spatial analysis on the selected two-dimensional image data of each band. Mean filtering, and then convert the format of the multi-band data corresponding to each pixel, and convert the one-dimensional vector into a square matrix, that is, each pixel corresponds to a square matrix data. Then design a six-layer classifier based on deep learning template, including input layer, first convolutional layer, maximum pooling layer, second convolutional layer, fully connected layer, output layer; extract the square matrix corresponding to several pixels The data is used as the training set, input the classifier and train the classifier; extract the square matrix data corresponding to several pixels as the test set, input it into the trained classifier, observe the classification results output by the trainer, and compare with the real The classification information is compared to verify the performance of the classifier. The classification accuracy rate of the present invention is higher than the existing 5-CNN method.

The invention provides an exponential signal denoising method achieved by means of prior information and relates to a denoising method for exponential signals.The prior exponential signals form a Hankel matrix according to a set sequence, the Hankel matrix is subjected to singular value decomposition, and a prior signal space and a prior singular value are obtained; then a Hankel matrix with the same size as the former Hankel matrix is established for the target exponential signals, and the matrix of the target signals is decomposed through the prior information space; after a weighted threshold is obtained from the prior singular value, a Hankel matrix of the denoised target signals is obtained from the singular value of the target exponential signals according to the weighted threshold; then the Hankel matrix of the target signals is solved, and finally the denoised signals are obtained.Through the prior information of the reference signals, the speed is high, the effect is good, and operation is easy.Denoising of a magnetic resonance spectrum can be achieved for signals with exponential features such as time domain signals of the magnetic resonance spectrum through the method, and the purposes of shortening the sampling time and increasing the signal to noise ratio of the spectrum are achieved.