57results about How to "Protection details" patented technology

The invention relates to an optimized human face recognition preprocessing method. The method comprises the following steps: firstly, converting a color human face image from a camera into a gray level image, then performing a scale normalization processing on the gray level image to cause the human face images to have the same size and posture, dividing the human face images into low-frequency components and high-frequency components by wavelet transformation, performing a histogram equalization processing on the low-frequency components only, executing wavelet reconstruction on the processed low-frequency components and the high-frequency components, and finally processing the reconstructed images by optimized median filtering. The method has the advantages of regulating the gray level range of the human face images, enhancing the contrast, better improving the human face gray level images with higher brightness, and enhancing the human face identification efficiency in a complicated illumination environment with different postures.

The invention carries out process in frame for image; then carries out process inter frame for continuous three frames of image, which are processed in frame. The method includes following steps: using Laplace operator to carry out edge detection for image so as to obtain edge information of each pixel; if a pixel is an edge pixel or its more than two neighbor pixels in four pixels around are as edge pixels, then value of the pixel is saved; otherwise, carrying out filtering process for the pixel by using median filter; based on difference between adjacent frames, movement detection is carried out to determine whether pixel in the current frame is moved; if yes, the value is saved; if no, then result of median filter for corresponding points in three frames is as result of filtering the pixel. The invention protects details of image and sharpness furthest so as to raise vision quality of output image.

A method for fusing visible light with infrared-ray image based on contourlet transform generates fusion coefficient of contourlet domain by utilizing different characteristic of visible light and infrared-ray image and by utilizing different fusion policy of high frequency detailed portion and low frequency background portion based on contourlet decomposition of inputted visible light and infrared-ray image then obtains fused image through contourlet counter transform.

The invention relates to a gas infrared image enhancing method based on anisotropic diffusion and belongs to the field of gas detection. The method comprises the following steps of: firstly, preprocessing a gas infrared video sequence image, and respectively processing by two ways, wherein one way uses a forward anisotropic diffusion algorithm so as to spread a gas cloud cluster region, and the other way uses a bidirectional anisotropic diffusion algorithm so as to reduce the noise, and protect and enhance the detail and edge of an image background; then, carrying out discontinuous frame difference on a first processing result, and accumulating difference results; and marking the gas cloud cluster region by the means that a K mean value is clustered in the accumulated result, confirming the position coordinate of the gas cloud cluster, and finally rendering the gas cloud cluster in a colorizing way according to the corresponding position of the coordinate in a second processing result, so that the interpretation property of the gas cloud cluster can be observably improved, the quality of the gas infrared image can be improved, and human eyes can quickly detect the formed gas cloud cluster when the gas leaks. The method can be used for detecting the leakage of the invisible hazardous gas.

Disclosed in the invention is a low-light and infrared night vision image fusion method. The method comprises: step one, for one same target scene, collecting an original low-light image and an original infrared image respectively; step two, carrying out de-noising processing on the collected original low-light and infrared images; step three, carrying out effective and precise registering on the low-light and infrared images after de-noising by using an edge-feature-based image registration method; and step four, carrying out image fusion on the registered image by using a wavelet transform method. According to the invention, the low-light night-vision image and the infrared night-vision image are fused and advantages of the two images can be fully realized; and the multifunctional advantage of the night-vision system is increased, so that the observer can obtain accurate, comprehensive and reliable image information of a certain scene. Therefore, with the method, a fused image with a super high quality can be obtained and the fusion speed is fast.

The present invention relates to a safety protection device for solitary old man or woman based on supply line of life. Said safety protection device includes four digital meters of water meter, electric meter, heating meter and gas meter and data acquisition interface connected with four meters which are mounted in the interior of solitary old man or woman room, and egression sensor, main controller nd communication interface which are mounted on the room door. Besides, said invention also provides the application method of said safety protection device.

The invention discloses an image noise detection and denoising method based on a Hessian matrix, and the method comprises the steps: detecting a noise point through different characteristics of the feature value of the Hessian matrix of the noise point, an edge point and a smooth region point for impulse noise, employing the median filtering idea in a 3*3 window which takes the detected point as the center, employing the central value of the window to replace the noise point, and carrying out no processing for other points; proposing the concept of a noise point detection rate for the algorithm evaluation. Because the smooth region point or the edge point can be taken as the noise point during the detection of the noise point, a third discrimination condition is set for further improving the detection accuracy and denoising effect. For an image with the large noise density, the method also can obtain a better effect through multi-iteration. The method is better in denoising effect, and keeps more image edge and detail information. Compared with the median filtering, the method is better in effect under the condition of large noise density.

A method for forming a whole geometric body with complex shape by covering powder and laser heating includes such steps as dividing the geometric model of whole moulding box into a target body and a body to be broken, decomposing the body to be broken into a series of stripped bodies, generating the 2D profile of target body and the isolating line of said body to be broken to form a tracing for laser heating, heating by laser, to fuse the covered powder, solidifying to become a whole target body and a series of small stripped bodies, and stripping said small stripped bodies by vibration or external force to obtain a whole target body.

The invention belongs to the field of medical science, and particularly relates to a denoising system of a medical x-ray image. The system comprises a preprocessing module, a decomposition module anda reconstruction processing module which are connected end to end in sequence. The preprocessing module is further connected with an acquisition module, an image storage module and an image transmission module. The image acquisition module is connected with the image storage module. The image storage module is connected with the preprocessing module and the image transmission module. The image transmission module is connected with the preprocessing module. The denoising system of the medical x-ray image can be compiled into operation codes and stored into an x computer hard disk. During the x-ray image acquisition process, the system is automatically operated, so that the denoising effect of x-ray images is effectively achieved. The defects that a hard threshold is not continuous and a soft threshold value has constant deviation can be overcome by adopting the improved wavelet threshold method. The local features of the wavelet can effectively protect the edge detail information of images. Therefore, the system is relatively suitable for noises with relatively small noise variance.

The invention discloses an image impulse noise self-adaptive filtering method based on a genetic algorithm-back propagation (GA-BP) neural network. The method includes a first step of optimizing an initial weight value and an initial threshold value of a BP neural network through a genetic algorithm, a second step of training the CA-BP neural network, a third step of conducting noise detection on image through the well trained GA-BP neural network, and a fourth step of conducting self-adaptive average filtering on image noise points which are detected out. Due to the fact that the initial weight value and the initial threshold value of the BP neural network are poetized through the genetic algorithm, the problems of local minimum and slow speed convergence and the like are solved through a BP algorithm, and generalization ability of the neural network can be effectively improved.

The invention discloses a self-adaptive median filtering denoising method applied to image processing, which comprises the following steps of: firstly preprocessing an image, and then, (1) sorting data in a filtering window by adopting a divide-and-conquer sorting strategy; And (2) adaptively adjusting the size of a filtering window by combining the noise density. The adaptive median filtering denoising method applied to image processing is high in denoising efficiency and easy to implement.

The invention discloses a non-local means filtering method for a speckle noise pollution image. The method comprises that iterative computation of a neurons firing state image series of the speckle noise pollution image is conducted through a pulse transmission cortex model, a Renyi entropy vector is extracted by the neurons firing state image series, and non-local means filtering is conducted on the speckle noise pollution image based on the Renyi entropy vector, so that a denoised gray value is obtained. By the aid of the method, rotation invariance, translation invariance and scaling invariance can be extracted from the image containing speckle noise, the method can use more image information for denoising than traditional methods, besides, the similarity between two image pixel blocks can be calculated reasonably, image noise can be suppressed significantly, and the peak signal-to-noise ratio of the image can be improved, so that detailed information of the image can be effectively protected.

The invention relates to a hybrid image filtering method based on target scale, which is mainly used for self-adaptive median filtering and anisotropic diffusion. The method is characterized in that: first anisotropic diffusion is carried out to a processed image, thereby inhibiting most noise other than pulse noise; then the target scale of all pixels of the diffused image are calculated; and finally the calculated target scale carries out self-adaptive median filtering to the diffused image so as to remove the rest pulse noise after diffusion and obtain the final result. The method can remove the noise of images with various gray levels, and simultaneously keeps the details of the images while removing the noise.

The invention relates to the technical field of digital image processing, in particular to a switch weighting vector median filter method utilizing edge detection. The switch weighting vector median filter method improves noise reduction capacity vector median filter, and protects image details. According to the technical scheme, the switch weighting vector median filter method comprises the following steps that the similarity of pixel points in a noise image are compared with that of the corresponding pixel points in a reference image obtained through scalar median filter which is carried out on R, G and B components of the noise image, so that noise monitoring is carried out, vector median filter is carried out on suspected noise points through the weighting method based on edge detection, the noise reduction capacity of vector median filter is improved, and the image tails are protected. The switch weighting vector median filter method is mainly applied to digital image processing.

The method includes steps: (1) calculating brightness graph based on the input image; (2) dividing frequency brightness graph of image so as to obtain high frequency segment and low frequency segment; (3) recognizing characters of each pixel in image including high frequency noise in small amplitude, not boundary noise, isolated noise point and useful signal of image, and making relevant process for removing noise. Based on characters of noise, using associated information between pixels, the disclosed method for removing isolated noise point in video carries out image classification. Moreover, using different treatment processes, the invention obtains good effect for removing noises including isolated noise point, not boundary noise and high frequency noise in small amplitude. Meanwhile, under basis of removing noises, the invention reserves detail of image and definition of edges farthest so that removing noise does not blur image.

The invention relates to a melting phenomenon reality simulation method based on physics. The melting phenomenon reality simulation method is suitable for modeling of a melting phenomenon in the natural world and comprises the steps that 1, the solid-to-liquid particle relation, heat transfer model, heat source radiation temperature model calculation mode of an object are established according toa particle model; 2, a liquid drop surface tension model is calculated by using a smooth particle fluid dynamical model and an interface theory of mechanics; 3, after liquid drops are generated, an adhering strength model when flow exists between the liquid drops and a solid also needs to be established so as to complete real flow of the liquid drops, and then a real effect is drawn. By adopting the melting phenomenon reality simulation method, a melting phenomenon protecting details can be quickly obtained, and the heat transfer model, a surface tension model constraining liquid drop behaviors and the adhering strength model embodying melting flow, which are needed for object melting, are designed. The melting phenomenon reality simulation method can achieve reality simulation of common object melting scenes, is simple and good in stability and has a certain practical value.

The invention relates to the technical field of digital image processing, and aims at avoiding a staircase effect and a block effect by conducting improvement on a traditional anisotropism filtering method. According to an anisotropism filtering method based on a self-adaptive averaging factor, in the image filtering process, edge fragmentary information is protected by reducing the smoothing degree of noise and marginal areas. Therefore, according to the technical scheme, the anisotropism filtering method based on the self-adaptive averaging factor comprises the steps that pretreatment is conducted on a noise image by adopting a Gaussian filter, the pretreatment formula comprises the step that an improved anisotropic filtering is utilized, the size of the value of a parameter K is determined according to differences of gradient values of each diffusion pixel to a central pixel, that is to say, a self-adaptive equation is utilized to replace the value of an original fixed parameter K, the value of the K of the improved anisotropic filtering is made to reduce on the noise and marginal areas, and the smoothing degree of the improved anisotropic filtering is reduced; the value of the K of the improved anisotropic filtering is increased on the smooth and flat areas, and the smoothing degree of the improved anisotropic filtering is increased. The anisotropism filtering method based on the self-adaptive averaging factor is mainly applied to digital image processing.

The invention provides a wavelet denoising method based on a self-adaptive non-local mean value. The method comprises the following steps: step S110: carrying out following steps on each color channelcontaining noise images: S111, extracting a low-frequency wavelet component and a high-frequency wavelet component of a noise-containing image under a color channel by adopting a wavelet transform algorithm; S112, calculating neighborhood calibration noise in a high-frequency wavelet component search window according to a constructed edge discrimination operator; S113, according to the domain calibration noise, sequentially calculating the similarity between a reference window and a target neighborhood window in a search window, and determining a weighting coefficient of a target wavelet coefficient based on the similarity; Step S114, according to the weighting coefficient of the target wavelet coefficient, updating the target wavelet coefficient, and based on the updated target wavelet coefficient, obtaining a denoised image containing the noise image under the color channel by using a wavelet inverse transformation algorithm; And S120, synthesizing the de-noised image under each color channel to obtain a de-noised image containing the noise image. The method improves wavelet denoising.

The invention discloses an improved TPS (thin plate spline) model based gel image correction algorithm. The algorithm includes: comparing a distorted image with a template image to acquire a deformation degree coefficient beta and a column number nclm of the distorted image, and introducing a height correction coefficient to obtain a 'smile' deformation model; acquiring n pairs of corresponding distorted image mark point coordinates ipts and template image mark point coordinates opts in the distorted image and the template image, and utilizing the 'smile' deformation model to transform the coordinates ipts to obtain mark point coordinates Sipts; utilizing the mark point coordinates Sipts and the template image mark point coordinates opts to obtain unknown parameters wi (i=1...n), a0, ax and ay of the TPS deformation model; creating an image wimg which is equal to the distorted image in size and zero in gray value, and performing 'smile' deformation and TPS deformation on the coordinates; assigning the gray value of the template image to the image wimg with the gray value being zero according to the corresponding relation between the distorted image coordinates and the template image coordinates; considering the assigned image wimg as the corrected image. Compared with the original correction algorithm, the improved TPS model based gel image correction algorithm has the advantages that better correction effect is obtained, and good preconditions are provided for matching of gel images.

The invention discloses a synthetic rain map rain removal method based on dictionary training and sparse representation, which is applied to rain removal recovery of a synthetic rain map by improvinga method applied to image super-resolution reconstruction. In the training stage, a pure rain template is used for adding rain to a group of rain-free images, a rain-rain-free training set is constructed, and a rain dictionary and a rain-free dictionary are obtained through training; and in the test stage, applying another pure rain template to rain the test rain-free image to obtain a test synthetic rain map, and based on the rain dictionary, performing sparse representation on the test synthetic rain map to obtain a sparse representation coefficient of the test synthetic rain map, obtaininga third different sparse representation coefficient of the pure rain template based on the rain dictionary, and subtracting the two sparse representation coefficients to further remove rain-related components in the representation coefficients, and finally, combining the subtracted sparse representation coefficients with the trained rain-free dictionary to obtain a final rain removal result of thetest synthesis rain map. According to the method, detail information in the image can be well protected while the image is subjected to rain removal.

The invention relates to the field of image processing and display and aims at providing a filtering method capable of effectively distinguishing image edges and noise. Therefore, according to the technical scheme, an edge-preserving-based weighted anisostropic diffusion filtering method comprises performing noise and edge detection; according to whether diffused pixels belong to an edge region, a flat region or a noise region, distributing different weighted values for the pixels. Therefore, the noise can be effectively eliminated, and meanwhile, losses of edge detail information of an image can be avoided. The edge-preserving-based weighted anisostropic diffusion filtering method is mainly applied to image processing and display.

A typhoon cloud picture enhancement method provided by the invention is characterized in that by decomposing an original typhoon image into a base layer and a detail layer, the image edges and detailscan be better protected, and the noise can be effectively removed; by using a corrected histogram equalization algorithm for the base layer and the detail layer, the brightness and contrast of the image are improved, and the saturated artifacts are removed at the same time; and finally, the enhanced base layer and detail layer are fused by adopting a Laplacian pyramid, so that the image edge information can be effectively protected to achieve a better enhancement effect.

The invention relates to a fluid surface detail protection method based on grid and particle coupling. The fluid surface detail protection method comprises the following steps: (1) solving a Navier-Stokes equation (N-S), and forming a simulation model of a main body fluid by adopting a grid method; (2) improving an LBM-VOF method, and tracking the surface of the fluid by using an improved VOF-LBMcoupling algorithm; (3) generating particles at the abnormal surface grid position, and then evolving the particles through a particle method; (4) designing a coupling algorithm of grids and particles, and integrating grid fluid and particle fluid into the same scene so as to ensure physical conservation of the whole flow field and reasonable physical information transmission between the grids andthe particles; and (5) finally, performing realistic rendering by using a screen space method, and realizing realistic and real-time fluid rendering on the GPU by drawing a sphere, calculating a depth value of each pixel point, performing depth filtering, solving a normal vector according to the depth values and position information of the pixel points, and performing illumination rendering. Realistic and real-time fluid rendering is realized on GPU.

The invention relates to a high spectral image strip noise elimination method with integration of wavelet transformation and local interpolation. The high spectral image strip noise elimination method with integration of wavelet transformation and local interpolation includes the following steps: inputting a high spectral image data cube; judging the high spectral image wave band data; rejecting the damaged high spectral image wave band data; reading the normal high spectral image data of the kth wave band; respectively using two dimension stationary wavelet and local interpolation to filtering strip noise of the high spectral image data; and then performing fusion processing. The high spectral image strip noise elimination method with integration of wavelet transformation and local interpolation can effectively eliminate common thin strip noise, can eliminate the strip noise of broadband and ultra wide band and the strip noise which is distributed without rules, can preferably keep some details and edge geometrical information in the images, at the same time can be applied to different high spectral images, and has relatively high universality.

The invention provides a CS image denoising reconstruction method based on hyperspectral total variation. The CS image denoising reconstruction method comprises the following steps: initializing a reconstructed image, an iterative index value and a noisy observation value; iteratively updating the obtained reconstructed image by using the noisy observation value to obtain an estimated value; respectively inputting the estimated values into a CS reconstruction model based on the l1-norm and the HTV to obtain an intermediate reconstruction image; performing sparse representation on the intermediate reconstructed image by using Starlet transform to obtain a Starlet coefficient; performing denoising filtering on the Starlet coefficient by using the new threshold operator and the improved BayeShrink threshold to obtain a curvelet coefficient; performing Starlet inverse transformation on the curvelet coefficient to obtain a reconstructed image; and judging whether an iteration stopping condition is met or not, and carrying out loop iteration. According to the method, while most noise information in the high-noise image is removed, details, textures and other feature information in the image can be effectively protected, the method is easy to implement and high in robustness, and the denoising reconstruction problem of the high-noise image is effectively solved.

The invention provides a nonlocal median value blind noise reduction method based on visual perception, comprising the following steps: constructing an impulse noise blind detector based on the visual outlier measure of the pixels in a digital image, wherein the visual outlier measure is obtained by quantizing the visual similarities in impulse noise of different models and fusing different visual feature quantization results; extracting the nonlocal information of the image, and constructing a nonlocal median value calculation model; calculating regularization parameters according to the visual outlier measure and the nonlocal information, and establishing a nonlocal median value regularization item; and building a nonlocal median value noise reduction functional model, and adaptively repairing the noise pixels in the image. According to the blind noise reduction method of the invention, impulse noise of different models and densities in the digital image is processed in a unified manner according to the visual features of impulse noise, image self-similarity and outlier data mining, and the problem that noise pixels are hard to repair effectively due to unknown impulse model, high-density noise and image multi-modal complexity in an actual noise reduction process is solved.

The invention relates to a high-precision rapid focus detection method based on a push-broom underwater hyperspectral original image. The method comprises the following steps: collecting the push-broom underwater hyperspectral original image; carrying out bilinear interpolation down-sampling to obtain a sampling graph; performing noise suppression on the sampling image; carrying out overlapped Laplacian operator weighting, and calculating a central pixel point gray value of each image sub-block; summing gray values, and identifying an optimal coarse focusing position; focusing a stepping motor, and collecting a push-broom underwater hyperspectral original image; carrying out edge maintenance; decomposing the filtered image by adopting frequency domain wavelet transform, and recombining the filtered image; obtaining a fine focusing evaluation value of the push-broom underwater hyperspectral original image; and completing focusing identification. The obtained image focusing evaluation value has good real-time performance, unimodality, environmental applicability and noise resistance, and the fine focusing identification algorithm has better function sensitivity than the coarse focusing identification algorithm.

The present invention discloses a method for effectively removing camera speckle noise suitable for hardware realization, wherein 3*3 window data in the image is obtained by a definite step length, and the window data is wave filtered in four directions of horizon, verticality, diagonal downward, and diagonal upward; slight difference between the pixel value of the center pixel piont and the pixel value of 8 adjacent pixel points are specifically analyzed, and then the relative darker points and lighter pixel points are respectively filtered ; even treating pictures rich in detail, the edge and detail information of the picture can be still effectively protected; while wave filtering operation only contains comparison and signed magnitude arithmetic, which is simply to calculate and much suitable for hardware realization. The method provided by the present invention can rapidly and effectively reduce speckle noise produced in camera, and has no ambiguity in detail. The treated picturenoise can essentially be inhibited and the visual effect is good.

The invention provides an image processing method and device, a terminal and a readable storage medium. The image processing method comprises the following steps: step 1, acquiring an initial image which comprises a plurality of blocks with the same size, wherein each block comprises at least one pixel point; step 2, using any one of the blocks as a target block, obtaining weight values of a plurality of adjacent blocks around the target block according to gain data corresponding to the target block in the expansion graph and a noise level of the target block, the adjacent blocks being blocks surrounding and connected to the target block, wherein the size of the expansion image is the same as that of the initial image; and step 3, performing noise reduction processing on the target block according to the plurality of weight values, and traversing all blocks in the initial image, taking each block as a primary target block, and repeatedly executing the second step and the third step to obtain a noise reduction image.

The invention relates to a peak envelope-based seismic data peak noise recognition and suppression method. The technical key point is that a half-width threshold value and an energy threshold value are set according to the differences between peak noise and a target seismic signal in the aspects of lasting time length and energy mean value in a time window so as to recognize the peak noise and airplane signals in seismic data and suppress the peak noise. Experiments prove that the peak envelope-based seismic data peak noise recognition and suppression method disclosed by the invention is capable of realizing the recognition and suppression of the peak noise and is particular effective for the conditions that the target seismic signal is long in lasting time and the peak noise is short in lasting time; and the method is capable of recognizing and suppressing the peak noise and protecting the details of the target seismic signal, thereby effectively improving the quality of the seismic data and reducing the acquisition cost.