939 results about "Gaussian filter" patented technology

The invention relates to a movement human abnormal behavior identification method based on template matching, which mainly comprises the steps of: video image acquisition and behavior characteristic extraction. The movement human abnormal behavior identification method is a mode identification technology based on statistical learning of samples. The movement of a human is analyzed and comprehended by using a computer vision technology, the behavior identification is directly carried out based on geometric calculation of a movement region and recording and alarming are carried out; the Gaussian filtering denoising and the neighborhood denoising are combined for realizing the denoising, thereby improving the independent analysis property and the intelligent monitoring capacity of an intelligent monitoring system, achieving higher identification accuracy for abnormal behaviors, effectively removing the complex background and the noise of a vision acquired image, and improving the efficiency and the robustness of the detection algorithm. The invention has simple modeling, simple algorithm and accurate detection, can be widely applied to occasions of banks, museums and the like, and is also helpful to improve the safety monitoring level of public occasions.

The invention discloses a color digital image enhancing and denoising method under random illumination. The method is characterized by comprising the following steps of: 1, image denoising processing, namely converting an original image from an RGB space to a YCbCr space, removing a Gaussian noise by using a Gaussian filter and removing a salt and pepper noise by performing median filter; 2, image brightness/contract stretch processing, converting the image from the RGB space to an HSI space, decomposing the image by using a two-sided filter, processing by an improved Retinex model algorithm, and obtaining a new image saturation by performing saturation compensation; and 3, performing fusion display on the image acquired after denoising the YCbCr space and the image acquired by HSI space processing. By the method, the color constancy of the image can be kept, the dynamic range of the image can be improved well, and simultaneously noises of the image can be inhibited and removed with less texture and detail information loss.

The invention discloses a mura defect detection method based on sample learning and human visual characteristics, which belongs to the TFT-LCD display defect detection field. According to the invention, the method comprises the following steps: firstly, utilizing the Gaussian filter smoothing and Hough transform rectangle to preprocess the TFT-LCD display image, removing a large amount of noise and segmenting the image areas to be detected; then, using the PCA algorithm to conduct learning to a large amount of defect-free samples; automatically extracting the differential characteristics between the background and the target and re-constructing a background image; and then, thresholding the differential characteristics between a testing image and the background; through the reconstructing of the background and the threshold calculating, jointly creating a model. According to the invention, based on the training sample learning, a relationship model between the background structure information and the threshold value is established; and a self-adaptive segmentation algorithm based on human visual characteristics is proposed. The main purpose of the invention is to detect different mura defects in a TFT-LCD, to raise the qualification rate and to increase accuracy for the detection of mura defects.

In order to prevent double-shadow and cavity phenomena from occurring in a target and rapidly obtain a complete outline of a moving target in a video sequence, the invention provides a moving object rapid detection method based on the video sequence. The method comprises the steps of denoising the video image sequence through a gaussian filter and solving an inter-frame differential image of any three frames of filtered adjacent video images through differential operation; carrying out iteration update on an initial background image through utilization of an inter-frame differential binary image and extracting a background image corresponding to a current frame; reestablishing a reference image corresponding to the current frame of image according to an inter-frame differential result of the three frames of adjacent video frames, thereby obtaining an inter-frame differential target detection image, and obtaining a moving target outline difference value image of the current frame through a background differential method; and combining target images extracted by a three-inter-frame differential method and a background differential method through an OR operation, and outputting a final moving target image. According to the method, noise interference can be effectively removed and the complete moving target can be detected rapidly and accurately.

The invention discloses a crowd counting method based on a coding-decoding structure multi-scale convolutional neural network, and the method is characterized in that the method comprises the following steps: considering the visual angle distortion of an image, and employing an adaptive Gaussian filter to calculate a true value density map of the image; adopting an encoding-decoding structure, building a multi-scale convolutional neural network model, wherein a loss function of the multi-scale convolutional neural network model comprises pixel space loss and counting error description; training and testing the built multi-scale convolutional neural network model to obtain a trained multi-scale convolutional neural network model; inputting a to-be-estimated image into the trained multi-scale convolutional neural network model, and predicting to obtain a crowd density map; and performing regression estimation on the crowd density map to obtain the number of people in the to-be-estimatedimage. According to the method, scale features and contextual information of the image can be reserved, and the output quality of the density map can be improved.

The invention discloses a method for improving an optical scanning holographic tomography effect, and belongs to the field of optical tomography. The method can solve the defect in the conventional optical scanning technology that restored section images have louder out-of-focus noise, in real time. The method comprises the following steps: utilizing a random phase in the optics encryption technique, changing a certain pupil function in a conventional optical scanning holographic system into a stochastic phase pupil function, and equaling restore of out-of-focus layer images into decryption under the condition of mistakenly decrypting a secret key, thereby enabling the out-of-focus layer images overlaid on the restored section images to be gaussian white noise with statistical independence and greatly reducing the influence of the out-of-focus noise on the focus layer images. Meanwhile the out-of-focus noise can also be filtered through design of a gaussian filter, so that the longitudinal resolution of the system can be improved; besides, through the adoption of the method, the bandwidth of the optical transfer function is wider, and the transverse resolution of the restored section images can be higher.

The invention discloses a face identification method based on multiscale weber local descriptor and kernel group sparse representation. The face identification method comprises the following steps: firstly normalizing the size of face images and smoothing the images by utilizing a gaussian filter; extracting differential excitation ingredients of the multiscale weber local descriptor of the images and extracting direction information by utilizing an Sobel operator; extracting the multiscale weber local descriptor of the face images according to the multiscale differential excitation and the direction information and mapping the multiscale weber local descriptor to a kernel space by utilizing a histogram intersection kernel; then with a kernel matrix obtained by a training sample as a sparse dictionary, calculating group sparse representation coefficients of a kernel vector obtained by a test sample; and finally reconstructing a multiscale weber local descriptor vector of the test sample according to the group sparse representation coefficients and distinguishing the test sample by utilizing the minimum reconstruction error. According to the face identification method, the multiscale weber local descriptor and the kernel group sparse representation algorithm are fused for face identification, and the identification accuracy rate is greatly improved.

The invention relates to a method for detecting and positioning the leakage of an oil-gas pipeline by utilizing an autonomous navigation robot. The method comprises a robot pipeline line-tracking detection method and a terminal computer detection positioning method; marking lines and correcting nodes are painted on the road surface near the pipeline along the axial direction of the pipeline; the robot navigation system adopts the photoelectric coding and correcting node identification method for navigation and line-tracking detection; a data acquisition combiner is formed by an ultrasonic sensor and a smell sensor to acquire a signal of the oil-gas pipeline, the signal passes an amplifying-filtering circuit and a control module, undergoes analog/digital conversion, modulated by a wireless communication module and sent to the computer; analytic software of the system adopts an algorithm combining the adaptive fuzzy noise detection smoothing algorithm and the multi-scale Gaussian filter algorithm to process data of ultrasonic strength and gas strength; and when the two signals are received at the same time and both higher than the threshold values, the leakage of the pipeline can be determined, and the alarm rings and the leakage can be positioned. The method is applicable for the line-tracking detection and leakage positioning of the oil-gas pipeline in an underground comprehensive pipe ditch.

The invention discloses a full-waveform laser radar echo data gaussian decomposition method based on a genetic algorithm. The method is mainly formed by five steps including echo denoising, initial wave crest extracting, initial parameter estimating, waveform fitting and precision judging; when the precision of initial fitting meets the requirements, second fitting is not carried out, when the precision of the initial fitting does not meet the requirements, iteration fitting is carried out, and in each iteration, the fitting precision is improved by a method of increasing a gaussian component until a stopping condition is met; a gaussian function is used for fitting echo data of a full-waveform laser radar; and a gaussian filter method is adopted to filter noise of the echo data of the full-wave form laser radar, the number of initial gaussian components is confirmed by extracting a peak value point, an initial parameter range is estimated, fitting and precision adjusting are carried out on the echo data of the full-waveform laser through the genetic algorithm, an echo gaussian decomposition software system based on the genetic algorithm is manufactured, and echo decomposition operation is simplified.

The invention discloses a three-dimensional medical image segmentation method. The method comprises the following steps: firstly performing three-dimensional superposition reconstruction on sequential two-dimensional medical image, and then performing Gaussian filter smoothing in a three-dimensional space; carrying out three-dimensional space contraction computation on the three-dimensional medical image, and calculating and recording the contraction orders and the energy characteristics of every three-dimensional pixel point; performing three-dimensional space extension on the three-dimensional medical image in a plurality of areas by a level set method; and finally extracting and segmenting the three-dimensional pixel points in the three-dimensional image according to information of interested areas, and respectively generating a plurality of three-dimensional images of the interested areas. The method has positive effects of strong timeliness, high computing efficiency, high segmentation accuracy, and clear and reliable image local features, and can achieve simultaneous segmentation of a plurality of areas.

The invention relates to a cloth defect detection method based on Fourier transform and image morphology. The cloth defect detection method comprises the following steps: (1) collecting an original image of cloth by using a linear array camera; (2) preprocessing the original image; (3) extracting image texture features by using LAWS texture filtering; (4) adopting a GMM classifier model to carry out defect pre-judgment; (5) constructing a band elimination filter by using a Gaussian filter; (6) generating a defect image by using Fourier transform and inverse transform; (7) extracting flaw positions and areas by adopting image morphology; And (8) outputting a cloth detection result. According to the method, more than ten types of cloth flaws including broken wefts, broken warps, broken holes, floating wefts and the like can be detected in real time, the detection speed is high, the accuracy rate is high, and the detection robustness is enhanced through the image processing method integrating the airspace and the frequency domain.

A method and apparatus detects one or more spiculated masses in an image using a processor. The image is received in the processor. The received image is filtered using one or more Gaussian filters to detect one or more central mass regions. The received image is also filtered using one or more spiculated lesion filters to detect where the one or more spiculated masses converge. In addition, the received image is filtered using one or more Difference-of-Gaussian filters to suppress one or more linear structures. An enhanced image showing the detected spiculated masses is created by combining an output from all of the filtering steps. The enhanced image is then provided to an output of the processor.

The invention relates to a motion detection method based on edge detection and frame difference. The method comprises the following steps of: 1, acquiring an image sequence and determining images of a reference frame and a current frame; 2, performing Gaussian filtering on the images of the reference frame and the current frame; 3, extracting the edge information of the reference frame and the current frame after the filtering; 4, acquiring converted images of the reference frame and the current frame; 5, performing the frame difference on the converted images of the reference frame and the current frame, and determining a motion region by using a double threshold value method; and 6, performing image post-processing on the motion region, and determining a motion object. The method has the advantages that: due to the adoption of the thinking of frame difference, the method is simple and practical, the complex computations of an optical flow method and hybrid Gaussian are avoided, computing time is only 1/3 of that of classical hybrid Gaussian, and requirements on the real-time performance of intelligent monitoring are met; incomplete frame difference detection results are improvedto a certain extent by the combination of the edge detection and the frame difference and a series of additional processing; and compared with the hybrid Gaussian serving as a mainstream method, the method is difficult to influence by illumination and external interference.

The invention discloses a machine vision image characteristic point detection and matching combination optimization method. The method mainly comprises the steps of firstly obtaining a template image and a search image, and splicing the template image and the search image into a workpiece image; secondly performing characteristic point detection on the workpiece image to obtain P characteristic points; thirdly performing characteristic point description on the P characteristic points separately, namely, selecting any characteristic point as a center for constructing a pixel block image, and performing Gaussian filtering on sampling points contained in the pixel block image separately to obtain a sampling point pair corresponding to the characteristic point so as to obtain sampling point pairs corresponding to the P characteristic points and sampling point pair distances corresponding to the P characteristic points; and fourthly obtaining corresponding aggregated model directions of long-distance sampling point pairs and corresponding binary descriptors of short-distance sampling point pairs, performing matching identification on the P characteristic points separately, calculating an affine transformation parameter between the template image and the search image, obtaining three-dimensional coordinates of a target workpiece in the search image, and performing accurate capture.

The invention relates to a deformation medical image registration method based on PCA and diffeomorphism Demons, and belongs to the technical field of medical image processing. According to the method, preprocessing is carried out on collected images, a multiresolution mechanism is then adopted, registration circulation is carried out under set iterations, a deformation field of movable images is calculated, smoothing is carried out through a Gaussian filter, interpolation is carried out on the movable images in an index domain by utilizing an obtained deformation field matrix, key features of the images are extracted by using PCA, and similarity measure between two images is calculated by combining SSD, Pearson, Spearman and Kendall. Compared with a traditional demons method, a PCA diffeomorphism log demons related registration method reduces calculated quantity, can restrain noise quite well, and is higher in robustness than a method that SSD similarity measure is merely adopted, the convergence speed is increased.

The invention, which belongs to the technical field of image processing, discloses a video segmentation method based on strong target constraint video saliency. According to the method disclosed by the invention, strong target constraint is introduced based on image saliency. The location and scale constraint of a target are obtained by a multi-scale tracking algorithm and optical flow correction,color constraint information of the target is obtained based on a historical frame segmentation result, and calculation is carried out obtain a video saliency result; histogram classification is carried out on the video saliency result to obtain a tag mask graph, and foreground/background prior probability models of a current frame are calculated; a super-pixel-based time-space continuum full connection condition random field model is constructed at the current frame, data items are defined by using the prior probability models, an intra-frame smooth item and an inter-frame smooth item are defined by combining color distances, space distances and edge relationships between super pixels, and optimized solution is carried out by using a fast high-dimensional Gaussian filter algorithm to complete video target segmentation. Therefore, the accuracy and the time efficiency of video segmentation are improved.

The invention provides a video moving target classification and identification method based on outline constraint. The video moving target classification and identification method includes the steps: (1) obtaining a realistic target region and a target outline through a level set partitioning algorithm which is based on color features, textural features and shape prior constraint; (2) conducting convolution operation on the realistic target region through Gaussian filter and obtaining space detail constituent of the target; (3) extracting a local binary pattern histogram of the space detail constituent and obtaining the textural features of the target; (4) extracting a directional gradient histogram of an outline constraint local region in the realistic target region and obtaining the edge gradient features of the target; (5) extracting the texture features and the edge gradient features of a training sample target, training the texture features and the edge gradient features of the training sample target through a machine learning method, obtaining a target classification model; and (6) extracting the texture features and the edge gradient features of a to-be-identified target, inputting the classification model and confirming the type of the target. By means of the video moving target classification and identification method based on outline constraint, classification accuracy under complex outdoor conditions is improved.

The invention provides a Gaussian filter-based real-time defogging method for a single image. The method is established on the basis of a physical model and can realize real-scene recovery of a degraded quality image. The method comprises the following four main steps of: step 1, calculating, inputting a dark channel image of a fogged image and estimating an overall atmosphere light value; steps 2, primarily estimating an atmosphere dissipative function; step 3, adopting Gaussian filtering to refine the atmosphere dissipative function; and step 4, recovering a scene radiancy. With the adoption of the method, defects that an existing defogging method is slower in defogging speed and cannot be applied to a real-time processing system are overcome; and the method is good in effect of an obtained defogged image and rapid in processing speed, can be applied to a real-time system, and has a better practical value and a wide application background.

The invention provides an automatic identification method and system of a pavement disease image. The method comprises: a shot pavement image is preprocessed, wherein preprocessing includes Gamma gray scale correction, Gaussian filter enhancement and local adaptive binarization; edge detection is carried out on a binarized image; connected domain contour detection is carried out on the image after edge detection to obtain the number of the connected domains and external rectangles of connected domain contours; according to the shapes of the external rectangles of the connected domain contours, a region of a pavement crack is determined; on the basis of position information of the region of the pavement crack, a crack region image is extracted from the image after edge detection and a black template is superposed to form a crack feature image with the size identical with that of the original image; and on the basis of a convolution neural network, the crack feature image is classified. Compared with the prior art, the method and system have the following advantages: the detection efficiency of crack positioning in the pavement disease is high; and for crack images with different features, robustness is high.

The invention relates to the field of image enhancement and discloses a Retinex-theory-based nonlinear image enhancement method and system. According to the method and the system, a gradient-domain-transformation-based edge-preserving filter manner is used for replacing the conventional Gaussian filter manner to obtain illumination components. On the one hand, the phenomenon of halo artifact near a high-contrast edge is effectively overcome; on the other hand, the edge-preserving filter of color images can be carried out in real time. In addition, a nonlinear compression function is constructed on account of the illumination components, a highlight area is suppressed and a dark area is effectively improved; the phenomenon of overexposure of the images is avoided; the image luminance is mapped to an ideal dynamic range; a self-adaptive local nonlinear enhancement manner is used for enhancing image details, so that the local details of reflection components are enhanced, the quality of the images is well improved; the enhanced images are closer to real natural sites; the details are more clear.