89 results about "Morphological gradient" patented technology

The invention discloses a human body motion feature extraction method based on a global remarkable edge area, comprising steps of using a contrast between an area and a whole image to calculate the significance, reducing the color quantity of the color space, smoothing the significance of the color space, calculating the significance area according to the space relation of the neighboring areas, performing morphology gradient changing on the foreground area segmented by a binarized threshold to generate a global remarkable edge area, traversing strong corner points of all grids of the video frames under various sizes, collecting key characteristic points, the light stream amplitude value of which is not 0, in the remarkable edge area, solving the displacement of the strong corner point according to the corrected light stream field, and forming the human body motion local time space characteristic by using the strong corner point continuous multi-frame displacement locus and the neighbourhood gradient vector. The invention extracts the motion characteristics through global remarkable edge area, eliminates the background noise points irrelevant to the human body motion, removes the affect on the light stream calculation by the camera motion, improves the accuracy of the human body motion local time space characteristic description and improves the human body motion recognition rate.

The invention relates to a multi-focus image fusion method based on contrast pyramid transformation. Firstly preprocessing and image enhancement are performed on an input left focus source image and a right focus source image; then information entropy of the obtained enhanced images is solved and the weighted fusion coefficient is determined so that a primary fusion image is obtained; then contrast pyramid decomposition is performed on the two enhanced images and the primary fusion image so that decomposition layer sequence images are obtained; the weighted fusion coefficient is determined for the top layer image according to the pixel information difference between the images, and the fusion rule of taking the maximum pixel absolute value is performed on other layers of images; image reconstruction is performed on the obtained fusion image sequences so that the fusion image is obtained; and finally morphological gradient processing is performed on the fusion image and the two enhanced images, and secondary fusion is performed on the obtained images so that the final fusion result image is obtained. Fusion of the multi-focus images can be performed, and the fusion result image has the characteristics of being high in contrast, comprehensive and abundant in information, outstanding in edge and clear in details.

The invention discloses a method for judging a single-phase grounding fault of a small-current grounding system, which is characterized in that single-phase grounding fault judgment is started according to a sudden change in electric field intensity and a sudden change in phase current, the fault starting time is accurately determined by using a morphological gradient method, and accurate synchronization of three-phase current is realized; filtering is performed on transient current signals so as to determine the duration time of a transient process, and the characteristic frequency of the transient current is calculated by using an autocorrelation method; a grounding fault is positioned according to the correlation of transient current of a fault phase and transient current of a non-fault phase, and detection for a single-phase grounding fault of a small-current grounding system applied to a fault indicator is realized. The method disclosed by the invention is not affected by the operating mode, the grounding site, the grounding resistance or the like, and on-site tuning and debugging are not required compared with a threshold comparison method adopted by the existing fault indicator, so that the fault judgment accuracy can be greatly improved, a problem of low false alarm and alarm omission rate of the existing fault indicator is solved, and the feeder automation level is greatly improved.

The invention discloses a morphological gradient-based identification method for the magnetizing inrush current of a transformer. The identification method comprises the following steps of: (1) acquiring a differential current signal I of current transformers arranged at two differential protection sides of the transformer; (2) sampling the differential current signal I; and (3) when the obtained differential current signal I exceeds a setting valve Icd of a differential-protection starting component, distinguishing the magnetizing inrush current. The step (3) specifically comprises the following steps of: (3-1) taking the differential current signal I of which a data window is a half of the length of a base frequency period to obtain a morphological gradient result of the I according to an improved mathematical morphology gradient algorithm; (3-2) introducing criterion sigma: determining that I is the magnetizing inrush current if sigma is greater than sigma set, and carrying out a step (3-3) otherwise; and (3-3) introducing criterion C: determining that I is the magnetizing inrush current if C is greater than C set, and determining that I is an inner fault signal otherwise. The method provided by the invention has the advantages of simple steps, small calculated quantity, short time delay, high reliability and the like, so that the magnetizing inrush current can be identified to realize the locking differential protection.

The invention discloses a multi-mode medical image registration method based on edge features and CS (Cauchy-Schwarz) information. Based on the principle of a generalized morphological filter, the method disclosed by invention improves the morphological gradient operator and performs image edge detection by use of a new gradient filtering operator so as to improve the computing speed and anti-noise performance of the image edge detection and quickly realize rough detection of the image edge, and determines the coordinates and direction of the image edge point to lay a foundation for the subsequent image registration; the method constructs a new CS geometrical information measure according to a CS inequality, wherein the measuring method does not involve division and logarithm arithmetic in the mathematical computation process, and can effectively reduce the computing complexity in image registration and improve the computing efficiency; and moreover, the method adopts a Powell method to optimize registration parameters, does not need to derivate a target function, has quadratic convergence and high convergence speed, and is conveniently realized by programming. Compared with the prior art, the method disclosed by the invention has better performance in computing standard deviation, information amount, average computing time and the like, and has better robustness.

The invention discloses a cotton fault detecting and identifying method based on a multi-spectrum technology. The method comprises the steps: (1) image acquiring; (2) image pre-processing; (3) the preliminary detection of faults; carrying out edge detection on an improved morphological gradient algorithm to obtain the comprehensive edge detection intensities of all faults to obtain an image with contrast enhancement of all faults; (4) precise location of faults: converting the image into a connected domain image through an improved iterative threshold and morphological processing, precisely locating the faults by selecting a main fault algorithm, and extracting fault areas by marking; (5) cotton ginning quality judgment: carrying out statistics on the amount of all faults, and calculating the total cotton faults to obtain a cotton ginning quality grade. The cotton fault detecting and identifying method provided by the invention has the advantages of being simple in theory, fast in detection speed and high in detection precision.

The invention relates to the field of image processing, and discloses a quick high-resolution remote sensing image segmentation method. The method comprises steps: S1, the high-resolution remote sensing image is read; S2, a multiband morphological gradient for the read high-resolution remote sensing image is calculated to obtain a multiband morphological gradient image formed by the multiband morphological gradient; S3, morphological rebuilding is carried out on the multiband morphological gradient image to obtain a gradient image after morphological rebuilding; S4, watershed segmentation is carried out on the gradient image after morphological rebuilding to obtain an image after watershed segmentation; and S5, region merging is carried out on the image after watershed segmentation. The remote sensing image with mass data can be quickly segmented, and over segmentation can be effectively reduced.

The invention provides an image retrieval method based on multi-feature fusion. The image retrieval method is used for solving the problem that an image retrieval method based on a single feature cannot meet the query requirement of a user. The method comprises the steps of performing noise reduction processing on a to-be-retrieved image by utilizing a filtering method; performing feature quantification by utilizing the improved HSV color space to extract global features of the to-be-retrieved image; performing multi-scale morphological gradient extraction on the denoised image to extract local features of the to-be-retrieved image; performing adaptive fusion on the global features and the local features to obtain an adaptive fusion image; carrying out hash coding on the self-adaptive fusion image, calculating the similarity between the to-be-retrieved image and all the images in the database through Hash codes, and selecting the first several images with the highest similarity with the to-be-retrieved image as retrieval results of the to-be-retrieved image. According to the method, the feature points of the image are fully extracted, and the edge information of the image is protected more comprehensively in the local feature extraction process, so that the retrieval accuracy is improved, and the retrieval time is shortened.

The invention relates to the technical field of image collection and recognition, and specifically relates to an automatic measurement method for morphological parameters of a prawn based on image recognition and a cascade classifier. The automatic measurement method comprises the steps of S1, collecting a plurality of prawn color images to serve as samples; S2, training the samples, and then classifying the trained samples to obtain an LBP feature based cascade strong classifier; S3, collecting a prawn color image A to be measured, and performing normalization correction on the prawn color image A to obtain a normalized image B; S4, performing image segmentation on the normalized image to obtain a segmented image B; S5, performing ruler measurement on a segmented image C obtained in the step S4 to obtain a measurement image D; S6, performing image correction on the measurement image D obtained in the step S5 to obtain a corrected image E; and S7, recognizing the corrected image E by using the cascade strong classifier, and calculating morphological parameters of the prawn to be measured. According to the algorithm, accurate measurement for the morphological parameters of the prawncan be accurately implemented, and the measurement accuracy and the measurement efficiency can be improved.

The invention relates to a paper defect classification method based on multi-scale morphology combined with a convolutional neural network. Morphological operation is performed on original images by using structural elements different in scale; multi-scale morphological gradient images are obtained through weighted fusion of gradient images under different scales. In order to increase paper defectcontrast and highlight defect gradient features and information features of defect edges, the multi-scale morphological gradient images and the original defect images are subjected to weighted fusionto achieve defect image enhancement; the images are input to convolutional neural network for feature extraction and classification, and the convolutional neural network is used in classification ofpaper defects so that accurate classification of the paper defects can be rapidly achieved. The method has the advantages of being simple, short in time consumption and high in recognition precision.

The invention discloses a method for detecting the consistency of different-source images, belonging to the technical field of computer image processing. The method comprises the following steps of: extracting contour images of the different-source images by adopting a two-dimensional cellular automaton model based on a morphological gradient method; acquiring characteristic points and descriptors of accelerating robust characteristics of the contour images; acquiring an initial matching dot pair set from the characteristic points of the accelerating robust characteristics; and screening out an accurately matched dot pair from the initial matching dot pair set. By using the method disclosed by the invention, the problem that the consistency cannot be directly detected from the different-source images by using a SURF algorithm is solved, and the detection speed and the accuracy rate are increased.

The invention discloses an image processing method which comprises road vanishing point detection steps. The road vanishing point detection steps comprise the steps that morphology gradient processing is conducted on an image to obtain a first intermediate image; binarization processing is conducted on the first intermediate image to obtain a second intermediate image; morphology processing is conducted on the second intermediate image to obtain a third intermediate image; pixel value cumsum is conducted on each row of the third intermediate image to obtain the sum of row pixel values, and the row with the maximum sum of the row pixel values serves as the ground line position of the image; edge detection is conducted on the portion, located below the ground line position, of the first intermediate image to obtain a fourth intermediate image; line detection is conducted on the fourth intermediate image to obtain a fifth intermediate image, and lines in the fifth intermediate image are obtained; a road vanishing point is determined from intersection points. By means of the image processing method, good road recognition and detection of targets in front of a vehicle can be achieved.

The invention discloses an image edge detection method. The method includes the steps of filtering out noise in spatial target images, calculating the morphological gradient of the images in a single scale, obtaining multi-scale morphological gradient images through calculation according to the morphological gradient of the images in the single scale, determining the position of a maximum value point in the multi-scale morphological gradient images, selecting a corresponding morphological gradient edge through a zero-crossing point position template, removing false edges according to a preset threshold, reserving the trunk of a target edge, removing fragments in the edge, and obtaining edges of the special target images. By means of the image edge detection method, edge information of the spatial target images can be effectively extracted.

An image fusion method based on contrast analysis comprises the following steps of: 1, constructing the morphological gradient graph of an input image and the morphological gradient graph of an enhanced image of the input image separately; 2, subjecting the morphological gradient graphs of the input image and the enhanced image thereof to subtraction to obtain a contrast difference graph of the unenhanced and enhanced input images; 3, creating a fusion coefficient matrix according to the contrast difference graph; and 4, fusing the input image and the enhanced image thereof according to the fusion coefficient matrix, subjecting fused processed image pixels to range control in order that the pixels are within a reasonable image range and are output. The single-graph fusion of the present invention aims at the fusion of the input image and the enhanced image thereof which have the same image information essence and is not fusion of different images. The image output by the image fusion method has uniform grayscale distribution and has completely retained details. The method is effective on not only grayscale images, but also color images, and is higher in fidelity of the restoration of the color images in associated color channels.

The invention relates to a method for carrying out inversion on urban area building sedimentation on the basis of a morphological gradient. The method considers that a building shows strong reflection characteristics and strong conherence in a high-resolution COSMO SkyMED radar image. The method comprises the following steps: firstly, through a correlation coefficient threshold value method based on spectrum statistical properties and an amplitude threshold value method based on back scattering intensity, a high-quality candidate point target is extracted from a high-resolution SAR (Synthetic Aperture Radar) image, and then, a SBAS (Small Baseline Subsets) method is adopted to carry out calculation to obtain the sedimentation information of the point target; an elevation threshold value is set for a result obtained by InSAR (Interferometric Synthetic Aperture Radar) calculation, and the point target on a building is roughly selected; and a morphological gradient method is used for extracting a building boundary in the SAR image, the extracted building boundary is used for removing non building points in the roughly selected point target, and finally, the automatic extraction of the building point target is finally realized. By use of the method, the identification accuracy and reliability of building sedimentation information can be improved.

The invention provides a bulk cargo yard stacker-reclaimer positioning method based on point cloud data processing technology and a system for realizing same. The method comprises the steps of acquiring the point cloud data of the bulk cargo yard, performing gridding on the point cloud data, representing elevation values of the grids by means of gray values for obtaining an elevation diaphragm, performing mathematical morphological operation on the elevation diaphragm, calculating a difference between expansion operation and corrosion operation for obtaining mathematical morphological gradient data; performing binaryzation, performing mathematical morphological open-and-close operation for obtaining position information, and realizing a purpose of positioning the stacker-reclaimer in the bulk cargo yard. The method of the invention has advantages of accurately and quickly extracting position of the stacker-reclaimer in the bulk cargo yard, realizing a purpose of positioning the stacker-reclaimer in the bulk cargo yard, realizing no requirement for artificial observation, and obtaining high accuracy.

The invention provides an image edge extraction method. According to the method, a two-dimensional nonseparable lowpass filter is constructed; first an original image is subjected to multi-scale nonseparable additive wavelet decomposition; then a gradient map of decomposed sub-images is obtained by utilizing a modern mathematical morphology gradient operator; high-frequency sub-images are subjected to addition, so that the edges of low-frequency and high-frequency sub-images are extracted; and a final edge image is formed again. A new image edge extraction method is provided by utilizing characteristics of richness, quickness and the like of two-dimensional nonseparable additive wavelet isotropy and morphology gradient detection of the edges. The method is simple in operation and good in detection effect; a target with a relatively small gray value in the image can be detected, more edge details can be obtained, and the edges with the integrity, multi-directionality, continuity and translation invariance can be extracted. Compared with other detection methods, the method has a relatively high processing speed.

The embodiment of the invention provides an image recognition method and device, and relates to the technical field of image recognition, and the method comprises the steps: carrying out the morphological gradient calculation of a to-be-recognized image, and obtaining a first gradient map; determining an area corresponding to the image area where the characters in the to-be-identified image are located in the first gradient image, and taking the area as a first image area; determining the number of characters in the first image area; determining a first grouping mode of the characters in the first image area based on the number of the characters; Based on the first grouping mode, performing character segmentation on the first image region to obtain a single character region; and carrying out character recognition on each single character region to obtain a character recognition result of the to-be-recognized image. When the scheme provided by the embodiment of the invention is used foridentifying the image, the anti-interference performance and the accuracy of image identification can be improved.

The invention provides a multispectral micro-imaging field depth extension method and system. The method comprises the following steps: setting the gray value of each pixel point as a medium value of gray values of all pixel points in a specific neighborhood window of the pixel point, and carrying out noise filtering on a microscopic image; comparing difference between an original image obtained after expansion and an original image obtained after corrosion to obtain a morphological gradient; on the basis of the precondition that a neighborhood region of a clear area is clear, expanding the clear area through a morphological expansion method; carrying out expansion on the image and then, carrying out corrosion, and eliminating isolated points or an isolated point group lower than high bright point neighborhood points through a morphological close operation to obtain edge images; carrying out vertical comparison on the gray values of the obtained edge images to construct a height graph; and giving RGB values of the pixel points of the original image to a new composite image through the height graph to generate a final all-depth-of-field-fused image. Axial imaging range of the microscope is expanded without artificial interference; and the multispectral micro-imaging field depth extension method and system are high in accuracy, fast in speed and high in practicability.

The invention discloses a novel distributed feed line automatic protection method applied to an FTU. The method includes the steps: accurately determining a fault moment through the zero sequence current of a feed line; calculating transient fault components of a phase current and the zero sequence current; performing four-layer wavelet decomposition for the transient fault components of the phase current and then obtaining a fault phase; performing four-layer wavelet decomposition for the transient fault components of zero sequence current, and determining whether grounding protection is started according to the ratio of energy of free vibration components of the transient fault components of zero sequence current; determining the ration of the transient capacitive current total energy of the fault phase to the average transient capacitive current total energy of non-fault phases, and determining whether the grounding protection is started; and determining whether fault phase grounding protection is started according to a cross correlation coefficient of an autocorrelation estimation sequence of the fault phase and the two non-fault phases. A morphological gradient technology is adopted to process the zero sequence current, extract current abrupt change information and accurately capture the moment when a fault occurs. Through the method, people can directly and accurately identify whether a feed line malfunctions on three-phase current information collected by an FTU.

The present invention discloses a road semi-automatic extraction method and system based on circular templates. The method comprises seven steps of filter pre-processing of road-containing images, acquisition of morphological gradient images, generation of an optimal circular template, acquisition of a reference template gray value, search of to-be-matched templates, encryption of road central points and generation of road center lines. The system comprises an image pro-processing module, a morphological gradient image acquisition module, a circular template generation module, a reference template gray value acquisition module, a matching template searching module and a center line generation module. By combining geometrical characteristics of a road, a multi-direction morphological filter method is adopted to pre-process the images and smooth noises of the road, facilitating follow-up extraction; the circular template is selected at the same time and can well fit the road; the searching speed of the to-be-matched templates is high; and a smooth center line is good in effect by polyline fitting.

The invention discloses a direct current transmission line protection method based on mathematical morphological gradient and mathematical morphological entropy, which comprises the following steps: 1) acquiring a direct current line voltage signal and carrying out line-mode conversion on the voltage signal; 2) constructing an improved morphological filter, filtering out noise singular points in aline mode voltage signal, performing multi-scale morphological gradient processing, and constructing a fault starting criterion according to the processing result; 3) carrying out multi-scale mathematical morphological gradient processing on the fault signals, constructing mathematical morphological entropy, a fault criterion of an inner region and an outer region and a fault criterion of an anode and a cathode by utilizing mathematical morphological gradient processing results under different scales. The direct current transmission line protection method based on mathematical morphological gradient and mathematical morphological entropy can quickly filter out the noise singular points in the signal, and quickly and accurately achieve fault identification through multi-scale mathematicalmorphological gradient and morphological entropy.

The invention discloses a vibration signal-based mechanical link looseness detection and early warning method. According to the method, the looseness condition of a mechanical link piece is judged bycollecting the vibration signals of the vibration device of the mechanical link piece, and then the early-warning information is given. The method comprises the following steps: step 1, constructing amechanical link looseness detection and early warning system based on vibration signals; step 2, acquiring the vibration signal of the mechanical link piece; step 3, processing the vibration signal to obtain the morphological gradient spectral entropy of the vibration signal, and then judging the operation state of the mechanical link piece; step 4, carrying out the performance degradation analysis of the mechanical link piece according to the information in the preceding steps, comprehensively evaluating the health condition of the mechanical link piece, and giving the early-warning information. The method provided by the invention is used for carrying out real-time and on-line monitoring and early warning analysis on the running state of the mechanical link. Therefore, the reliability and the safety of the operation of the mechanical link piece are improved. The better implementation performance is achieved.

The invention discloses a business license information extraction method based on morphological gradient multi-scale fusion. The business license information extraction method comprises the followingsteps: sequentially carrying out scaling processing of multi-scale transformation, denoising of a bilateral filtering algorithm and graying processing on a business license original image to obtain agrayscale image; performing black cap operation on the gray-scale image and negation according to pixels to obtain a black cap image; subtracting the gray-scale image from the black-cap image according to pixels to obtain an edge enhancement image; performing open operation on the edge enhancement graph to obtain an open operation graph; detecting the open operation diagram by using a maximum extreme value stabilization algorithm to obtain a candidate region diagram; detecting and de-duplicating the candidate area map to obtain text area maps with different scales; and comparing the text areagraphs under different scales, and keeping the text area detected under each scale, thereby extracting the image text information. The method solves the technical problem that the accuracy, the hardware cost and the calculation cost of the existing image processing information in the aspect of extracting the business license information cannot be considered at the same time, and greatly improves the administrative efficiency.

The invention discloses a rolling bearing fault damage degree identification method comprising the steps of fault vibration signal acquisition, calculating the mathematical morphological gradient spectral value, calculating the change rate of the mathematical morphological gradient spectral value, determining the optimal scale range of the structural elements, calculating the high-order differencemathematical morphological gradient spectral value, calculating the high-order difference mathematical morphological gradient spectral entropy, defining the fault damage degree of discrimination, calculating the fault damage degree of discrimination and judging the fault damage degree. The damage degree of the bearing inner race fault can be effectively identified so that the method has high correct rate of damage degree identification and can greatly enhance the identification efficiency and is an effective fault degree quantitative identification method, and a new method can be provided forrotating machinery fault damage degree identification and fault prediction and the method is great in practicality and worthy of popularization.

The invention provides an irregular beverage bottle stacking image edge detection method based on an improved Roberts operator. The method comprises the following steps of S1, performing filtering, edge information retaining and denoising on an irregular beverage bottle stacking image by utilizing a morphological smoothing algorithm; S2, calculating a morphological gradient of the image under a single scale; performing calculation to obtain a multi-scale morphological gradient image according to the morphological gradient of the image under the single scale; determining the position of a maximum value point in the multi-scale morphological gradient image; selecting a corresponding morphological gradient edge by using a zero crossing point position template, and removing a pseudo edge according to a preset threshold value; and S3, carrying out edge detection on the multi-scale morphological gradient image by adopting an improved Roberts edge detection operator to obtain edge informationof the irregular beverage bottle stacking image. A communication base station can improve the spectrum distribution rate and the transmission efficiency.

The invention provides an improved Prewitt operator-based fire complex environment image detecting method. The method comprises the steps of: S1, filtering an analog fire complex environment image byutilizing a morphological smoothing algorithm, preserving edge information and removing noise; S2, calculating single-scale morphological gradient and multi-scale morphological gradient of the image;and S3, carrying out edge detection on the multi-scale morphological gradient image by adopting an improved Prewitt edge detection operator to obtain the edge information of the analog fire complex environment image.

The invention discloses a method for segmenting a video object based on fixed period regional compensation, comprising the following steps: firstly, noise suppression and dilation connection are carried out on a frame difference image so as to obtain a change testing module; secondly, open and close reconstruction simplification is carried out on an original image so as to obtain a morphology gradient; thirdly, non-linear transformation and gradient level division are carried out on the morphology gradient image so as to finally obtain an exact boundary which is obtained in a watershed way, and the initiate binaryzation module of the video object is extracted by a scaling; and finally, a final complete video object module is obtained by the fixed period regional compensation. The method for segmenting the video object has better adaptability to various testing video sequences, and can eliminate or reduce a local deletion phenomenon which easily happens when the video object slowly moves or the internal texture of the object has higher consistency.