157 results about "Variance method" patented technology

The invention relates to the field of image processing and computer vision technologies, in particular to an extraction method for characters in a form document image. The extraction method includes a first step of extracting line segments in the image through edge detection and Hough transformation algorithm, a second step of estimating an inclined angle of the whole image according to direction distribution of the line segments and carrying out inclination correction on the image, a third step of connecting the line segments in the horizontal direction and the perpendicular direction, and locating table cells of a form, a fourth step of carrying out image binaryzation and segmenting a full line of characters in the table cells through a maximum between-cluster variance method, and extracting the characters in the table cells through a window sliding method, and a fifth step of carrying out restoration on deletion of strokes of the characters according to statistics features of frame lines of the table cells. The extraction method is good in flexibility and capable of effectively solving the problems of adhesion between the characters and overlap between the characters and form lines, and greatly reduces the influence of the adhesion and overlap on optical character recognition (OCR).

The invention relates to an empty bottle wall defect detection method, which comprises the following steps: shooting two images by using a camera before and after the same empty bottle is rotated for 90 degrees during advancing, then transmitting the images to a CPU of an industrial personal computer to perform analysis on the images, and judging whether the bottle wall has defect. Image processing comprises the following steps: A, scanning edge point pairs on a bottle neck, positioning the bottle wall, and dividing a detection area of the bottle wall to perform subarea processing; B, pre-processing image data in the positioned area by adopting a grey stretching method; C, partitioning the images by adopting a maximum between-cluster variance method to acquire target information; and D, performing connectivity analysis on the partitioned bottle wall images, extracting characteristic data of each defect, and judging whether each detected connected domain has real defect according to the mass center position, posture ratio and area characteristic of the connected domain. The invention discloses a detection device used for the detection method at the same time. The detection method and the detection device are easily applied in detection equipment of an industrial flow line so as to realize automatic high-speed accurate detection on the empty bottle wall defect.

The invention discloses a feature analysis-based video flame detecting method. For a video image acquired by a monitoring camera in a fixed visual field. the method comprises the following steps of: firstly, performing binary cutting of the image by maximum between-cluster variance method and acquiring a foreground region; secondly, removing noise points by morphological filtering, and counting a plurality of features of the foreground region, such as colour illumination, geometry, diffusion property, irregularity, sparkling frequency and the like; and finally, comprehensively judging the calculated features of the foreground region to judge whether the video image has flame or not. The binary cutting of the image by maximum between-cluster variance method can reduce the influence of a large amount of disturbances. At the same time, in order to overcome the defect of single or few discriminating features of the currently usually used flame detecting method, the method extracts the plurality of features of the foreground region acquired by the maximum between-cluster variance method to make a comprehensive judgment, so the method realizes quick and accurate judgment that whether the video image has flame or not and greatly reduces false alarm rate.

The invention discloses a method for extracting a text in a complex background image. The method comprises the steps of firstly, using the susan operator to detect and identify angular points in a source image, after removing isolated angular points, conducting integral projection transformation to cut out suspected text regions, and screening and removing non-text regions according to priori knowledge; then, judging the background complexity of a text region by the utilization of gray level jump information, when a background is judged to be complex, conducting color clustering on the text region by the utilization of the kmeans clustering algorithm, and determining the type which a text belongs to and extracting the text according to the color information at the position where the angular points are the densest; when the background is judged to be simple, conducting binaryzation on the image by the utilization of a largest-between-class variance method; finally, realizing accurate extraction of the text region. The text extraction method can locate the text region in the complex background image, and finally extracts characters after removing the background.

The invention relates to an electric car path optimal seeking method considering the range and charging station position. The method overcomes defects in the prior art. According to the technical scheme, a seek idea for clearing the optimal path through an electric car auxiliary line is provided; paths, mileages and time consumption functions by adopting starting and final positions as variables are built; according to the current status of electric cars, the electric cars are divided into four types; the potential charging station attributes in the paths are considered, a slow charging station and a rapid charging station are included and distinguished through a charging efficiency variance method, and the optimal path for four types of electric cars can be given according to the types of the electric cars through the seeking method.

The invention discloses a method for detecting diseases of crop leaves. The method comprises the following steps: acquiring a leaf image of a crop to be detected, uploading the leaf image to an on-line detection platform with a disease image automatic identification function and a professional diagnosis system function, performing scab image partitioning and identification on the leaf of the crop to be detected, outputting a detection result, and giving a control suggestion, wherein the scab image partitioning is as followings: converting an original image from a red, green and blue (RGB) model space to a horizontal situation index (HSI) space, respectively extracting an H component image and an I component image in the HIS space, and performing dynamic threshold value partitioning on the H component image by using a maximum between-cluster variance method to preliminarily obtain a scab region image; superimposing the I component image on the partitioning result of the H component image to eliminate misjudgment caused by a background region on the scab partitioning, thus obtaining a binary image only comprising the scab region; and performing subsequent treatment on the partitioning result by using a morphological method, and finally obtaining a complete scab image of the leaf of the crop to be detected.

The invention discloses a substation inspection robot-based circuit breaker state template-matching identification method, wherein a substation inspection robot is used for automatically acquiring an image of a circuit breaker, and a template matching method is used for identifying the state of the circuit breaker. The method comprises the following specific steps: respectively extracting areas where characters of open and close are positioned as templates by using the template matching method, and carrying out matching on follow-up shot images; carrying out image segmentation on the open template, the close template and matched windows by using maximum between-cluster variance method (OTSU method) to obtain binary images which contain targets and backgrounds; and comprehensively judging the final state of the circuit breaker based on the difference degree between the templates and the targets of the matched windows. Proved by experiments, the method disclosed by the invention has the advantages of fast processing speed, high reliability, no misjudgment on given open-close conclusion and capability of satisfying with the requirements of real time and high reliability for field inspection.

The invention requests to protect a vehicle detection and tracking method based on monocular vision, and belongs to the field of machine vision. The method comprises the steps that in a vehicle detection stage, a rod driving region is determined by combining interest region extracting, self-adaptive Canny edge detection and lane line detection, then a vehicle bottom shadow is obtained by means of local tonal value statistics and a thresholding method in which a maximum between-class variance method is used twice in an integrated manner, furthermore, a supposed vehicle is proposed, and then the supposed vehicle is verified through a texture description co-occurrence matrix method; and in a vehicle tracking stage, an improved algorithm combining Kalman filtering with Cam shift is adopted to carry out multi-target tracking, then new target judging, whether searching is successful, and whether a vehicle is out of an edge are used as three standards, and if a new target is detected twice, the new target is processed as a new tracking target, and the tracking target is updated continuously. By adopting the method, vehicle detection and multi-target tracking under a dynamic background are realized, and the instantaneity, the accuracy and the reliability are relatively high.

The invention discloses a marine oil spill information identification and extraction method based on an SAR image. Noise and distortion of a spaceborne SAR image are eliminated through preprocessing. The universal processing is performed on the image for covering non-sea-surface areas. Then the image is divided by means of a two-dimensional maximum between-class variance method for obtaining an interested area. Based on the divided image, the mis-divided image is corrected by means of morphological filtering. Furthermore a selected characteristic parameter is extracted from an image dividing area. Finally according to priori knowledge, an image dividing result is classified by means of a similarly processed sample image set trained BP neural network, thereby differentiating sea surface oil spill and suspected oil spill and accurately extracting marine oil spill information.

The invention discloses an aggregated white blood cell segmentation counting system. The system comprises an image acquisition module for dyeing white blood cells in a blood sample, dissolving red blood cells in the blood sample by using red blood cell lysate and acquiring a white blood cell image, an image preprocessing module used for performing image background removal on the white blood cellimage and obtaining an optimal segmentation threshold by using a maximum inter-class variance method and roughly segmenting a white blood cell region, an aggregated cell determination module used forobtaining a coarse segmentation image according to the rough segmentation of the white blood cell region, setting a discriminant function of a cell area and obtaining a multi-cell aggregation region,and an aggregated cell segmentation counting module used for extracting a cytoskeleton in each aggregation region and a gray curve at the cytoskeleton by using a morphological refinement method. According to the invention, by analyzing the gray scale characteristics of various white blood cell areas under a low power microscope, an adaptive threshold function is constructed, while a white blood cell count is obtained, the number of oxyphil cells is obtained, the cells in the aggregation region are quickly and accurately divided and counted, the method is quick and simple and is easy to implement.

The invention relates to a fast noise-containing image two-dimensional maximum between-class variance threshold value method, which comprises the steps of firstly solving a gray average value and a gray standard deviation of a noise image; smoothing each pixel of the image by adopting an average gray value of a 3*3 neighborhood to acquire a smooth image; then calculating the between-class variance of the smooth image by using a maximum between-class variance threshold value method, reducing the search space of a solution of the between-class variance through the gray average value and the standard deviation, traversing the search space, and recording a solution, which enables the between-class variance to be the maximum, to be an optimal one-dimensional threshold value T0; and calculating a trace of a between-class variance dispersion matrix of a target class and a background class by using a two-dimensional maximum between-class variance method, reducing the search space of a solution of the trace through the optimal one-dimensional threshold value T0 and the gray standard deviation of the noise image, traversing the search space of the solution, and recording a gray value binary group, which enables the trace of the dispersion matrix to be the maximum, to be an optimal two-dimensional cutting threshold value. The method provided by the invention can avoid traversal for all gray levels, and also can acquire an accurate solution while greatly reducing the calculation amount.

A printer performs printing after appropriately correcting irregularities, thin-line patchiness, and isolated pixels. A first image-quality corrector unit detects slanting-line irregularities represented by the black and white pixel data generated through binary processing according to a method other than the error-variance method to thereby perform smoothing processing therefor. A second image-quality corrector unit detects defects such as irregularities and patchiness specific to the binary processing according to the error-variance method to thereby perform smoothing processing therefor. A third image-quality corrector unit detects isolated pixels in gray fields that are specific to the binary processing according to the error-variance method to thereby distribute the isolated pixels to peripheral pixels. When an original image is reduced in size according to pixel-removal prior to the binary processing, the pixel-removal is not performed if a gradient variation is relatively great with respect to peripheral pixels.

The invention relates to a KNN (K-Nearest Neighbor) sorting algorithm based method for correcting and segmenting the grayscale nonuniformity of an MR (Magnetic Resonance) image, belonging to the field of image processing. The method comprises the following steps of: firstly constructing a grayscale nonuniform field model by utilizing surface fitting knowledge and using a group of orthonormalization basis functions, and establishing energy functions; and then solving model parameters according to an energy function minimization principle to realize grayscale nonuniformity correction and image segmentation, wherein subordinate functions are solved by adopting an iterative algorithm and the KNN algorithm in the model parameter solving process, therefore a partial volume effect is greatly reduced while a grayscale nonuniform field is eliminated, and the influence of noises on the correction and the segmentation of the grayscale nonuniformity of the MR image is reduced. The subordinate functions are solved with KNN through the following steps of: firstly acquiring an accurate smooth normalization histogram by using a kernel estimation algorithm; then respectively solving a threshold value TCG between cerebrospinal fluids and gray matters and a threshold value TGW between the gray matters and white matters by using a maximum between-cluster variance method; carrying out rough sorting on the KNN sorting algorithm by utilizing the two threshold values; and finally accurately sorting points to be fixed by adopting the traditional KNN sorting algorithm.

The invention discloses an all-angle and rapid fingerprint recognition method that makes recognition according to fingerprint details: a. image preprocessing: using local gray level variance method to segment out the area of the fingerprint image; using a directional filtering method to realize binaryzation of fingers in the area of the fingerprint image; using an improved blend OPTA algorithm to make refining treatment to the binary image; b. fingerprint feature extraction: using Poincare method to test and make statistic of the number of core points of fingerprint to make sure of the reference point of fingerprint and the reference direction of fingerprint, and to extract the information of fingerprint feature point in the thinned fingerprint image; c. feature matching: taking the reference point as the point of origin, and taking the reference direction as the principal axis; using a polar coordinate to signify the fingerprint feature point axis, which is taken as characteristic data of template fingerprints; matching the data with the characteristic data of master plate fingerprints to judge whether to anastomose. The novel method can effectively restrain the influence brought from the parallel movement and rotation of the fingerprint image, and can rapidly make out judgement, thereby being suitable for a plurality of running platforms such as large-scale computers, DPS, embedded systems, satisfying the requirements on various personal identification occasions, and having broad market prospect and extremely high economic value.

The invention discloses an improved Canny image edge detection method with noise immunity. On the basis of a Canny algorithm, the improved algorithm combines adaptive median filter to replace Gaussianfilter to denoise the image, so as to filter salt and pepper noise interference better. Then combined with the maximum inter-class variance method and the maximum entropy method, the double thresholdselection method is improved to get the high and low thresholds, so as to achieve the edge detection of the image, so that the target image retains the edge information as much as possible, while filtering out the unnecessary interference edges. The invention realizes the edge detection of an image under the condition of salt and pepper noise pollution and a double threshold selection method withstrong adaptability, and has high application reference value.

The invention discloses a fabric fuzzing and pilling image segmentation method based on wavelet transformation and the morphological algorithm. The method comprises the steps that firstly, histogram equalization is carried out on an acquired fabric fuzzing and pilling image, and eight-stage multi-resolution wavelet decomposition is carried out; secondly, after the coefficient of a first-stage wavelet decomposition layer, the coefficient of a second-stage wavelet decomposition layer and the coefficient of an approximate wavelet decomposition layer are set to be zero, the image is reconstructed, wherein the first-stage wavelet decomposition layer and the second-stage wavelet decomposition layer express the highest-frequency information, and the approximate wavelet decomposition layer expresses low-frequency information; thirdly, binaryzation is carried out on the constructed image according to the maximum between-cluster variance method, and erosion is carried out on the image through a horizontal linear structure element, a 45-degree linear structure element, a vertical linear structure element and a 135-degree linear structure element in sequence; fourthly, dilation is carried out on the eroded image through structure elements with the sequence opposite to that of the structure elements applied to erosion, namely, the 135-degree linear structure element, the vertical linear structure element, the 45-degree linear structure element and the horizontal linear structure element in sequence. By means of the method, the influence of the texture of fabric, illumination unevenness, the surface roughness of the fabric and the size difference and shape difference of fuzzy balls on fuzzing and pilling grade assessment on the fabric can be effectively avoided.

The invention discloses a small cell network edge partial caching method taking user time delay into consideration and belongs to the technical field of wireless communication. The method comprises the steps that a caching video file alternative set is rearranged according to popularity and corresponding popularity values and corresponding video file sizes are acquired; the maximum number K of thevideo files capable of being completely cached is calculated; iteration is carried out through utilization of a maximum between-class variance method according to the maximum number K, complete caching is carried out on the video files of which total size is smaller than or equal to a caching threshold Sc, and partial caching is carried out on the video files of which total size is greater than the caching threshold Sc; when a user initiates a content request to a small cell base station, if the video files are cached completely, the user acquires complete content from the small cell base station and finishes the process; and if the video files are partially cached or are not cached, content is requested from a content server of a core network through a backhaul link. According to the method, the access time delay of the user can be ensured, a cache hit rate can be improved, a utilization rate of a caching space of a small base station can be improved, and the bandwidth pressure of the backhaul link can be effectively mitigated.

The invention provides a slant vehicle detection and tracking system and method based on machine vision. The system comprises a CCD camera, a USB data transmission port and a computer terminal. The method comprises the following steps of: in vehicle detection phase, firstly, image preprocessing is performed, region-of-interest extraction and improved lane line detection are combined, an inclined vehicle detection area is divided; an adaptive threshold value and a maximum between-class variance method are adopted; a shadow area is extracted, further a shadow line is extracted at the bottom of the vehicle, left and right boundaries of the vehicle are determined in combination with Sobel vertical edge extraction to obtain a suspected rectangular frame of a vehicle target, then features in the rectangular frame are extracted, dimensionality reduction is performed on the features by adopting kernel principal component analysis, and detection confirmation is performed by utilizing an Adaboost cascade classifier; in the vehicle tracking stage, mean shift and Kalman filtering are combined, a vehicle detection result is used as an initial tracking target, and the tracking target is screened by using rectangular frame coincidence. The system and the method can realize real-time vehicle detection and tracking, and have high accuracy.

The invention discloses a binary image fusion method for a train license plate. The method comprises the following steps of obtaining channel data, wherein the channel data are composed of the R, G, B three-channel map of the color image of a train license plate, and the grayscale map of the color image of the train license plate; acquiring a plurality of binary images of the channel data, subjecting a binary image obtained based on the maximum interclass variance method, a binary image obtained based on edges, and a binary image obtained based on the maximum interclass variance of adaptive blocks to the logic 'or' operation; subjecting the operation result of the 'or' operation and a binary image obtained based on adaptive weighted Gaussian threshold method to the logic 'and' operation; and adopting a final result to be an ultimate ideal binary image. According to the technical scheme of the invention, diversified train license plates are binarized effectively, and the binarized train license plates are subjected to the logic operation. Therefore, the high accuracy of a train license plate recognition system is ensured. Meanwhile, the more accurate and real-time traffic information is provided for the railway traffic.

The invention relates to a river section degradation evaluation method, which effectively solves the problems that the evaluation excessively depends on data, the compatibility is complicated and the accuracy and efficiency of the evaluation are affected. The method comprises the following steps: establishing river section degradation evaluation elements and indexes, performing in-situ monitoring, carrying out water quality chemical index analysis and biological authentication for a collected sample, determining a weight of criterion layer evaluation elements for the river degradation by utilizing the level analysis characteristic value method, determining a sub-weight for each representative evaluation index of the criterion layer evaluation elements by utilizing a variance method, and obtaining an absolute weight of each index by considering the weight of the criterion layer evaluation elements; evaluating the river section degradation situation according to each evaluation index absolute weight and a matter-element model, calculating a correlation function and correlation degree, and obtaining the degradation degree of the river section according to the evaluation standard. The method is novel and unique, easy to operate and use, good in effect and capable of accurately evaluating the river section degradation and instructing the river governing and ecological restoration.

The invention discloses a sea antenna detection method based on image division and texture features, comprising the following steps: 1) obtaining a color sea surface image; 2) converting the originalcolor image into a gray-scale image; 3) carrying out Gaussian filtering processing on the gray-scale image by using a Gaussian filtering frame; 4) uniformly dividing the filtered image into a plurality of regions along the vertical direction, respectively calculating the gray level co-occurrence matrix and the contrast parameters of the gray level co-occurrence matrix with the distance between thepoint pairs being 1 vertically downward in each region direction; 5) calculating the ratio of the contrast of the co-occurrence matrix of each region and the previous region, wherein the region withthe lar ratio is the region where the sea antenna is located; 6) calculating the gradient in the vertical direction of the selected region, obtaining a threshold value by using a maximum inter-class variance method, and taking pixel points whose gradient value is great than the threshold value as candidate points for straight line detection; 7) hough transform straight line detection is carried out, and that best straight line detect is sea antenna. The method of the invention can effectively detect the position of sea antenna in sea surface image.