57 results about "Gabor filter bank" patented technology

The invention discloses a method for detecting and classifying fabric defects and mainly aims to solve the problem of automatic detection and classification of fabric defects. The method comprises the following steps of: firstly, detecting a picture of the fabric defects, filtering the picture by using a Gabor filter group, selecting an optimal filtering result and performing binaryzation on the optimal filtering result by using a reference picture so as to position the positions of the defects in the picture; secondly, extracting a compound characteristic consisting of a Gabor characteristic and a partial binary model characteristic according to the positions of the defects; thirdly, performing pre-treatment on the compound characteristic by main constituent analysis and generalized discriminant analysis algorithm; fourthly, training a neural network classifier by using a pre-treated defect characteristic; and lastly, realizing accurate classification of a fabric defect characteristic by using a trained classifier. The method has the advantages of accurate defect positioning and high classification accuracy and can be used for detecting and classifying the fabric defects in a textile mill.

The invention provides a magnetic tile surface defect feature extraction and defect classification method based on machine vision. A concrete algorithm comprises a first step of building a 5-scale and 8-direction Gabor filter bank suitable for magnetic tile surface defect feature extraction, conducting filtering to an original image and obtaining a 40-width component plot, a second step of respectively extracting a gray average and a variance feature of the component plot and forming a 80-dimension feature vector, a third step of conducting dimensionality reduction to the original 80-dimension feature vector through a principal component analysis (PCA) method and an independent component analysis (ICA) method, removing relevance and redundancy and obtaining a 20-dimension feature vector, a fourth step of conducting normalization pretreatment to feature vector data, wherein the original data are normalized between zero and one, and a fifth step of adopting a grid method and a K-CV method to achieve SVM parameter optimization at first and training an SVM model using training sample data offline, wherein pretreated testing sample data are input into a support vector machine during online testing, and automatic classification and identification of defects can be achieved. The feature extraction method can effectively filter interference and prominent defects of magnetic tile surface texture, extracted features can reflect defect information accurately, data values are small, and a classifier used for classifying the defects can achieve defect identification fast and accurately online.

The invention belongs to the technical field of image processing and pattern recognition, and relates to a cloth defect detecting method based on machine vision. The cloth defect detecting method includes performing power spectral density analysis for an image of a normal cloth texture and acquiring central frequency F and an azimuthal angle theta of the texture; constructing an SXL adaptive Gabor filter bank; filtering the image of the normal cloth texture to obtain a feature image group, and computing the mean value and the variance of each image of the feature image group; acquiring an image of to-be-detected cloth; filtering the image of the to-be-detected cloth to obtain a feature image group; performing threshold post-processing for the feature image group of the image of the to-be-detected cloth to obtain an absolute feature image group; carrying out normalization processing; fusing images and performing binarization processing for the images to obtain a detected binary image; and removing noise interference to obtain a final detection result. The S represents the number of the selected central frequency, and the L represents the number of the selected azimuthal angle. The cloth defect detecting method has the advantages of high universality and efficiency.

The invention belongs to the technology for detecting a contour of an image target object by adopting a simulated vision mechanism in the bioinformatics technology, which comprises the following steps: determining an azimuth of a filter corresponding to a nonclassical receptive field and restraint quantities of a lateral area and an end area thereof to a central pixel by adopting large and small two scale parameters and performing Gabor filtering in multiple directions, and preparing a restrained image; and performing conventional binarization processing on the restrained image to obtain a target contour plot. In the technology, a Gabor filter bank respectively filters the image in different azimuths under two different scale parameters so as to obtain a high frequency information distributing map and a low frequency information distributing map, a filter of the nonclassical receptive field is utilized to perform restraint processing on non-contour information such as textures, and the like. Therefore, the technology has the characteristics of strong adaptability along with the change of outside input information, capacity of effectively improving the capability of a contour detection system of quickly and accurately extracting the target contour from a complex scene, effect and contour definition, and the like.

The invention provides a voice identifying method based on deep neural network characteristic training. The method involves the realization and identification of Gabor filter bank characteristics, Gabor filter sub-banks, and a deep neural network (DNN), which is achieved through the following steps: a Gabor filter extracting automatic voice identifying characteristics from a voice signal, firstly on the basis of a distributive voice identification standard, extracting a logarithm Me 1 spectrogram from the voice signal, then conducting convolution on the spectrogram and each 2D filter from the Gabor filter bank; selecting a specific modulation frequency, such that a transfer function of the filter exhibits constant overlapping in a modulated frequency field; an automatic voice identifying system, on the basis of character error rate in a test set, carrying on evaluation, and finally acquiring an identification result. According to the invention, the Gabor filter sub-bank can reduce character and word identification errors, and exhibit the channel distortion resistance and low signal-to-noise ratio. The method uses a voice identifier having a high time modulation filter, has low error rate, and increases the distinctiveness among object types.

The present invention relates to a network embedded multi-channel license plate detection system and a license plate detecting method, which is a system and method that can flexibly configure various detection parameters through networks to realize multi-channel and high-definition license plate detection. The system comprises a central processor module, a memory module and an input-output module which are connected with each other through a bus. The method is based on multi-channel image shooting, which has the advantages of more comprehensive image shooting and more information collecting for more useful data so as to contribute to significant improvement in the analysis of image. As the present invention uses the character identification technology based on Gabor filter groups to detect license plates, and as the method has the advantages of strong anti-interference performance, accurate identification, etc., the influence of the background on license plate images is eliminated. Therefore, even under less favorable conditions, the license plate images can be clearly analyzed, which improves the recognition rate of license plates.

The invention relates to a Gabor human face recognizing method based on a simplified intelligent single-particle optimizing algorithm, comprising a simplified intelligent single-particle optimizing algorithm and a link of selecting a Gabor filter group by the simplified intelligent single-particle optimizing algorithm, wherein the simplified intelligent single-particle optimizing algorithm comprises the following steps of: carrying out optimizing search on a solution space of a problem function by a particle and respectively carrying out intelligent updating on all dimensionality components of the single particle during iteration; and the link of selecting the Gabor filter group by the simplified intelligent single-particle optimizing algorithm comprises the following steps of: constructing a particle structure according to the quantity of Gabor filters, determining a particle search range and using a Fisher criterion as a fitness function; during the human face recognition, carrying out characteristic extraction on a human face image conforming to the authentication to form a characteristic database; after a new human face image is input, carrying out the characteristic extraction on the human face image by the selected Gabor filter; and comparing the obtained characteristic vector with the characteristics in the database by a minimal adjacent classifier to judge whether the input image is contained in the database or not and a corresponding identity. The invention improves the recognition precision and the characteristic extracting pertinence.

The invention provides a textile defect detection method based on hierarchical clustering and Gabor filtering. A hierarchical clustering algorithm is combined with a Gabor filter set; digital image pixel gray-level information of a flat textile surface based on an illumination light source is analyzed to automatically position surface defects of a textile; the textile defect detection method comprises the main three steps of dividing a grid pattern, extracting characteristics and comparing the characteristics. The textile defect detection method provided by the invention is especially suitable for automatic recognition of textile surface defects in digital images, acquired under the stable illumination light source, on the flat surface of the textile; the textile defect detection method is a method for automatically dividing the grid pattern from the textile image based on the hierarchical clustering algorithm and carrying out characteristic extraction and defect recognition on the grid pattern based on the Gabor filter set.

The invention discloses a method for extracting a lung lobe contour from a DR image. The method comprises the following steps: a representative template of a lung lobe contour is obtained through offline training; a chest DR image lung lobe area extraction system is initialized; according to the size of a DICOM image, the image is subjected to three-layer pyramid decomposition; a Gabor filter set is used to reconstruct the to-be-processed image, and the residual error of the reconstructed image after Gabor filter is converted into a black and white image; the black and white image is refined with a Zhan-Suen refinement algorithm; with each offline training template called as a convolution kernel operator, the contour image is subjected to convolution; a local optimal convolution value of the optimal possibility is filtered out of the convolution results and subjected to combined evaluation; and a lung lobe contour shape is generated by combining the most matching upper and lower templates and the most matching positions. The method improves the work efficiency and inspection precision of lung disease inspection by doctors, supports further deepening the informatization of tuberculosis monitoring, and facilitates popularization of regular resident infectious disease examination screening of tuberculosis.

The invention discloses an aurora image classification method based on biological stimulation characteristics and manifold learning. The method comprises the following steps of: (1) carrying out preprocessing of edge denoising on an input aurora image; (2) carrying out Gabor filtering on the aurora image subjected to preprocessing by using a multi-directional Gabor filter group, so as to obtain C1-layer characteristic graphs, and taking the sum of pixel gray level values of each characteristic graph as a C1 characteristic of the aurora image; (3) extracting a Gist characteristic of the aurora image; (4) fusing the C1 characteristic and the Gist characteristic so as to obtain a BIFs characteristic of the aurora image; (5) carrying out fuzzy C-mean value clustering on the BIFs characteristic, and subsequently carrying out dimensionality reduction by using a manifold learning algorithm so as to obtain the expression of the BIFs in a low-dimension space; and (6) respectively classifying aurora images by using a support vector machine (SVM) and a nearest neighbor (NN) classifier. By utilizing the method, the recognition process of human brain visual cortex can be well simulated, data redundancy is reduced, classification accuracy rate is improved, and therefore the method can be used for scene classification and object recognition.

The invention relates to a non-reference 3D stereoscopic image quality evaluation method based on asymmetric distortion, comprising the following steps: decomposing a stereoscopic image pair by using a Gabor filter group to obtain Gabor pyramids of left view and right view respectively. The fused cyclopean image is obtained by adding pixel-by-pixel points of the left-view and right-view Gabor pyramids. The difference map is generated by subtracting the absolute difference values ​​of the left and right view Gabor pyramids. The generalized Gaussian model GGD is used to fit the cyclopean diagram and the difference diagram, and the statistical characteristic parameters after fitting are extracted, including the mean value, variance and shape parameters of the GGD distribution. Calculate the structural similarity SSIM score between the left view and the right view decomposed by the Gabor filter, as the asymmetric information of the stereo image, combined with the statistical feature parameters extracted in the previous step, use support vector regression SVR to predict the image quality score .

The invention provides a GrabCut texture image segmentation method based on Gabor transform. The method comprises the following steps that: transforming a two-dimensional image into frequency domain space representation through Gabor wavelet transform, calculating the parametric transform of a frequency spectrum space two-dimensional Gabor filter group to obtain multi-scale multi-direction texture features, constructing a multi-scale texture energy generic function, calculating the GMM (Gaussian Mixture Model) probability distribution model of the texture features, and introducing the GMM of the texture features into GrabCut for image segmentation. The GrabCut texture image segmentation method based on the Gabor transform has the characteristics of being simple in artificial interaction, high in processing speed, good in image segmentation effect and the like.

The invention discloses a method for detecting and classifying defects of non-woven fabrics, and the problems of the automatic detection and classification of four defects including holes, oil stains,foreign objects and scratches of the non-woven fabrics are solved. The method comprises a step of detecting a non-woven fabric defect image, filtering the image by an optimized Gabor filter group, fusing a filtering result, binarizing the result by using an adaptive threshold segmentation method, eliminating noise interference by a pseudo-defect culling algorithm, and thus accurately determiningthe positions of the defects in the image, a step of segmenting a region of interest in the image according to the position of the defects, and extracting a composite feature vector formed by a shapefeature, a first-order moment feature and a second-order moment feature based on the region of interest, a step of training an SVM classifier by using a composite feature vector group and a one-to-onedesign strategy, and a step of finally accurately classifying the defect characteristics of the non-woven fabrics by using the trained classifier group. The method has the advantages of the accuratepositioning of the defects and high accuracy of classification and is used for detecting and classifying cloth defects of non-woven fabric manufacturers.

The invention relates to an image characteristic description method based on a Gabor synthetic characteristic, which comprises the following steps: a first step, acquiring and uploading a human face image signal; a second step, performing resolution adjustment and matrix expression of a face image; a third step, extracting an image characteristic; and a fourth step, synchronously outputting a processing result. According to the image characteristic description method, through using the amplitude part and the phase part which are converted by the Gabor wave filter, wherein the phase part comprises direction information in the Gabor filtering result, a certain characteristic discriminating meaning is realized; the filtering result of a Gabor wave filter set is sufficiently utilized; more abundant characteristic information is extracted for facilitating afterward identification; furthermore for aiming at a defect of averaging image blocks, different importance degrees of sub image blocks to the integral image are considered; and furthermore the face characteristic can be better described in combination with texture contribution degrees.

The invention discloses a target detection method based on a sparse representation and visual cortex attention mechanism. The target detection method is realized by the following steps of: (1) utilizing a multi-scale and multi-directional Gabor filter group to carry out Gabor filtering on an input image to be detected and extracting to obtain S1 layer cells of an HMAX model; (2) utilizing independent component analysis (ICA) and utilizing C1 layer cells of the HMAX model to construct a sparse group; and (3) utilizing C2 layer cells which have invariance on scales and translation, and a Shannon entropy theory to obtain information entropy of the C2 layer cells, and furthermore, constructing a measure for detecting a visual significance region to obtain an obvious image of a previous image. According to the target detection method disclosed by the invention, the perception property of human eyes in finding the significance of a natural scene image can be better satisfied, so that the accurate rate is higher and the false detection rate is lower; and compared with the traditional image detection method, the target detection method has the advantages that the effectiveness and the accuracy in a psychological stimulation mode image are better.

Described is a novel method for feature extraction for automatic gait recognition. This method uses Multi-kernel Fuzzy-based Local Gabor Binary Pattern. From a captured gait video sequence, the gait period is determined then a gait energy image is constructed to represent the spatial-temporal variations during one motion cycle of the gait sequence. Then, each gait sequence is represented with a feature vector. The computation of this vector is conducted by first applying the 2D Gabor filter bank then encoding the variations in the Gabor magnitude using a multi-kernel fuzzy local binary pattern operator. Finally, gait classification is performed using a support vector machine.

The invention discloses a visual perception enlightening high-resolution remote-sensing image segmentation method, and the method enables an object boundary in an image to be divided into an intensity boundary corresponding to a spectrum homogeneous area, and a texture boundary corresponding to a spectrum change area, and respectively extracts two types of visual information which serve as the main basis of segmentation. The method comprises the steps: filtering noise and texture information in the image through a nonlinear filtering method, and obtaining the intensity gradient through a gradient operator; analyzing the texture features of the image through employing a Gabor filter, enabling the filtering output of a plurality of channels to be merged and processed through the gradient operator, and obtaining the texture gradient of the image; carrying out the fusion of the intensity gradient and the texture gradient, carrying out the conversion of the gradient image after fusion through watershed conversion, and achieving the segmentation of the image. Compared with the prior art, the method improves the boundary accuracy of image segmentation, reduces the over-segmentation and under-segmentation phenomena, and can be effectively used in the field of high-resolution remote-sensing image information processing.

The invention is applicable to the technical field of image processing, and provides a skin texture detecting method and apparatus based on Gabor features. The method comprises the steps of dividing an image to be detected into a plurality of image blocks; performing convolution in the space for color channels of the image blocks and pre-established Gabor filter banks sequentially, to obtain k*s*r arrays having the same size as that of the image blocks; calculating mean values and variances of the arrays, and taking the mean values and variances as feature values to obtain k*s*r*2 feature values; and taking the k*s*r*2 feature values as feature vectors of the image blocks, and detecting the feature vectors through a pre-established classifier to determine whether the image blocks have skin texture features. The invention can be used for quickly and accurately detecting images having skin texture features.

The invention provides an improved iris recognition method and aims to improve the accuracy of iris recognition and shorten the time of iris recognition. The method includes decomposing an iris image through wavelet transform to obtain low frequency sub-band information of the iris image, extracting texture features of the iris by utilizing two-dimensional log-Gabor filter banks in different directions and dimensions, and matching generated iris feature codes by utilizing a support vector machine. The iris recognition method is more suitable for real-time iris recognition.

The invention relates to a method and a system for achieving an intelligent bus stop board. The method comprises the steps of locating images obtained by a camera and obtaining a locating picture of a bus license board, carrying bus license board cutting to the locating picture to obtain a character binary picture, obtaining a character grey-scale picture from the locating picture of the bus license board according to the boundary location of the character binary picture, filtering the character grey-scale picture through a Gabor filter set to obtain Gabor characteristics, obtaining a digital group by the Gabor characteristics through a corresponding batch processing (BP) neural network, when the largest numerical value of the digit group is larger than a threshold value, obtaining a character, corresponding to the largest numerical value, in a recognition network as a recognition result, when the largest numerical value of the digit group is smaller than or equal to the threshold value, recognizing and processing the character binary picture corresponding to the Gabor characteristics, obtaining average moving speed of a corresponding bus according to the character recognition result of the bus license board and multiple frames of images obtained by the camera, and obtaining the shortest waiting time of all bus stations according to the average moving speed and displaying the shortest waiting time to passengers waiting for the bus through the bus stop board.

The invention relates to a method for identifying a flaw of a textile based on a Gabor filter and an RBF support vector machine. The method comprises: collecting a textile image and carrying out pretreatment; generating a Gabor filter group and filtering the textile image; carrying out preferential fusion on a filter image group; carrying out binaryzation processing on the fused image to display aflaw region; generating a feature vector for the flaw region; and classifying the feature vector by using a pre-trained RBF kernel function support vector machine classifier. According to the invention, the image is analyzed by using the Gabor filter and texture information of each scale and each angle is covered; and the flaw type is analyzed by using the RB support vector machine. The method has the positive significance in improving the production quality of the textile.

The invention discloses a local edge feature integration-based target object contour extraction method. The method comprises the steps of firstly performing filtration processing on an input original image by adopting a group of Gabor filter banks different in direction to obtain a directional information graph and a corresponding energy distribution image; and secondly dividing an image region into a central region and a peripheral region according to processed pixel point coordinates, and designing a peripheral center action mechanism based on a directional distribution difference, so that the robust contour extraction method is realized through flexible local edge feature integration, wherein the peripheral region consists of a plurality of sub-regions capable of independently perceiving local features and performing nonlinear modulation in response to the central region. According to the contour extraction method, feature integration parameters can be adaptively calculated according to a context relationship in a large-range region where pixels are located, and the robustness and validity of contour extraction of target objects in complex scenes can be effectively improved.

The invention discloses an image processing method for chest X-ray DR (digital radiography) image rib inhibition. The method comprises the following steps: acquiring a chest X-ray DR image; performing pyramid decomposition on the DR image, performing a down sampling process to obtain a Gaussian image pyramid S, and performing an up sampling process to obtain a Laplacian image pyramid difference chart D(S); taking the minimum S as a current to-be-processed image I; performing filtering processing on the image I by using an adjustable Gabor filter bank so as to obtain a reconstructed image R; differencing the to-be-processed image I and the reconstructed image R to obtain a processing result image E with weakened segment-shaped textures under the scale; and doubling the processing result image E, adding the processing result image E and a corresponding Laplacian image pyramid difference chart D(S) under the size together, and repeating the processing procedure until the size is the same as that of an original DR image, thereby obtaining an image after rib inhibition. According to the method disclosed by the invention, the visual saliency of pulmonary shadows is improved, the workload of doctors is reduced, automatic processing can be realized, and an analysis conclusion is more objective and stable.

The invention belongs to the technology for detecting a contour of an image target object by adopting a simulated vision mechanism in the bioinformatics technology, which comprises the following steps: determining an azimuth of a filter corresponding to a nonclassical receptive field and restraint quantities of a lateral area and an end area thereof to a central pixel by adopting large and small two scale parameters and performing Gabor filtering in multiple directions, and preparing a restrained image; and performing conventional binarization processing on the restrained image to obtain a target contour plot. In the technology, a Gabor filter bank respectively filters the image in different azimuths under two different scale parameters so as to obtain a high frequency information distributing map and a low frequency information distributing map, a filter of the nonclassical receptive field is utilized to perform restraint processing on non-contour information such as textures, and the like. Therefore, the technology has the characteristics of strong adaptability along with the change of outside input information, capacity of effectively improving the capability of a contour detection system of quickly and accurately extracting the target contour from a complex scene, effect and contour definition, and the like.

The invention provides a different-source image matching method based on Gabor coding, which comprises the steps of carrying out filtering on a reference image and a real-time image by using a Gabor filter group, and acquiring a group of reference image filtering images and a group of real-time image filtering images, wherein the reference image and the real-time image are different-source images; carrying out pooling on each reference image filtering image and each real-time image filtering image, and acquiring pooling data of each reference image filtering image and pooling data of each real-time image filtering image; and carrying out binaryzation and binary presentation on the pooling data of each reference image filtering image and the pooling data of each real-time image filtering image, and acquiring a Gabor coding characteristic of the reference image and a Gabor coding characteristic of the real-time image; calculating the similarity between the Gabor coding characteristic of the reference image and the Gabor coding characteristic of the real-time image by using bit manipulation of binary data, and acquiring a matching result of the reference image and the real-time image. The method provided by the invention can realize effective matching for the different-source images, and reduce the amount of calculation in the matching process.

The invention provides a dropper image processing method and a dropper capturing system based on a linear array camera. The method comprises steps: a dropper image is acquired; the dropper image is preprocessed; a GABOR filter bank is used to carry out GABOR transformation on the dropper image; the dropper image is subjected to dynamic binary processing; and the area of a white point in the imageafter binaryzation is calculated, and if the area is larger than a predetermined threshold, the dropper is judged. The system comprises a linear array antenna and a server. A clear image of a tunnel surface can be acquired, the detection efficiency is high, the reliability is good, and manpower can be saved.

The invention discloses a remote-sensing image perceptual hash authentication method based on a Gabor filter bank and DWT converting. Through the feature that a Gabor filter can enhance an image in different directions under different frequencies and the great effect of edge features in remote-sensing image using, a remote-sensing image is subjected to grid dividing first, then based on a self-adaptation strategy, the parameters of the Gabor filter bank are determined in a self-adaptation mode according to the information entropy of a grid unit, then the edge features of the grid unit are extracted and compressed by DWT converting and a PCA method, and finally the perceptual hash of the image is generated; an by comparing the normalization Hamming distance between image perceptual hash sequences, an authentication process of the remote-sensing image is achieved. Authentication based on content of the remote-sensing image is achieved, remote-sensing image local tampering can be effectively detected, robustness is achieved on image operation with content unchanged is achieved, and powerful safety guarantee is provided for effective using of the remote-sensing image.

The invention discloses an ELM-based multi-granularity iris recognition method and solves the problems of incomplete extracted features and low recognition speed in an existing iris recognition method. The method comprises the steps of performing image acquisition and marking; performing image preprocessing; performing a gray-level co-concurrence matrix feature extraction process; performing a 2D-Gabor filter group feature extraction process; constructing a multi-granularity eigenvector; obtaining an iris recognition model; performing iris category testing; and calculating recognition precision. According to the method, a gray-level co-concurrence matrix is combined with a 2D-Gabor filter group to generate the multi-granularity eigenvector, and the multi-granularity eigenvector contains not only high-frequency texture information but also low-medium-frequency texture information, so that the multi-granularity eigenvector contains relatively comprehensive iris features, the iris recognition characteristics are enhanced, and the iris recognition precision is improved; and an ELM (Extreme Learning Machine) is applied to the iris recognition process, so that the iris recognition speed is increased. The method is suitable for the security information field with relatively high requirements on the recognition precision and the real-time property.

The invention requests to protect an image enhancement method based on improved non-subsampled Contourlet transform (NSCT). The method comprises the following steps: firstly, carrying out NSCT on an image to obtain low-pass sub-bands and band-pass sub-bands in various scales and directions; carrying out linear enhancement processing on the low-pass sub-band, so that the overall contrast of the image is improved; for a band-pass sub-band, adaptively determining a denoising threshold value of each sub-band according to energy distribution, and a weak edge enhancement algorithm is provided, so that the effects of enhancing detail texture and suppressing noise are achieved. and acquiring a clear palm vein image through Contourlet inverse transformation. And finally, the fractured part of the palm vein is bridged through a Gabor filter bank. Experimental results show that the palm vein detail texture information is effectively enhanced through the algorithm, and the contrast ratio, the information entropy, the average gradient and the variance are improved to 47.9, 9.1, 5.1 and 2594 respectively.

The invention discloses a composite sleeper pore defect identification method based on Gabor multi-feature extraction and optimization. The method comprises: firstly, carrying out X-ray transillumination detection on the composite sleeper through optimized detection system parameters on the basis of the attenuation characteristic of the composite sleeper to X rays, and obtaining an original projection image; secondly, preprocessing the original ray image, including line minimum noise reduction and fuzzy enhancement; secondly, performing multi-scale and multi-direction feature extraction on theray image by utilizing a designed Gabor filter bank, and performing screening according to an energy criterion to obtain an image with prominent pore defect features; vectorizing the feature image, performing mean value removal and normalization, and completing clustering and segmentation of the pore defects by using a fuzzy C-means method; and finally, removing miscellaneous points and pseudo defects by using morphological opening and closing operation, and calculating defect geometric characteristic parameters to realize quantitative identification of the pore defects.