CN109359659A - A kind of car insurance piece classification method based on color characteristic - Google Patents

A kind of car insurance piece classification method based on color characteristic Download PDF

Info

Publication number
CN109359659A
CN109359659A CN201811597480.9A CN201811597480A CN109359659A CN 109359659 A CN109359659 A CN 109359659A CN 201811597480 A CN201811597480 A CN 201811597480A CN 109359659 A CN109359659 A CN 109359659A
Authority
CN
China
Prior art keywords
image
vector
color
classification
pixel
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201811597480.9A
Other languages
Chinese (zh)
Inventor
尤波
陈国杰
梁强
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Harbin University of Science and Technology
Original Assignee
Harbin University of Science and Technology
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Harbin University of Science and Technology filed Critical Harbin University of Science and Technology
Priority to CN201811597480.9A priority Critical patent/CN109359659A/en
Publication of CN109359659A publication Critical patent/CN109359659A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/56—Extraction of image or video features relating to colour
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Multimedia (AREA)
  • Image Analysis (AREA)

Abstract

A kind of car insurance piece classification method based on color characteristic, belongs to image procossing and area of pattern recognition.The present invention solves the classification problem of efficiently and accurately of the car insurance piece under fixed complex environment in existing automatic automobile safety box assembling link using the color characteristic of car insurance piece as classification foundation.It is the following steps are included: Step 1: eliminate the background in image by average background method using CCD industrial camera acquisition car insurance picture information；Step 2: carrying out image enhancement using stain recovery technique；Step 3: the extraction foundation characterized by the coloration center vector of different classes of car insurance piece, completes the feature extraction clustered to colouring information based on chroma vector；Step 4: inputting support vector machines by the characteristic vector that will be extracted, reach the classification of the identification to different type safety plate.The present invention is suitable for the classification problem between different type car insurance piece.

Description

automobile safety disc classification method based on color characteristics
Technical Field
The invention belongs to the field of image processing and pattern recognition, and particularly relates to a method for classifying automobile safety discs based on color characteristics.
Background
In the production of most of current automobile fuse boxes, the automobile fuse pieces are inserted manually, the automation degree in the production process is very low, the problems of low insertion efficiency and easy occurrence of insertion errors exist, and the efficient and accurate classification of the automobile fuse pieces in a fixed complex environment is one of the main problems of limiting the application of automatic production equipment in the production process of the automobile fuse boxes. The color difference among automobile safety plates of different models is an important basis for distinguishing various automobile safety plates, the classification of the automobile safety plates by distinguishing the colors of the automobile safety plates of different models is one of the main feasible methods at present, and the main methods at present comprise methods such as a color space threshold, a color histogram, a color distance and the like aiming at the extraction and characterization of color features. The method for characterizing the color threshold has a simple structure and is easy to implement, but the method has the following defects: firstly, the method performs sampling statistics on different colors, determines the value range of each component of each color in a color space according to the statistical result, and then sets a color threshold value to distinguish different colors, so that the sampling statistical process is complicated, and when the number of statistical samples is insufficient, the classification effect is greatly influenced. Secondly, the method has poor classification effect when distinguishing similar colors or more classifications. The method for representing colors by using the histogram is simple in calculation, has scale, translation and rotation invariance, and is widely applied to image retrieval, but the identification precision of the method is directly influenced by the quantization degree, if the quantization degree is too large, a large amount of color information is lost, and the problem of inaccurate color classification occurs, and if the quantization degree is too small, the extracted feature vector dimension is too high, and the problem of difficulty in use occurs. The color distance characterization method has been successfully applied to many content-based image retrieval and identification systems, but the method has poor effect on distinguishing similar colors and is generally only used for distinguishing images with larger feature differences. Therefore, the invention provides a method for classifying automobile safety discs based on color characteristics.
Disclosure of Invention
The invention provides a method for classifying automobile safety discs based on color characteristics, which aims to solve the problem that the automobile safety discs in the existing automatic automobile safety box assembly link are difficult to classify efficiently in a fixed complex environment.
A method for classifying automobile safety discs based on color features is characterized in that the method for classifying the automobile safety discs based on the color features comprises the following implementation steps:
step one, acquiring image information of the automobile safety disc by using a CCD industrial camera, and eliminating a background in the image by using an average background method, wherein the method specifically comprises the following steps:
(1) creating a background model based on statistics, acquiring 500 frames of images as statistical samples through an industrial camera, wherein the time interval between the frames is 1s, and then taking a mean image of the samples as the background model, wherein the calculation formula is as follows:
wherein,is the pixel value of the background model at (x, y), is the number of samples, fi(x, y) is a pixel value of the ith frame image at (x, y),
(2) and calculating the threshold value of the pixel point. Taking 3 times of the sample standard deviation sigma as the error range of the background model pixel point, the calculation formula is as follows:
where σ (x, y) is the sample standard deviation at pixel point (x, y),
(3) the background is segmented by using a background model, the pixel value of the image of a new frame acquired by the camera at (x, y) is g (x, y), the pixel value after segmentation by the background model is g' (x, y), and the computing formula of the segmented pixel value is as follows:
step two, performing image enhancement based on stain repair on the image segmented by the background in the step one, setting p as a point to be repaired on a boundary delta omega, setting B (epsilon) as a known pixel area in the field with a p-scale parameter of epsilon, and when epsilon is small enough, determining a new pixel value at the p point by the field B (epsilon), wherein the calculation formula is as follows:
wherein q is the pixel point in B (ε), ▽ I (q) is the gradient at the point q, w (p, q) is the weighting function used to define the contribution size of each pixel in B (ε), and w (p, q) is defined as follows:
w(p,q)＝dir(p,q)·dst(p,q)·lev(p,q) (5)
in the formula:
wherein d is0And T0Respectively, a distance parameter and a level set parameter, where n (p) is a normal direction at p, and T (·) represents a distance from a pixel point to a boundary δ Ω, and is defined as:
and step three, extracting the color characteristics of the automobile safety disc by using a chrominance vector clustering method. In the HSI color space, color information and luminance information of an image are separated, the color information being represented by a hue component H and a saturation component S, and thus colors are characterized by image color information, where a plane formed by the hue component and the saturation component is referred to as a chromaticity plane H-S, and a point on the chromaticity plane is referred to as a chromaticity vector ZhsColor of image under HSI spaceThe color information consists of a color metric. The chrominance vectors of each pixel point in the image are counted, and the chrominance center vector Z of the image is calculated according to the statistical distribution condition of the chrominance vectorscenterDefined as:
in the formula:as a chrominance vector ZhsThe weighted value of (1), the total number of N pixel points, N (h, s) is the statistic value of the chromaticity vector,
considering the color difference existing among the automobile insurance pieces of the same type in reality, defining the chromaticity center vector of the image as the clustering center of the chromaticity center vectors of the m automobile insurance piece samples, and defining as follows:
the K-type automobile safety discs are arranged for classification, and the specific method for extracting the color features through the chroma vector clustering is as follows:
(1) calculating a chromaticity center vector for each colorAnd i represents the type of the automobile safety disc.
(2) The chrominance vectors Z of all pixel points in the image are calculatedhsIs classified intoIn the chroma vector pattern class which is the clustering center, the classification basis is as follows:
in the formulaCalled euclidean distance, defined as:
(3) after the classification of the pixel points in the image is finished, counting the classification result, and classifying the pixel points to a chromaticity centerThe statistical value of the pixel point is used as the ith dimension characteristic value of the color characteristic vector to finally obtain a K dimension color characteristic vector x,
step four, inputting the extracted feature vectors into a support vector machine to achieve the specific identification and classification method of the different types of safety plates as follows:
constructing a training sample set by adopting the feature vectors extracted in the third step for random samples, training an SVM (support vector machine) two-classification classifier by using the training sample set, classifying and identifying a plurality of types of automobile insurance pieces by a plurality of trained SVM two-classification classifiers, setting M types of automobile insurance pieces, taking the training sample set as T, and dividing the training sample set into X types1Andtwo types, denoted as (x)i,yi)，i＝1,2,…,n，yiN is the total number of training sample sets, then the samples and labels satisfy:
selecting a proper Gaussian radial kernel function K (x, z) and a proper parameter C, and constructing and solving an optimization problem:
obtaining an optimal solution from equation (15)Selection α*A positive component ofAnd (3) calculating:
from α*And b*Determining a construction decision function:
the class w can be done by a decision function1Andinto two classes of1And (4) from the training sample set, repeating the process to solve the decision function of the residual categories, and finally, inputting the extracted feature vector into the trained SVM to classify the insurance plates of different types.
Has the advantages that: the inter-class distance of the color feature vectors extracted by utilizing the chroma vector clustering is far greater than the intra-class distance, and the feature vectors extracted by the method have good separability. The automobile insurance pieces can be characterized by characteristic vectors extracted by chromaticity vector clustering, and high-accuracy classification of various automobile insurance pieces can be realized by combining a plurality of SVM classifiers.
Drawings
FIG. 1 shows a conventional class 9 automobile safety guard
FIG. 2 is an average background image of a fixed complex environment
FIG. 3 is an original image of a yellow type automobile safety disc
FIG. 4 is a foreground image of a yellow type automobile safety patch
FIG. 5 is a diagram showing the effect of the yellow automobile safety disc after spot repairing
FIG. 6 is a chromaticity center vector corresponding to 20 samples of a class 9 automotive rupture disc
FIG. 7 is a schematic diagram of a multi-classification process of the automobile rupture disc
Detailed Description
Preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
The specific implementation mode is as follows: referring to fig. 1, the method for classifying the automobile insurance plates based on the color characteristics is realized by the following steps:
step one, acquiring image information of the automobile safety disc by using a CCD industrial camera, and eliminating a background in the image by using an average background method, wherein the method specifically comprises the following steps:
(1) creating a background model based on statistics, acquiring 500 frames of images as statistical samples through an industrial camera, wherein the time interval between the frames is 1s, and then taking a mean image of the samples as the background model, wherein the calculation formula is as follows:
wherein,is the pixel value of the background model at (x, y), is the number of samples, fi(x, y) is a pixel value of the ith frame image at (x, y),
(2) and calculating the threshold value of the pixel point. Taking 3 times of the sample standard deviation sigma as the error range of the background model pixel point, the calculation formula is as follows:
where σ (x, y) is the sample standard deviation at pixel point (x, y),
(3) the background is segmented by using a background model, the pixel value of the image of a new frame acquired by the camera at (x, y) is g (x, y), the pixel value after segmentation by the background model is g' (x, y), and the computing formula of the segmented pixel value is as follows:
step two, performing image enhancement based on stain repair on the image segmented by the background in the step one, setting p as a point to be repaired on a boundary delta omega, setting B (epsilon) as a known pixel area in the field with a p-scale parameter of epsilon, and when epsilon is small enough, determining a new pixel value at the p point by the field B (epsilon), wherein the calculation formula is as follows:
wherein q is the pixel point in B (ε), ▽ I (q) is the gradient at the point q, w (p, q) is the weighting function used to define the contribution size of each pixel in B (ε), and w (p, q) is defined as follows:
w(p,q)＝dir(p,q)·dst(p,q)·lev(p,q) (5)
in the formula:
wherein d is0And T0Respectively, a distance parameter and a level set parameter, where n (p) is a normal direction at p, and T (·) represents a distance from a pixel point to a boundary δ Ω, and is defined as:
and step three, extracting the color characteristics of the automobile safety disc by using a chrominance vector clustering method. In the HSI color space, color information and luminance information of an image are separated, the color information being represented by a hue component H and a saturation component S, and thus colors are characterized by image color information, where a plane formed by the hue component and the saturation component is referred to as a chromaticity plane H-S, and a point on the chromaticity plane is referred to as a chromaticity vector ZhsThe color information of an image under the HSI space consists of the amount of color measurement. The chrominance vectors of each pixel point in the image are counted, and the chrominance center vector Z of the image is calculated according to the statistical distribution condition of the chrominance vectorscenterDefined as:
in the formula:as a chrominance vector ZhsThe weighted value of (1), the total number of N pixel points, N (h, s) is the statistic value of the chromaticity vector,
considering the color difference existing among the automobile insurance pieces of the same type in reality, defining the chromaticity center vector of the image as the clustering center of the chromaticity center vectors of the m automobile insurance piece samples, and defining as follows:
the K-type automobile safety discs are arranged for classification, and the specific method for extracting the color features through the chroma vector clustering is as follows:
(1) calculating a chromaticity center vector for each colorAnd i represents the type of the automobile safety disc.
(2) The chrominance vectors Z of all pixel points in the image are calculatedhsIs classified intoIn the chroma vector pattern class which is the clustering center, the classification basis is as follows:
in the formulaCalled euclidean distance, defined as:
(3) after the classification of the pixel points in the image is finished, counting the classification result, and classifying the pixel points to a chromaticity centerThe statistical value of the pixel point is used as the ith dimension characteristic value of the color characteristic vector to finally obtain a K-dimension color characteristic vector x, and the step four, the extracted characteristic vector is input into a support vector machine to achieve the specific identification and classification method of different types of insurance pieces as follows:
constructing a training sample set by adopting the feature vectors extracted in the third step for random samples, training an SVM (support vector machine) two-classification classifier by using the training sample set, classifying and identifying a plurality of types of automobile insurance pieces by a plurality of trained SVM two-classification classifiers, setting M types of automobile insurance pieces, taking the training sample set as T, and dividing the training sample set into X types1Andtwo types, denoted as (x)i,yi)，i＝1,2,…,n，yiN is the total number of training sample sets, then the samples and labels satisfy:
selecting a proper Gaussian radial kernel function K (x, z) and a proper parameter C, and constructing and solving an optimization problem:
obtaining an optimal solution from equation (15)Selection α*A positive component ofAnd (3) calculating:
from α*And b*Determining a construction decision function:
the class w can be done by a decision function1Andinto two classes of1And (4) from the training sample set, repeating the process to solve the decision function of the residual categories, and finally, inputting the extracted feature vector into the trained SVM to classify the insurance plates of different types.
Example (b):
the following description will be further made with reference to the embodiments, and referring to fig. 1, there are 9 types of commonly used automobile safety discs, and the corresponding specifications are as follows: 2A (grey), 3A (light purple), 5A (brown), 7.5A (dark red), 10A (red), 15A (blue), 20A (yellow), 25A (light white), 30A (green).
And (3) constructing an average background image (see the attached figure 2) of the fixed complex environment according to the step one, and performing background elimination on the original image (see the attached figure 3) of the yellow automobile safety piece by using the constructed average background image to obtain a foreground image (see the attached figure 4) of the yellow automobile safety piece.
And (3) performing image enhancement on the foreground images (see the attached figure 4) of the yellow automobile safety disc according to the two pairs of steps, and obtaining the effect image of the yellow automobile safety disc subjected to stain repair, which is shown in the attached figure 5.
The 9 types of automobile insurance plates are sampled for 20 times according to the formula (10), the chromaticity center corresponding to each sample is shown in the attached figure 6, and the clustering center of the 20 samples according to the formula (11) is shown in the following table:
respectively randomly sampling the various automobile insurance sheets for 100 times, performing feature extraction according to the third step, constructing a training sample set T, finally solving according to the fourth step to obtain 8 SVM two-classification classifiers, using a plurality of classifiers to perform a multi-classification process schematic diagram referring to the attached figure 7, randomly sampling the various automobile insurance sheets for 100 times, performing feature extraction according to the third step, inputting extracted feature vectors into the SVM two-classification classifiers according to the classification process, repeating the steps for three times, wherein the classification accuracy is shown in the following table:

Claims (1)

1. A method for classifying automobile safety discs based on color features is characterized in that the method for classifying the automobile safety discs based on the color features comprises the following implementation steps:
step one, acquiring image information of the automobile safety disc by using a CCD industrial camera, and eliminating a background in the image by using an average background method, wherein the method specifically comprises the following steps:
(1) creating a background model based on statistics, acquiring 500 frames of images as statistical samples through an industrial camera, wherein the time interval between the frames is 1s, and then taking a mean image of the samples as the background model, wherein the calculation formula is as follows:
wherein,is the pixel value of the background model at (x, y), is the number of samples, fi(x, y) is a pixel value of the ith frame image at (x, y),
(2) calculating a threshold value of a pixel point, and taking 3 times of the standard deviation sigma of the sample as an error range of the pixel point of the background model, wherein the calculation formula is as follows:
where σ (x, y) is the sample standard deviation at pixel point (x, y),
(3) the background is segmented by using a background model, the pixel value of the image of a new frame acquired by the camera at (x, y) is g (x, y), the pixel value after segmentation by the background model is g' (x, y), and the computing formula of the segmented pixel value is as follows:
step two, performing image enhancement based on stain repair on the image segmented by the background in the step one, setting p as a point to be repaired on a boundary delta omega, setting B (epsilon) as a known pixel area in the field with a p-scale parameter of epsilon, and when epsilon is small enough, determining a new pixel value at the p point by the field B (epsilon), wherein the calculation formula is as follows:
in the formula: q is a pixel point within B (epsilon),for the gradient at point q, w (p, q) is a weighting function defining the contribution of each pixel within B (epsilon), and is defined as follows:
w(p,q)＝dir(p,q)·dst(p,q)·lev(p,q) (5)
in the formula:
wherein d is0And T0Respectively, a distance parameter and a level set parameter, where n (p) is a normal direction at p, and T (·) represents a distance from a pixel point to a boundary δ Ω, and is defined as:
extracting color characteristics of the automobile safety disc by using a chroma vector clustering method, wherein in an HSI (hue, saturation and intensity) color space, color information and brightness information of an image are separated, the color information is represented by a hue component H and a saturation component S, so that colors are represented by the color information of the image, a plane formed by the hue component and the saturation component is called a chroma plane H-S, and points on the chroma plane are called a chroma vector ZhsThe color information of the image in the HSI space consists of the color measurement, the chrominance vector of each pixel point in the image is counted, and the chrominance center vector Z of the image is calculated according to the statistical distribution condition of the chrominance vectorscenterDefined as:
in the formula:as a chrominance vector ZhsThe weighted value of (1), the total number of N pixel points, N (h, s) is the statistic value of the chromaticity vector,
considering the color difference existing among the automobile insurance pieces of the same type in reality, defining the chromaticity center vector of the image as the clustering center of the chromaticity center vectors of the m automobile insurance piece samples, and defining as follows:
the K-type automobile safety discs are arranged for classification, and the specific method for extracting the color features through the chroma vector clustering is as follows:
(1) calculating a chromaticity center vector for each colori represents the type of the automobile safety disc;
(2) the chrominance vectors Z of all pixel points in the image are calculatedhsIs classified intoIn the chroma vector pattern class which is the clustering center, the classification basis is as follows:
in the formulaCalled euclidean distance, defined as:
(3) after the classification of the pixel points in the image is finished, counting the classification result, and classifying the pixel points to a chromaticity centerThe statistical value of the pixel point is used as the ith dimension characteristic value of the color characteristic vector to finally obtain a K dimension color characteristic vector x,
step four, inputting the extracted feature vectors into a support vector machine to achieve the specific identification and classification method of the different types of safety plates as follows:
constructing a training sample set by adopting the feature vectors extracted in the third step for random samples, training an SVM (support vector machine) two-classification classifier by using the training sample set, classifying and identifying a plurality of types of automobile insurance pieces by a plurality of trained SVM two-classification classifiers, setting M types of automobile insurance pieces, taking the training sample set as T, and dividing the training sample set into X types1Andtwo types, denoted as (x)i,yi)，i＝1,2,…,n，yiN is the total number of training sample sets, then the samples and labels satisfy:
selecting a proper Gaussian radial kernel function K (x, z) and a proper parameter C, and constructing and solving an optimization problem:
obtaining an optimal solution from equation (15)Selection α*A positive component ofAnd (3) calculating:
from α*And b*Determining a construction decision function:
the class w can be done by a decision function1Andinto two classes of1And (4) from the training sample set, repeating the process to solve the decision function of the residual categories, and finally, inputting the extracted feature vector into the trained SVM to classify the insurance plates of different types.
CN201811597480.9A 2018-12-26 2018-12-26 A kind of car insurance piece classification method based on color characteristic Pending CN109359659A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811597480.9A CN109359659A (en) 2018-12-26 2018-12-26 A kind of car insurance piece classification method based on color characteristic

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811597480.9A CN109359659A (en) 2018-12-26 2018-12-26 A kind of car insurance piece classification method based on color characteristic

Publications (1)

Publication Number Publication Date
CN109359659A true CN109359659A (en) 2019-02-19

Family

ID=65329480

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811597480.9A Pending CN109359659A (en) 2018-12-26 2018-12-26 A kind of car insurance piece classification method based on color characteristic

Country Status (1)

Country Link
CN (1) CN109359659A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111639589A (en) * 2020-05-28 2020-09-08 西北工业大学 Video false face detection method based on counterstudy and similar color space

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050126505A1 (en) * 2003-12-10 2005-06-16 Gallager Scott M. Optical method and system for rapid identification of multiple refractive index materials using multiscale texture and color invariants
CN102663357A (en) * 2012-03-28 2012-09-12 北京工业大学 Color characteristic-based detection algorithm for stall at parking lot
CN108304878A (en) * 2018-02-05 2018-07-20 河北工业大学 A kind of photovoltaic cell color classification algorithm based on aberration histogram and DAG-SVMs

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20050126505A1 (en) * 2003-12-10 2005-06-16 Gallager Scott M. Optical method and system for rapid identification of multiple refractive index materials using multiscale texture and color invariants
CN102663357A (en) * 2012-03-28 2012-09-12 北京工业大学 Color characteristic-based detection algorithm for stall at parking lot
CN108304878A (en) * 2018-02-05 2018-07-20 河北工业大学 A kind of photovoltaic cell color classification algorithm based on aberration histogram and DAG-SVMs

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
尤波 等: "基于色度向量聚类的保险片识别及插接方法" *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111639589A (en) * 2020-05-28 2020-09-08 西北工业大学 Video false face detection method based on counterstudy and similar color space
CN111639589B (en) * 2020-05-28 2022-04-19 西北工业大学 Video false face detection method based on counterstudy and similar color space

Similar Documents

Publication Publication Date Title
CN111179251B (en) Defect detection system and method based on twin neural network and by utilizing template comparison
CN107545239B (en) Fake plate detection method based on license plate recognition and vehicle characteristic matching
Jagadev et al. Detection of leukemia and its types using image processing and machine learning
CN108108761B (en) Rapid traffic signal lamp detection method based on deep feature learning
CN106846316B (en) Automatic identification method for typical defect images in GIS
CN103824091B (en) A kind of licence plate recognition method for intelligent transportation system
CN107610114A (en) Optical satellite remote sensing image cloud snow mist detection method based on SVMs
CN106610969A (en) Multimodal information-based video content auditing system and method
Niu et al. Image segmentation algorithm for disease detection of wheat leaves
CN109948625A (en) Definition of text images appraisal procedure and system, computer readable storage medium
DK2447884T3 (en) A method for the detection and recognition of an object in an image and an apparatus and a computer program therefor
CN107590500A (en) A kind of color recognizing for vehicle id method and device based on color projection classification
CN102184404B (en) Method and device for acquiring palm region in palm image
CN109509188A (en) A kind of transmission line of electricity typical defect recognition methods based on HOG feature
Yingthawornsuk et al. Automatic Thai Coin Calculation System by Using SIFT
CN116416206A (en) Defect detection method and device based on training of small amount of normal sample data set
CN109359659A (en) A kind of car insurance piece classification method based on color characteristic
Fajas et al. Automatic Number Plate Recognition for indian standard number plates
CN108615039A (en) Cartridge case defect automatic testing method based on computer vision
CN111738310B (en) Material classification method, device, electronic equipment and storage medium
Pant et al. Automatic Nepali number plate recognition with support vector machines
CN108765426A (en) automatic image segmentation method and device
CN112164057A (en) Qualified label detection method, storage medium and electronic equipment
CN110929740A (en) LGBM model-based tongue quality and tongue coating separation method
CN117036314A (en) Method for detecting oxidation area of high-density flexible IC substrate

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20190219

WD01 Invention patent application deemed withdrawn after publication