CN115880297B - Quilt cover dyeing quality assessment method based on machine vision - Google Patents
Quilt cover dyeing quality assessment method based on machine vision Download PDFInfo
- Publication number
- CN115880297B CN115880297B CN202310180939.XA CN202310180939A CN115880297B CN 115880297 B CN115880297 B CN 115880297B CN 202310180939 A CN202310180939 A CN 202310180939A CN 115880297 B CN115880297 B CN 115880297B
- Authority
- CN
- China
- Prior art keywords
- super
- quilt cover
- pixel block
- super pixel
- detected
- 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.)
- Active
Links
Images
Classifications
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Abstract
The invention relates to the technical field of image processing, and provides a quilt cover dyeing quality assessment method based on machine vision, which comprises the following steps: acquiring a surface image of a quilt cover to be detected and a surface image of a standard quilt cover, and acquiring a gray image of the surface of the quilt cover to be detected and a gray image of the standard quilt cover; obtaining first color difference similarity, normalizing the sum of the first color difference similarity of all pixel points to obtain the color difference of the quilt cover to be detected and the standard quilt cover; obtaining the duty ratio of each chrominance value of each super pixel block; obtaining the quilt cover color distance of the first super pixel block and the second super pixel block; obtaining a surface dyeing significance value of each super pixel block; and obtaining a dyeing quality evaluation value according to the surface dyeing remarkable values of all the super pixel blocks, and judging whether the surface dyeing quality of the quilt cover is good or not according to the evaluation value. The invention solves the quality problem of difference from expected color which may occur in the dyeing and coloring process of the quilt cover.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to a quilt cover dyeing quality assessment method based on machine vision.
Background
The quilt cover is used as a common textile product and is closely related to our daily life. With the continuous breakthrough of related industrial theoretical technology and the continuous perfection and development of industrial manufacturing production level, manufacturers have further pursued the quality of quilt cover products so as to meet the current continuous improvement of various consumer groups along with the economic development living standard. The quality of the surface printing and dyeing of the quilt cover directly reflects the quality of the quilt cover, and the surface dyeing process of the quilt cover serving as a textile is a relatively complex procedure, and factors influencing the result of the surface dyeing quality of the quilt cover, such as temperature, dyeing formula, dyeing process, working state of a dyeing machine, operating state of related staff and the like, are more generated in the whole coloring and dyeing process of the quilt cover. The surface of the dyed quilt cover is interfered by the factors, and the color difference phenomenon inconsistent with the expected dyeing and finishing color can occur. Meanwhile, common dyeing quality problems such as color cloud and color spots can also occur.
Disclosure of Invention
The invention provides a machine vision-based quilt cover dyeing quality assessment method for solving the problem of different quality problems possibly occurring in the dyeing and coloring process of the existing quilt cover and different from expected colors, and the adopted technical scheme is as follows:
one embodiment of the invention provides a machine vision-based quilt cover dyeing quality assessment method, which comprises the following steps:
acquiring a surface image of a quilt cover to be detected and a surface image of a standard quilt cover, and acquiring a gray image of the surface of the quilt cover to be detected;
obtaining a chromaticity value, a saturation value and a brightness value of each pixel point in an HSV space of a surface image of a to-be-detected quilt cover and a surface image of a standard quilt cover, normalizing the sum of the first color difference similarities of all the pixel points of the to-be-detected quilt cover and all the pixel points of the standard quilt cover according to the difference value of each pixel point in each HSV channel of the surface image of the to-be-detected quilt cover and the surface image of the standard quilt cover and the standard difference value of all the pixel points of the surface image of the to-be-detected quilt cover and the surface image of the standard quilt cover in different channels to obtain a quilt cover color difference between the to-be-detected quilt cover and the standard quilt cover;
dividing a gray level image of the surface of the to-be-detected quilt into a plurality of super-pixel blocks, and obtaining the pixel point duty ratio of each chromaticity value of each super-pixel block according to the number of pixels corresponding to the chromaticity value of each pixel point in each super-pixel block and the ratio of the number of pixels corresponding to the chromaticity value in the whole surface image of the to-be-detected quilt;
obtaining a chromaticity proportion vector of each super pixel block according to the pixel point occupation ratio of each chromaticity value of each super pixel block, and performing diffusion growth based on the chromaticity proportion vector of each super pixel block to obtain a diffusion coefficient of each super pixel block;
recording any two super pixel blocks as a first super pixel block and a second super pixel block, acquiring a pixel point duty ratio corresponding to the chromaticity value of each pixel point of the first super pixel block as a first ratio, acquiring a pixel point duty ratio corresponding to the chromaticity value of each pixel point of the second super pixel block as a second ratio, acquiring second color difference similarity of the pixel points corresponding to the first ratio and the second ratio, and acquiring a sleeved color distance of the first super pixel block and the second super pixel block according to the accumulated sum of products of the first ratio, the second ratio and the second color difference similarity;
obtaining a surface dyeing significant value of each super pixel block according to the covered color distance between each super pixel block and all super pixel blocks except the super pixel block;
and obtaining a dyeing quality evaluation value of the to-be-detected quilt according to the surface dyeing significant values, the diffusion coefficients and the quilt color difference between the to-be-detected quilt and the standard quilt of all super pixel blocks of the to-be-detected quilt surface image, and judging whether the dyeing quality of the quilt surface is good or not according to the evaluation value.
Preferably, the method for obtaining the first color difference similarity between each pixel point of the to-be-detected quilt cover and the pixel point corresponding to the standard quilt cover according to the difference value of each pixel point on the to-be-detected quilt cover surface image and the standard quilt cover surface image in the HSV respective channels and the standard difference value of all the pixel points of the to-be-detected quilt cover surface image and the standard quilt cover surface image in different channels comprises the following steps:
obtaining a pixel point i from a quilt cover to be detected, finding a pixel point j at a position corresponding to a standard quilt cover surface image, obtaining values of two pixel points in three channels in HSV space, namely a chromaticity value, a saturation value and a brightness value, obtaining standard deviation values of the chromaticity value, the saturation value and the brightness value of all the pixel points of the quilt cover surface image to be detected and the standard quilt cover surface image in the HSV space, marking the standard deviation values as first standard deviations, obtaining a first ratio value by comparing the difference between the ith pixel point and the jth pixel point with the first standard deviation in the channel, and obtaining the first color difference similarity between the ith pixel point and the jth pixel point according to the first ratio value of the three channels.
Preferably, the obtaining the chrominance proportion vector of each super pixel block according to the pixel point duty ratio of each chrominance value of each super pixel block, and performing diffusion growth based on the chrominance proportion vector of each super pixel block to obtain the diffusion coefficient of each super pixel block includes:
obtaining each chromaticity value and corresponding proportion of each super pixel block to form a chromaticity proportion vector of the super pixel block; and then performing diffusion growth based on the super pixel block, wherein the diffusion growth is that when the cosine similarity of the chromaticity proportion vector of any neighborhood super pixel block of the super pixel block and the chromaticity proportion vector of the current super pixel block is larger than a threshold G, marking the neighborhood super pixel block as 1, then performing diffusion growth based on the neighborhood super pixel block until the condition of the threshold G is not met, stopping, and finally taking the number of super pixel blocks marked as 1 after the super pixel block performs diffusion growth as the diffusion number, and normalizing the diffusion number of all the super pixel blocks to obtain the diffusion coefficient of each super pixel block.
Preferably, the calculating method for obtaining the covered color distance of the first superpixel block and the second superpixel block according to the sum of the products of the first ratio, the second ratio and the second color difference similarity comprises the following steps:
in the method, in the process of the invention,
for the number of pixels with the same chromaticity value as the c pixel in the z-th super pixel block of the surface image of the to-be-detected quilt cover, the number of pixels is +.>
For the number of pixels with the same chromaticity value of the whole surface image of the quilt cover to be detected and the c pixel, the number of pixels is +.>
For the proportion of the pixel points with the same chromaticity value as the c pixel point in the z-th super pixel block,/the pixel points are in the same chromaticity value as the c pixel point>
Indicating the +.f in the z1 th super pixel block>
Pixel points with the same chromaticity valueThe ratio, noted as the first ratio,
indicating the +.f in the z2 th super pixel block>
The proportion of the pixels with the same chromaticity value of each pixel is recorded as a second ratio +.>
Super-pixel block for representing surface image of quilt cover to be detected>
Is>
Super-pixel block for representing surface image of quilt cover to be detected>
Is>
Super-pixel block for representing surface image of quilt cover to be detected
Pixel dot +.>
And super pixel block->
Pixel dot +.>
Second color difference similarity of>
First super-pixel block for representing surface image of quilt cover to be detected>
And a second super pixel block->
The color distance between the two quilt covers.
Preferably, the method for obtaining the surface dyeing significance value of each super-pixel block according to the color distance between each super-pixel block and the quilt cover of all super-pixel blocks except the super-pixel block comprises the following steps:
and for the super pixel blocks of the covered surface image, marking any one super pixel block as a third super pixel block, marking each other super pixel block as a fourth super pixel block, obtaining the space distance between the third super pixel block and the fourth super pixel block according to the Euclidean distance between the third super pixel block and the centroid coordinates of each fourth super pixel block, and obtaining the salient parameters of the third super pixel block and each fourth super pixel block according to the space distance between the third super pixel block and each fourth super pixel block and the covered color distance, wherein the accumulated sum of all salient parameters is the surface dyeing salient value of each third super pixel block.
Preferably, the method for obtaining the salient parameters of the third superpixel block and each fourth superpixel block according to the space distance between the third superpixel block and each fourth superpixel block and the quilt cover color distance, wherein the summation of all salient parameters is the surface dyeing salient value of each third superpixel block comprises the following steps:
in the method, in the process of the invention,
for the third super pixel block in the surface image of the quilt cover to be detected +.>
And a fourth super pixel block->
N represents the number of all super pixel blocks, exp () is an exponential function based on a natural constant, ++>
Is spatial weight>
Third super-pixel block for representing surface image of quilt cover to be detected>
And fourth super pixel block->
The color distance of the quilt cover between the two parts is->
The surface dye significance value of the quilt cover of the third super pixel block z1 is represented.
Preferably, the method for obtaining the dyeing quality evaluation value of the to-be-detected quilt cover according to the surface dyeing significant values, the diffusion coefficient and the quilt cover color difference between the to-be-detected quilt cover and the standard quilt cover of all super pixel blocks of the to-be-detected quilt cover surface image comprises the following steps:
judging whether the super pixel block is defective or not according to the surface dyeing significant value of each super pixel block, obtaining a defect mark of each super pixel block, marking the super pixel block with the defect as 1, obtaining the number of defective super pixel blocks, marking the accumulated sum of the products of the covered surface dyeing significant values of all the super pixel blocks and the respective defect marks as a first parameter, taking the ratio of the number of the super pixel blocks with the defect mark as 1 in the super pixel block to the number of all the super pixel blocks as a second parameter, taking the color difference to be detected of the covered as a third parameter, taking the accumulated sum of the diffusion coefficient reciprocal of all the super pixel blocks to be detected as a fourth parameter, and taking the reciprocal of the product of the first parameter, the second parameter, the third parameter and the fourth parameter as a dyeing quality evaluation value.
The beneficial effects of the invention are as follows: according to the embodiment of the invention, the surface of the quilt cover is divided into different super-pixel block areas, and the color distances of the different super-pixel block areas are calculated according to the color characteristic information in the different super-pixel block areas. Compared with the traditional method for comparing differences between direct colors of different pixel points, the color distance of different pixel block areas avoids invalid calculation of differences between pixel points with little difference to a certain extent in calculation. Further, the embodiment of the invention combines the color distance of the pixel block area to construct the corresponding significance value, judges the surface quality abnormal area of the quilt cover according to the significance value, avoids the influence of poor quality evaluation result caused by the fact that the surface quality abnormal area of the quilt cover cannot be normally divided when the traditional global threshold segmentation algorithm is used, and improves the overall accuracy and real-time effect. A plurality of super pixel blocks with similar pixel characteristics can be obtained by adopting super pixel segmentation, then the measurement among the super pixel blocks is carried out so as to rapidly judge the dyeing quality of the quilt cover, the method can be suitable for dyeing quality evaluation of different color quilts, the classification inspection for the different color quilts is not required in the existing method, and the dyeing quality of the quilt cover can also be effectively judged by adopting the iteration measurement among the super pixel blocks when patterns exist in the quilt cover.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions of the prior art, the drawings which are used in the description of the embodiments or the prior art will be briefly described, it being obvious that the drawings in the description below are only some embodiments of the invention, and that other drawings can be obtained according to these drawings without inventive faculty for a person skilled in the art.
Fig. 1 is a flow chart of a machine vision-based method for evaluating dyeing quality of a quilt cover according to an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Referring to fig. 1, a flowchart of a machine vision-based quilt cover dyeing quality assessment method according to an embodiment of the invention is shown, the method includes the following steps:
and S001, shooting and collecting the surface of the quilt cover by using an image collecting device, and processing the obtained surface image of the quilt cover.
The surface of the quilt cover with dyeing quality is different from the standard quilt cover in color, and meanwhile, the surface of the quilt cover is also subjected to quality problems such as color cloud and color spots. In order to obtain a quilt cover surface image with clear colors, a CCD camera is used for shooting and collecting the quilt cover surface to obtain the quilt cover surface image under RGB color space. Considering the condition that insufficient illumination possibly appears in the shooting and collecting working environment, the LED strip-shaped structure light source is used for carrying out illumination treatment on the surface of the quilt cover. Meanwhile, in order to compare and analyze with the standard quilt cover surface, the standard quilt cover surface without large chromatic aberration and colorless cloudy spots on the quilt cover surface needs to be photographed and collected.
In order to reduce the calculation complexity in the subsequent quilt cover dyeing quality evaluation process and avoid errors in the calculation process, different R, G, B channel coefficients are set according to the perception degree of human eyes on different colors, and 0.299, 0.586 and 0.114 are obtained through experience calculation to obtain the gray level image of the quilt cover surface. Meanwhile, in consideration of larger interference generated in calculation of a subsequent quilt cover dyeing quality evaluation process of noise gray levels possibly generated in an acquisition working environment, a Gaussian filtering method is used for eliminating noise influence generated in the acquisition process as much as possible, and a processed quilt cover surface gray level image to be detected and a standard quilt cover surface gray level image are obtained.
And step S002, calculating the color difference of the quilt cover according to the color characteristics of the surface of the quilt cover and the color characteristics of the surface of the standard quilt cover.
First, the color of the whole quilt cover surface may be different from the expected color of the standard quilt cover surface due to the complexity of the process in the whole dyeing and coloring process of the quilt cover. The color difference on the surface of the quilt cover is sometimes not easy to perceive, and the color difference is greatly perceived by a human eye vision system, so that the color difference between the standard quilt cover and the quilt cover to be detected is used as one of judging standards.
Because the surface image of the quilt cover to be detected and the surface image of the standard quilt cover are obtained through the same image acquisition and processing process, the size of the surface image of the quilt cover to be detected is the same as that of the surface image of the standard quilt cover, namely the same number of pixels exists. The surface image of the quilt cover to be detected and the surface image of the standard quilt cover are converted into HSV space, and the difference is more easily detected in the space in saturation and brightness, so that the similarity of the color difference of the pixel points at the same position of the quilt cover to be detected and the standard quilt cover is calculated through the following formula:
in the method, in the process of the invention,
representing the chromaticity value of the ith pixel point of the surface image of the quilt cover to be detected, < + >>
Representing the saturation value of the ith pixel point of the surface image of the quilt cover to be detected,/for the detected surface image>
Indicating the brightness value of the ith pixel point of the surface image of the quilt cover to be detected, < + >>
Color value of j pixel point of standard quilt cover surface image, < >>
Saturation value of jth pixel point of surface image of representing standard quilt cover,/for>
Brightness value of j-th pixel point of standard quilt cover surface image>
、/>
And
the standard deviation values of the chromaticity value, the saturation value and the brightness value of all different pixel points in the surface image of the quilt cover to be detected and the surface image of the standard quilt cover are respectively->
、/>
、/>
The chrominance value weight factor, the saturation value weight factor and the brightness value weight factor are respectively, and an experience value is usually adopted>
=2、/>
，/>
The color difference similarity in HSV color space of the ith pixel point in the surface image of the to-be-detected quilt cover and the jth pixel point in the surface image of the standard quilt cover at the corresponding position is obtained, wherein i=j,/and/or #>
The larger the value, the more similar the two pixels are, and the smaller the color difference.
Calculating the color difference degree of the quilt cover to be detected and the standard quilt cover through the color difference similarity of the corresponding pixel points of the two images:
in the method, in the process of the invention,
representing the number of pixels in the surface image of the quilt cover to be detected, < >>
For the color difference similarity in HSV color space of the ith pixel point in the surface image of the quilt cover to be detected and the jth pixel point in the surface image of the standard quilt cover at the corresponding position, the color difference similarity is->
Representing the difference of the quilt cover color of the to-be-detected quilt cover and the standard quilt cover, and calculating the difference of the quilt cover color between the to-be-detected quilt cover and the standard quilt cover>
When the numerical value of (2) is larger, the quilt cover is more likely to have the problem of dyeing quality in the dyeing process; on the contrary, when the difference of the quilt cover color between the quilt cover to be detected and the standard quilt cover is calculated>
When the number of the quilt cover is smaller, the difference between the quilt cover at the position and the standard quilt cover is not large, namely the expected color obtained on the surface of the quilt cover to be detected is not found, and the problem of dyeing quality is avoided.
And S003, constructing a significance index according to the phenomenon of uneven surface color of the quilt cover.
First, the surface of the quilt cover may be uneven and nonuniform in dyeing, so that color spots appear on the surface of the quilt cover, and the color spots appear in a region aggregation state on the surface of the quilt cover. Therefore, in order to better identify the dyeing quality problem of the surface of the quilt cover caused by the uneven dyeing. In the embodiment of the invention, the surface of the to-be-detected quilt cover is firstly processed by using a super-pixel Segmentation (SLIC) algorithm to obtain N super-pixel blocks, and the embodiment is described with N=100. The super-pixel Segmentation (SLIC) algorithm is a fast segmentation algorithm that divides pixels with high similarity in the whole image into a more representative region according to the feature similarity of different pixels. The image of the surface of the quilt cover to be detected can be divided into different small areas through the super-pixel segmentation algorithm, and if the problem of color cloud and color spots caused by uneven dyeing occurs on the surface of the quilt cover, the color cloud and color spots can occur in the same small area.
In order to identify whether the color cloud and color spot defects exist in the surface image of the quilt cover, according to the dividing result, calculating and analyzing each super pixel block to obtain the color distribution change condition of different super pixel blocks in the surface image of the quilt cover to be detected, firstly calculating the proportion of pixel points with different chromaticity values in each super pixel block, and then calculating the distance between the two super pixel blocks of the quilt cover to be detected, wherein the calculating formula is as follows:
in the method, in the process of the invention,
for the number of pixels with the same chromaticity value as the c pixel in the z-th super pixel block of the surface image of the to-be-detected quilt cover, the number of pixels is +.>
For the number of the pixels with the same chromaticity value as the c-th pixel in the whole surface image of the quilt cover to be detected,
for the proportion of the pixel points with the same chromaticity value as the c pixel point in the z-th super pixel block,/the pixel points are in the same chromaticity value as the c pixel point>
Indicating the +.f in the z1 th super pixel block>
The proportion of the pixels with the same chromaticity value of each pixel is recorded as a first ratio,
indicating the +.f in the z2 th super pixel block>
The proportion of the pixels with the same chromaticity value of each pixel is recorded as a second ratio +.>
Super-pixel block for representing surface image of quilt cover to be detected>
Is>
Super-pixel block for representing surface image of quilt cover to be detected>
Is>
Super-pixel block for representing surface image of quilt cover to be detected
Pixel dot +.>
And super pixel block->
Pixel dot +.>
Second color difference similarity of>
Super-pixel block for representing surface image of quilt cover to be detected>
And super pixel block->
The color distance between the two quilt covers.
And simultaneously acquiring each chroma value and corresponding proportion (the number of the pixels of the chroma value is divided by the number of the pixels of the current super-pixel block) of each super-pixel block to form a chroma proportion vector of the super-pixel block. And then performing diffusion growth based on the super pixel block, firstly marking the current super pixel block as 1, wherein the diffusion growth is that when the cosine similarity of the chromaticity proportion vector of any neighborhood super pixel block of the super pixel block and the chromaticity proportion vector of the current super pixel block is larger than a threshold G, marking the neighborhood super pixel block as 1, then performing diffusion growth based on the neighborhood super pixel block until the threshold G condition is not met, stopping, and taking the empirical value of the threshold G to 0.9, thereby obtaining the diffusion quantity of each super pixel block, namely finally obtaining the quantity (including the self) of the super pixel blocks with the super pixel blocks marked as 1 after the diffusion growth of the super pixel block. For example, if the super pixel block A has 9 neighborhood super pixel blocks, wherein 4 neighborhood super pixel blocks and the super pixel block A meet that the cosine similarity of the chrominance proportion vector is larger than the threshold G, the 4 are marked as 1, and then the four super pixel blocks are used for continuing diffusion growth until stopping. The number of diffusions represents the chrominance spatial distribution of the super-pixel block, the larger it is, the more the chrominance scale vector of the super-pixel block in the image is close to the current super-pixel block, the more it is not isolated, the more likely it is the body color of the quilt cover. After the diffusion number of all the super pixel blocks is obtained, the diffusion number of all the super pixel blocks is subjected to linear normalization to obtain the diffusion coefficient of each super pixel block.
Two different super-pixel blocks of the surface image of the quilt cover to be detected obtained through calculation
And->
The quilt cover color distance between the two>
The number value of the color shade can be effectively judged, and the color shade quality problem caused by uneven dyeing at different positions in the surface of the quilt cover to be detected can be effectively judged. When the color distance of the quilt cover is->
Numerical values of (2)When the difference is larger, the color difference value of all pixel points of the two super-pixel blocks is calculated by the color distance of the jacket, that is, the larger the value is, the larger the difference between all pixel points in the two super-pixel blocks is, the more the difference is, the uneven dyeing is likely to occur, because the super-pixel segmentation is to gather the high-similarity area in the whole image into one super-pixel block, the pixel characteristics in each super-pixel block are very close, and the difference possibly exists between different super-pixel blocks.
In order to further judge the dyeing quality problem of the super pixel blocks, the dyeing significance value of the surface of the quilt cover is obtained by combining the distribution space information of different super pixel blocks through the following calculation
Is of the size of (2):
in the method, in the process of the invention,
for the spatial distance between the z1 th super pixel block and the z2 nd super pixel block of the surface image of the quilt cover, the calculation method is that Euclidean distance of barycenter coordinates of the two super pixel blocks is used as the spatial distance between two different super pixel blocks of the surface image of the quilt cover>
Taking the checked value as the width of the surface image to calculate the spatial information item of the surface of the quilt cover>
The values lie in [0,1]Applying; />
For the number of all super pixel blocks, +.>
Representing the super-image of the surface image of the quilt cover to be detectedPlain block->
1 and super pixel block->
The color distance between the two is covered, meanwhile, the spatial information item value of the super pixel block at a far position is smaller, the spatial information item value of the super pixel block at a near position is larger, and the surface dyeing significance value of the quilt cover is calculated for the current super pixel block>
The higher the contribution of (c).
According to the analysis, the obvious dyeing value of the surface of the quilt cover of different super pixel blocks can be obtained
And when the surface of the quilt cover of the super pixel block is dyed with obvious value +.>
When the size is larger, the super pixel block on the surface of the quilt cover is more likely to generate the defect of color cloud color spots caused by uneven dyeing of the surface of the quilt cover.
And S004, evaluating the dyeing quality of the quilt cover according to the difference of the color of the quilt cover and the obvious value of the dyeing of the surface of the quilt cover.
Due to the obvious value of the surface dyeing of the quilt cover when the super pixel block
When the super pixel block on the surface of the quilt cover is bigger, uneven dyeing of the surface of the quilt cover is more likely to occur, the obvious values of all the super pixel blocks are subjected to linear normalization, and according to an empirical threshold value J=0.7, when:
dyeing significance values of different super pixel blocks according to the surface of the quilt cover
Marking different super pixel blocks, and marking +.>
The method comprises the steps of carrying out a first treatment on the surface of the Otherwise, mark
。
Calculating a quilt cover surface dyeing quality evaluation value qua, wherein the calculation formula is as follows:
calculating according to the above formula to obtain the evaluation value of the surface dyeing quality of the quilt cover
In the formula (I), in the formula (I)>
Marking defect in all super pixel blocks of quilt cover with +.>
The number of super pixel blocks is 1; n is the number of all super pixel blocks on the whole quilt cover surface,
the larger the diffusion coefficient representing the P-th super-pixel block, the more the chrominance scale vector of the super-pixel block in the image is close to the current super-pixel block, the more it is not isolated, the more likely it is the body color of the quilt cover, and the less likely it is the abnormal part. When the surface defect of the quilt cover is more serious, the calculated evaluation value of the dyeing quality of the surface of the quilt cover is +.>
The smaller and closer to 0; on the contrary, when the surface of the quilt cover has no serious dyeing quality defect, the calculated evaluation value of the dyeing quality of the surface of the quilt cover is +.>
The larger and the closer to 1.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (7)
1. The machine vision-based quilt cover dyeing quality assessment method is characterized by comprising the following steps of:
acquiring a surface image of a quilt cover to be detected and a surface image of a standard quilt cover, and acquiring a gray image of the surface of the quilt cover to be detected;
obtaining a chromaticity value, a saturation value and a brightness value of each pixel point in an HSV space of a surface image of a to-be-detected quilt cover and a surface image of a standard quilt cover, normalizing the sum of the first color difference similarities of all the pixel points of the to-be-detected quilt cover and all the pixel points of the standard quilt cover according to the difference value of each pixel point in each HSV channel of the surface image of the to-be-detected quilt cover and the surface image of the standard quilt cover and the standard difference value of all the pixel points of the surface image of the to-be-detected quilt cover and the surface image of the standard quilt cover in different channels to obtain a quilt cover color difference between the to-be-detected quilt cover and the standard quilt cover;
dividing a gray level image of the surface of the to-be-detected quilt into a plurality of super-pixel blocks, and obtaining the pixel point duty ratio of each chromaticity value of each super-pixel block according to the number of pixels corresponding to the chromaticity value of each pixel point in each super-pixel block and the ratio of the number of pixels corresponding to the chromaticity value in the whole surface image of the to-be-detected quilt;
obtaining a chromaticity proportion vector of each super pixel block according to the pixel point occupation ratio of each chromaticity value of each super pixel block, and performing diffusion growth based on the chromaticity proportion vector of each super pixel block to obtain a diffusion coefficient of each super pixel block;
recording any two super pixel blocks as a first super pixel block and a second super pixel block, acquiring a pixel point duty ratio corresponding to the chromaticity value of each pixel point of the first super pixel block as a first ratio, acquiring a pixel point duty ratio corresponding to the chromaticity value of each pixel point of the second super pixel block as a second ratio, acquiring second color difference similarity of the pixel points corresponding to the first ratio and the second ratio, and acquiring a sleeved color distance of the first super pixel block and the second super pixel block according to the accumulated sum of products of the first ratio, the second ratio and the second color difference similarity;
obtaining a surface dyeing significant value of each super pixel block according to the covered color distance between each super pixel block and all super pixel blocks except the super pixel block;
and obtaining a dyeing quality evaluation value of the to-be-detected quilt according to the surface dyeing significant values, the diffusion coefficients and the quilt color difference between the to-be-detected quilt and the standard quilt of all super pixel blocks of the to-be-detected quilt surface image, and judging whether the dyeing quality of the quilt surface is good or not according to the evaluation value.
2. The machine vision-based quilt cover dyeing quality assessment method according to claim 1, wherein the method for obtaining the first color difference similarity between each pixel point of the quilt cover to be detected and the pixel point corresponding to the standard quilt cover according to the difference value of each pixel point on the quilt cover surface image to be detected and the standard quilt cover surface image in the HSV respective channels and the standard difference value of all pixel points of the quilt cover surface image to be detected and the standard quilt cover surface image in different channels is as follows:
obtaining a pixel point i from a quilt cover to be detected, finding a pixel point j at a position corresponding to a standard quilt cover surface image, obtaining values of two pixel points in three channels in HSV space, namely a chromaticity value, a saturation value and a brightness value, obtaining standard deviation values of the chromaticity value, the saturation value and the brightness value of all the pixel points of the quilt cover surface image to be detected and the standard quilt cover surface image in the HSV space, marking the standard deviation values as first standard deviations, obtaining a first ratio value by comparing the difference between the ith pixel point and the jth pixel point with the first standard deviation in the channel, and obtaining the first color difference similarity between the ith pixel point and the jth pixel point according to the first ratio value of the three channels.
3. The machine vision-based quilt cover dyeing quality assessment method according to claim 1, wherein obtaining a chromaticity proportion vector of each super-pixel block according to a pixel point duty ratio of each chromaticity value of each super-pixel block, performing diffusion growth based on the chromaticity proportion vector of each super-pixel block, and obtaining a diffusion coefficient of each super-pixel block comprises:
obtaining each chromaticity value and corresponding proportion of each super pixel block to form a chromaticity proportion vector of the super pixel block; and then performing diffusion growth based on the super pixel block, wherein the diffusion growth is that when the cosine similarity of the chromaticity proportion vector of any neighborhood super pixel block of the super pixel block and the chromaticity proportion vector of the current super pixel block is larger than a threshold G, marking the neighborhood super pixel block as 1, then performing diffusion growth based on the neighborhood super pixel block until the condition of the threshold G is not met, stopping, and finally taking the number of super pixel blocks marked as 1 after the super pixel block performs diffusion growth as the diffusion number, and normalizing the diffusion number of all the super pixel blocks to obtain the diffusion coefficient of each super pixel block.
4. The machine vision-based quilt cover dyeing quality assessment method according to claim 1, wherein the method for calculating the quilt cover color distance of the first super-pixel block and the second super-pixel block according to the sum of the products of the first ratio, the second ratio and the second color difference similarity is as follows:
in the method, in the process of the invention,
for the number of pixels with the same chromaticity value as the c pixel in the z-th super pixel block of the surface image of the to-be-detected quilt cover, the number of pixels is +.>
For the number of pixels with the same chromaticity value of the whole surface image of the quilt cover to be detected and the c pixel, the number of pixels is +.>
For the proportion of the pixel points with the same chromaticity value as the c pixel point in the z-th super pixel block,/the pixel points are in the same chromaticity value as the c pixel point>
Indicating the +.f in the z1 th super pixel block>
The proportion of the pixels with the same chromaticity value of each pixel is recorded as a first ratio, +.>
Indicating the +.f in the z2 th super pixel block>
The proportion of the pixels with the same chromaticity value of each pixel is recorded as a second ratio +.>
Super-pixel block for representing surface image of quilt cover to be detected>
Is>
Super-pixel block for representing surface image of quilt cover to be detected>
Is>
Super-pixel block for representing surface image of quilt cover to be detected>
In (3) pixels
And super pixel block->
Pixel dot +.>
Second color difference similarity of>
First super-pixel block for representing surface image of quilt cover to be detected>
And a second super pixel block->
The color distance between the two quilt covers.
5. The machine vision-based quilt cover dyeing quality assessment method according to claim 1, wherein the method for obtaining the surface dyeing significance value of each super-pixel block according to the quilt cover color distance of each super-pixel block and all super-pixel blocks except for the super-pixel block is as follows:
and for the super pixel blocks of the covered surface image, marking any one super pixel block as a third super pixel block, marking each other super pixel block as a fourth super pixel block, obtaining the space distance between the third super pixel block and the fourth super pixel block according to the Euclidean distance between the third super pixel block and the centroid coordinates of each fourth super pixel block, and obtaining the salient parameters of the third super pixel block and each fourth super pixel block according to the space distance between the third super pixel block and each fourth super pixel block and the covered color distance, wherein the accumulated sum of all salient parameters is the surface dyeing salient value of each third super pixel block.
6. The machine vision based quilt cover dyeing quality assessment method according to claim 5, wherein the method for obtaining the salient parameters of the third super pixel block and each fourth super pixel block according to the space distance between the third super pixel block and each fourth super pixel block and the quilt cover color distance, wherein the sum of all salient parameters is the surface dyeing salient value of each third super pixel block is as follows:
in the method, in the process of the invention,
for the third super pixel block in the surface image of the quilt cover to be detected +.>
And a fourth super pixel block->
N represents the number of all super pixel blocks, exp () is an exponential function based on a natural constant, ++>
Is spatial weight>
Third super-pixel block for representing surface image of quilt cover to be detected>
And fourth super pixel block->
The color distance of the quilt cover between the two parts is->
The surface dye significance value of the quilt cover of the third super pixel block z1 is represented.
7. The machine vision-based quilt cover dyeing quality assessment method according to claim 1, wherein the method for obtaining the dyeing quality assessment value of the quilt cover to be detected according to the surface dyeing significance values, the diffusion coefficients and the quilt cover color difference between the quilt cover to be detected and the standard quilt cover of all super pixel blocks of the surface image of the quilt cover to be detected is as follows:
judging whether the super pixel block is defective or not according to the surface dyeing significant value of each super pixel block, obtaining a defect mark of each super pixel block, marking the super pixel block with the defect as 1, obtaining the number of defective super pixel blocks, marking the accumulated sum of the products of the covered surface dyeing significant values of all the super pixel blocks and the respective defect marks as a first parameter, taking the ratio of the number of the super pixel blocks with the defect mark as 1 in the super pixel block to the number of all the super pixel blocks as a second parameter, taking the color difference to be detected of the covered as a third parameter, taking the accumulated sum of the diffusion coefficient reciprocal of all the super pixel blocks to be detected as a fourth parameter, and obtaining a dyeing quality evaluation value according to the products of the first parameter, the second parameter, the third parameter and the fourth parameter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310180939.XA CN115880297B (en) | 2023-03-01 | 2023-03-01 | Quilt cover dyeing quality assessment method based on machine vision |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310180939.XA CN115880297B (en) | 2023-03-01 | 2023-03-01 | Quilt cover dyeing quality assessment method based on machine vision |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115880297A CN115880297A (en) | 2023-03-31 |
| CN115880297B true CN115880297B (en) | 2023-05-05 |
Family
ID=85761709
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310180939.XA Active CN115880297B (en) | 2023-03-01 | 2023-03-01 | Quilt cover dyeing quality assessment method based on machine vision |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115880297B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116612110B (en) * | 2023-07-14 | 2023-10-24 | 微山县振龙纺织品有限公司 | Intelligent quality assessment method for gradual change printing and dyeing effect |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111127437A (en) * | 2019-12-25 | 2020-05-08 | 浙江传媒学院 | Full-reference image quality evaluation method based on color space decomposition |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110084782B (en) * | 2019-03-27 | 2022-02-01 | 西安电子科技大学 | Full-reference image quality evaluation method based on image significance detection |
| CN113298763B (en) * | 2021-05-09 | 2022-08-19 | 复旦大学 | Image quality evaluation method based on significance window strategy |
| CN115035111B (en) * | 2022-08-10 | 2022-10-25 | 江苏永银化纤有限公司 | Colored silk color difference detection method based on image processing |
| CN115294116B (en) * | 2022-10-08 | 2022-12-30 | 南通梦洁家纺有限公司 | Method, device and system for evaluating dyeing quality of textile material based on artificial intelligence |
-
2023
- 2023-03-01 CN CN202310180939.XA patent/CN115880297B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111127437A (en) * | 2019-12-25 | 2020-05-08 | 浙江传媒学院 | Full-reference image quality evaluation method based on color space decomposition |
Non-Patent Citations (1)
| Title |
|---|
| 基于Sigmoid 核函数的纺织品色差分类检测研究;谈波 等;西昌学院学报（自然科学版）;第35卷(第2期);第89-93页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115880297A (en) | 2023-03-31 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US9750326B2 (en) | Transparency evaluation device, transparency evaluation method and transparency evaluation program | |
| CN108323204B (en) | Method for detecting face flaw point and intelligent terminal | |
| CN103034838B (en) | A kind of special vehicle instrument type identification based on characteristics of image and scaling method | |
| CN104636759B (en) | A kind of method and picture filter information recommendation system for obtaining picture and recommending filter information | |
| CN115880297B (en) | Quilt cover dyeing quality assessment method based on machine vision | |
| CN110335312A (en) | A kind of object space localization method neural network based and device | |
| CN106503644B (en) | Glasses attribute detection method based on edge projection and color characteristic | |
| US20170079599A1 (en) | Moisture feeling evaluation device, moisture feeling evaluation method, and moisture feeling evaluation program | |
| CN105574514B (en) | The raw tomato automatic identifying method in greenhouse | |
| CN108537239A (en) | A kind of method of saliency target detection | |
| CN107180439B (en) | Color cast characteristic extraction and color cast detection method based on Lab chromaticity space | |
| WO2019168142A1 (en) | Image classification method, device, and program | |
| CN111062891A (en) | Image processing method, device, terminal and computer readable storage medium | |
| CN107993209A (en) | Image processing method, device, computer-readable recording medium and electronic equipment | |
| CN111624203A (en) | Relay contact alignment non-contact measurement method based on machine vision | |
| CN109583330B (en) | Pore detection method for face photo | |
| CN101996317A (en) | Method and device for identifying markers in human body | |
| CN107909542A (en) | Image processing method, device, computer-readable recording medium and electronic equipment | |
| CN111738931A (en) | Shadow removal algorithm for aerial image of photovoltaic array unmanned aerial vehicle | |
| CN113642358B (en) | Skin color detection method, device, terminal and storage medium | |
| CN107239761B (en) | Fruit tree branch pulling effect evaluation method based on skeleton angular point detection | |
| Park et al. | Automated defect inspection systems by pattern recognition | |
| CN116612110B (en) | Intelligent quality assessment method for gradual change printing and dyeing effect | |
| CN110751690B (en) | Visual positioning method for milling machine tool bit | |
| CN107886549A (en) | A kind of dermatoglyphic pattern of the fabric color transfer method based on braiding grain details enhancing |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |