CN110047074A - The fiber of textile mixes content detection, reverse engineering analysis method and equipment - Google Patents
The fiber of textile mixes content detection, reverse engineering analysis method and equipment Download PDFInfo
- Publication number
- CN110047074A CN110047074A CN201910412007.7A CN201910412007A CN110047074A CN 110047074 A CN110047074 A CN 110047074A CN 201910412007 A CN201910412007 A CN 201910412007A CN 110047074 A CN110047074 A CN 110047074A
- Authority
- CN
- China
- Prior art keywords
- fiber
- categories
- textile
- image
- fibre
- 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
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/36—Textiles
- G01N33/367—Fabric or woven textiles
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30108—Industrial image inspection
- G06T2207/30124—Fabrics; Textile; Paper
Abstract
The fiber that the present invention discloses a kind of textile mixes detection method of content, including inspection textile is disassembled into fiber dust；Digital image acquisition is carried out to all fibres in fiber dust, obtains several fibre images；The instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and calculates the average diameter and variance for obtaining fiber of all categories simultaneously；Computer deep learning and image, semantic parted pattern are marked and combined according to preliminary, semantic segmentation is carried out to fibre image, obtains semantic segmentation image；In conjunction with average diameter and variance, all fibres of all categories in semantic segmentation image are analyzed and calculated, obtains the mixing content of fiber of all categories in inspection textile.The detection method effectively improves the detection efficiency and accuracy of textile fiber content, can also help to establish a set of reverse-engineering for being applicable in textile designs, be conducive to the processing and production of textile by combining computer deep learning and image, semantic to divide.
Description
Technical field
This application involves the fibers of the technical field of quality of textile products detection more particularly to a kind of textile to mix content inspection
Survey method, reverse engineering analysis method and relevant device.
Background technique
In the textile industry, the specific of kinds of fibers in detection blending product and statistics wherein various fibers is generally required
Content, currently used fiber, which mixes detection method of content, mainly manual Split Method, chemical dissolution method, microscopic method.
Manual Split Method is to differentiate the fiber distinguished for range estimation, using manual fractionation, drying, is weighed, to calculate
Fiber mass content out.This method is time-consuming and laborious, and is easy to appear the feelings that missing inspection, false retrieval are caused because of operator's experience deficiency
Condition influences the testing result that fiber mixes content；
Chemical dissolution method is to make the fiber of blended fabric using solubility property of the various fibers in different chemical solvents
Component separation, insoluble residue is weighed, the ratio of dissolved constituent is calculated by mass loss.Relative to manual fractionation
Method, although the testing result of chemolysis method is more accurate, chemical dissolution method needs to predict in advance may in tested fabric
Containing kinds of fibers, then corresponding chemical solvent and experimental group are set, it is seen that it compares the prediction for relying on early period, and forecasting inaccuracy is true
It is easy for causing chemical solvent and experimental group quantity that insufficient or excessive situation is arranged, testing result is caused to have deviation or part molten
The problem of agent, experimental group extra waste.
Microscopic method is the kind for differentiating each fiber in tested fabric by microscope (microscopic projector) by operator
Class and the radical for counting various fibers, measure the diameter of fiber, calculate the volume of fiber, are calculated according to known fiber specific gravity
The weight of fiber, so that concrete content of the various fibers in blending product is calculated, although this method can be obtained more intuitively
Know that the fiber of tested fabric constitutes situation, but it has subjectivity and differentiates that type of fibers, measuring speed is slow, working efficiency is low, measurement
The low disadvantage of precision.
As it can be seen that existing fiber mixes the equal existing defects of detection method of content, it is real content can not to be mixed to textile fiber
It applies and fast and accurately detects, be unfavorable for the processing and production of textile, therefore, how to obtain one kind can solve the above problem
Detection method, it has also become the important subject of those skilled in the art.
Summary of the invention
The technical problem to be solved in the present invention is that in view of the deficiencies of the prior art, the fiber for providing a kind of textile is mixed
With detection method of content, reverse engineering analysis method and relevant device, the detection method by combine computer deep learning and
Image, semantic segmentation, it is effective to improve the detection effect that fabric fibre mixes content to overcome problem in above-mentioned background technique
Rate and accuracy can also help to establish a set of reverse engineering analysis method for being applicable in textile designs, be conducive to adding for textile
Work and production.
To achieve the above object, the present invention provides technical solution below:
In a first aspect, the present invention provides a kind of fibers of textile to mix detection method of content, comprising:
Inspection textile is disassembled into fiber dust；Wherein, all fibres in fiber dust are single fiber state；
Digital image acquisition is carried out to all fibres in fiber dust, until obtaining that inspection textile entirety can be represented
Several fibre images of performance；
The instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and is counted simultaneously
Calculate the average diameter and variance for obtaining fiber of all categories；Wherein, the classification of fiber includes material, color, shape, technique；
Computer deep learning and image, semantic parted pattern are marked and combined according to preliminary, fibre image is carried out semantic
Segmentation, all fibres of all categories in fibre image are respectively divided and are identified, obtain semantic segmentation image；
In conjunction with the average diameter and variance of fiber of all categories, all fibres of all categories in semantic segmentation image are divided
It analyses and calculates, obtain the mixing content of fiber of all categories in inspection textile.
Optionally, the average diameter and variance of combination fiber of all categories, to institute of all categories in semantic segmentation image
There is fiber to be analyzed and calculated, obtain the mixing content of fiber of all categories in inspection textile, specifically include:
Pixel shared by the longitudinal cross-section to the fiber of all categories in semantic segmentation image counts respectively, obtains each
The sum of the longitudinal cross-section area of classification fiber；
According to the following formula, fibre of all categories in inspection textile is calculated in conjunction with the average diameter and variance of fiber of all categories
The weight percent of dimension obtains the mixing content of fiber of all categories in inspection textile:
Wherein, ρ1、ρ2...ρmThe density of fiber respectively of all categories；n1、n2...nmThe longitudinal direction of fiber respectively of all categories
The sum of area of section；ξd1、ξd2...ξdmThe stochastic variable of the diameter of fiber respectively of all categories, E are mathematic expectaion, and D is variance.
Optionally, described to disassemble inspection textile at fiber dust, it specifically includes:
Inspection textile is disassembled into single fiber state；
All fibres are carried out to cut sample preparation, obtain fiber dust；
Wherein, each fibre length of fiber dust is equal.
Optionally, if inspection textile includes achromatic fibrils, all fibres in fiber dust carry out digitized map
As acquisition, until several fibre images of inspection textile overall performance can be represented by obtaining, further includes:
Fibre staining differentiating solvent is instilled toward fiber dust, the fiber of all categories in fiber dust is dyed.
Optionally, the past fiber dust instills fibre staining differentiating solvent, carries out to the fiber of all categories in fiber dust
Dyeing, specifically includes:
The first coloring solution that fabric sample can be completely covered is instilled toward fiber dust；
The second coloring solution is instilled toward fiber dust；
Wherein, first coloring solution is 6.25*10-3The halide or polyhalide and 2.5*10 of~1.44g/ml- 3Mg/ml glycerol mixed aqueous solutions, second coloring solution be 0.200lg/ml halide or polyhalide it is water-soluble
Liquid.
Second aspect, the present invention provides a kind of reverse engineering analysis method of textile, suitable for textile design,
Production, detection, comprising:
Detection method of content is mixed by the fiber of any of the above-described kind of textile to detect inspection textile, is sent
Examine the mixing content of fiber of all categories in textile；
The corresponding design of mixing content configuration, the production decision obtained according to detection, produces textile.
Optionally, the corresponding design of mixing content configuration, the production decision obtained according to detection, produces textile,
It specifically includes:
Several different classes of fibers are chosen according to the mixing content that detection obtains；
Several different classes of fibers are subjected to weight proportion, mixing, obtain mixed raw material；
Textile is produced using mixed raw material.
The third aspect, the present invention also provides a kind of fibers of textile to mix content detection equipment, for realizing such as originally
The mixing content detection of fiber provided by invention any embodiment or reverse engineering analysis method, including textile disassembling apparatus,
Fibre image acquisition device, processing terminal and the computer input terminal for receiving operator command；
The textile disassembling apparatus, for disassembling inspection textile at fiber dust；Wherein, the institute in fiber dust
Having fiber is single fiber state；
The fibre image acquisition device, for carrying out digital image acquisition to all fibres in fiber dust, until
Obtain to represent several fibre images of inspection textile overall performance；
The processing terminal includes:
First processing units carry out the fiber of all categories in fibre image for the instruction of response computer input terminal
Preliminary mark, and the average diameter and variance for obtaining fiber of all categories are calculated simultaneously；Wherein, the classification of fiber includes material, face
Color, shape, technique；
The second processing unit, for marking and combining computer deep learning and image, semantic parted pattern according to preliminary,
Semantic segmentation is carried out to fibre image, all fibres of all categories in fibre image are respectively divided and are identified, obtains semantic point
Cut image；
Third processing unit, for combining the average diameter and variance of fiber of all categories, to all kinds of in semantic segmentation image
Other all fibres are analyzed and are calculated, and the mixing content of fiber of all categories in inspection textile is obtained.
Compared with prior art, the invention has the following advantages:
Detection method of content and equipment are mixed the present invention provides a kind of fiber of textile, and this method is by being disassembled
Digital image acquisition is carried out for the inspection textile of single fiber state, using operator to different classes of in few fibers image
Fiber tentatively marked, accurately mark reference can be made for the actual conditions that detect every time, subsequent computer root again
Mark and combine computer deep learning and Image Segmentation Model according to preliminary, to the fiber of all categories in all fibres image into
Line identifier, finally by counting and being calculated the mixing content of fiber of all categories in textile, thus using artificial experience and
The mode that computer technology combines overcomes traditional fibre to mix the defect of detection method of content, improves textile fiber and mixes
The efficiency and accuracy of content detection work；The present invention also provides a kind of reverse engineering analysis method of textile, this method
Inspection textile is detected by above-mentioned detection method, configures corresponding production decision further according to obtained mixing content,
It is advantageously implemented the production and processing of fabric product.
Detailed description of the invention
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technical description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
Some embodiments of invention without any creative labor, may be used also for those of ordinary skill in the art
To obtain other attached drawings according to these attached drawings.
Fig. 1 is that a kind of fiber for textile that the embodiment of the present invention one provides mixes the process signal of detection method of content
Figure；
Fig. 2 is that a kind of fiber of textile provided by Embodiment 2 of the present invention mixes the process signal of detection method of content
Figure；
Fig. 3 is that a kind of fiber for textile that the embodiment of the present invention three provides mixes the process signal of detection method of content
Figure；
Fig. 4 is that a kind of fiber for textile that the embodiment of the present invention four and embodiment five provide mixes detection method of content
Flow diagram；
Fig. 5 is a kind of method flow diagram for the reverse engineering analysis method that the embodiment of the present invention six provides textile；
Fig. 6 is that a kind of fiber for textile that the embodiment of the present invention seven provides mixes the structural representation of content detection equipment
Figure.
Specific embodiment
To enable the purpose of the present invention, feature, advantage more obvious and understandable, implement below in conjunction with the present invention
Attached drawing in example, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that reality disclosed below
Applying example is only a part of the embodiment of the present invention, and not all embodiment.Based on the embodiments of the present invention, this field is common
Technical staff's all other embodiment obtained without making creative work belongs to the model that the present invention protects
It encloses.
In the description of the present invention, it is to be understood that, when a component is considered as " connection " another component, it can
To be directly to another component or may be simultaneously present the component being centrally located.When a component is considered as " setting
Set " another component, it, which can be, is set up directly on another component or may be simultaneously present the component being centrally located.
In addition, the indicating positions such as term " length " " short " "inner" "outside" or positional relationship for the orientation that is shown based on attached drawing or
Person's positional relationship is merely for convenience of the description present invention, rather than the device or original part of indication or suggestion meaning must have this
Specific orientation is operated with specific orientation construction, should not be understood as limitation of the invention with this.
To further illustrate the technical scheme of the present invention below with reference to the accompanying drawings and specific embodiments.
Embodiment one
Fig. 1 is please referred to, which mixes content detection side
The flow diagram of method, the detection method can be by being realized, by combining meter based on the detection system of optical microscopy
Calculation machine deep learning and image, semantic segmentation, overcome the defect that traditional fibre mixes detection method of content, and effective improve is knitted
Fibres mix the detection efficiency and accuracy of content, this detection method includes:
S101: inspection textile is disassembled into fiber dust；
Wherein, all fibres in fiber dust are single fiber state；
S102: carrying out digital image acquisition to all fibres in fiber dust, until obtaining that inspection weaving can be represented
Several fibre images of product overall performance；
The present embodiment preferably uses optical micro imaging device plant fiber image collecting device, fibre image acquisition device
It is electrically connected transmission and calling that computer is achieved in data；Specifically, fiber dust is placed on optical micro imaging device
Platform specimen holder on, fibre image is amplified by optical microscope system, the X of sample, Y direction adjustment feeding,
Fiber enlarged drawing is not repeatedly acquired, fiber longitudinal cross-section image is converted by digital picture using imaging sensor, as
Sample image is transmitted to computer；
S103: the instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and same
When calculate the average diameter and variance for obtaining fiber of all categories；
Wherein, the classification of fiber includes material, color, shape, technique；
It is some for carrying out human-computer interaction device, meter with computer that the input terminal can be keyboard, mouse or touch screen etc.
The memory of calculation machine is stored with software or readable program for marking type of fibers；Specifically, operator by computer and
Relevant computer input terminal observes the fibre image by amplification, and computer response operator inputs in input terminal
Instruction carries out preliminary artificial mark to the fiber of all categories in small part sample image, and calculates and obtain unlike material, face
Color, shape, technique fiber average diameter and variance；
Preferably, at least 300 fibers in fibre image are tentatively marked, to be subsequent computer depth
Study provides sufficient learning sample, can improve computer learning efficiency while simplifying the complexity of computer learning, have
Help the operational paradigm of subsequent image semantic segmentation；
S104: marking and combine computer deep learning and image, semantic parted pattern according to preliminary, to fibre image into
All fibres of all categories in fibre image are respectively divided and are identified, obtain semantic segmentation image by row semantic segmentation；
Specifically, computer will tentatively mark the fiber of resulting unlike material, color, shape, technique, as study sample
This carries out semantic segmentation to all fibre images, by sample drawing in conjunction with computer deep learning and image, semantic parted pattern
The fiber of background and unlike material, color, shape, technique as in is respectively divided and identifies, and obtains semantic segmentation image；
S105: in conjunction with the average diameter and variance of fiber of all categories, to all fibres of all categories in semantic segmentation image
It is analyzed and is calculated, obtain the mixing content of fiber of all categories in inspection textile.
It present embodiments provides the present invention provides a kind of fiber of textile mixing detection method of content, this method passes through
Digital image acquisition is carried out to the inspection textile being disassembled as single fiber state, using operator in few fibers image
Different classes of fiber is tentatively marked, and accurately mark reference can be made for the actual conditions detected every time, is then counted
Calculation machine is further according to preliminary mark and combines computer deep learning and Image Segmentation Model, to of all categories in all fibres image
Fiber be identified, finally by counting and being calculated the mixing content of fiber of all categories in inspection textile, thus sharp
The manually mode that experience and computer technology combine overcomes traditional fibre to mix the defect of detection method of content, improves and spins
Fabric fiber mixes the efficiency and accuracy of content detection work.
Embodiment two
Referring to figure 2., which mixes content detection side
The flow diagram of method；The present embodiment on the basis of example 1, to step " average diameter and side in conjunction with fiber of all categories
Difference is analyzed and is calculated to all fibres of all categories in semantic segmentation image, obtains fiber of all categories in inspection textile
Mixing content " implement further optimization, it may be assumed that
Pixel shared by the longitudinal cross-section to the fiber of all categories in semantic segmentation image counts respectively, obtains each
The sum of the longitudinal cross-section area of classification fiber；
According to the following formula, fibre of all categories in inspection textile is calculated in conjunction with the average diameter and variance of fiber of all categories
The weight percent of dimension obtains the mixing content of fiber of all categories in inspection textile:
Wherein, ρ1、ρ2...ρmThe density of fiber respectively of all categories；n1、n2...nmThe longitudinal direction of fiber respectively of all categories
The sum of area of section；ξd1、ξd2...ξdmThe stochastic variable of the diameter of fiber respectively of all categories, E are mathematic expectaion, and D is variance.
Specifically, the fiber that such as inspection textile contains two different classifications；
The percentage composition of one of fiber are as follows:
And the percentage composition of another fiber are as follows:
Further, containing there are many calculation method of the inspection textile of different classes of fiber can with and so on.
Based on above-mentioned optimization, as shown in Fig. 2, a kind of fiber of textile provided in this embodiment mixes content detection side
Method may comprise steps of:
S201: inspection textile is disassembled into fiber dust；
Wherein, all fibres in fiber dust are single fiber state；
S202: carrying out digital image acquisition to all fibres in fiber dust, until obtaining that inspection weaving can be represented
Several fibre images of product overall performance；
S203: the instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and same
When calculate the average diameter and variance for obtaining fiber of all categories；
Wherein, the classification of fiber includes material, color, shape, technique；
Preferably, at least 300 fibers in fibre image are tentatively marked, to be subsequent computer depth
Study provides sufficient learning sample, can improve computer learning efficiency while simplifying the complexity of computer learning, have
Help the operational paradigm of subsequent image semantic segmentation；
S204: marking and combine computer deep learning and image, semantic parted pattern according to preliminary, to fibre image into
All fibres of all categories in fibre image are respectively divided and are identified, obtain semantic segmentation image by row semantic segmentation；
S205: pixel shared by the longitudinal cross-section to the fiber of all categories in semantic segmentation image counts respectively,
Obtain the sum of the longitudinal cross-section area of fiber of all categories；
It, can be intuitively anti-using pixel since computer cannot directly measure the longitudinal cross-section size of various fibers
Mirror the longitudinal cross-section area of various fibers；
S206: the weight of fiber of all categories in inspection textile is calculated in conjunction with the average diameter and variance of fiber of all categories
Percentage obtains the mixing content of fiber of all categories in inspection textile；
Specifically, the weight percent with bigger Practical significance is converted by the area of different fibers, so that detection knot
By being easier to understand, be conducive to the production and processing of textile.
The fiber for present embodiments providing a kind of textile mixes detection method of content, and this method is by being disassembled as list
The inspection textile of fiber condition carries out digital image acquisition, using operator to fibre different classes of in few fibers image
Dimension is tentatively marked, and accurately mark reference can be made for the actual conditions detected every time, subsequent computer is further according to first
Step marks and combines computer deep learning and Image Segmentation Model, marks to the fiber of all categories in all fibres image
Know, finally by counting and being calculated the mixing content of fiber of all categories in inspection textile, thus using artificial experience and
The mode that computer technology combines overcomes traditional fibre to mix the defect of detection method of content, improves textile fiber and mixes
The efficiency and accuracy of content detection work.
Embodiment three
Referring to figure 3., a kind of fiber of the textile for being illustrated as the offer of the embodiment of the present invention three mixes content detection side
The flow diagram of method；Inspection textile on the basis of example 1, " is disassembled into fiber dust " step by the present embodiment
Implement further optimization, specifically, this detection method may include steps of:
S301: inspection textile is disassembled into single fiber state；
S302: all fibres are carried out to cut sample preparation, obtain fiber dust；
Wherein, the present embodiment can cut inspection textile using Kazakhstan formula food slicer as textile disassembling apparatus
Piece sample preparation；The equal length of each fiber in the fiber dust being disassembled, therefore using the longitudinal area of not classification fiber
With the ratio between replace the ratio between different classes of number of fiber in conventional method, thus avoid fiber is interlaced, stacking and lead to fiber
The problem of radical judgement inaccuracy, improves the accuracy of detection；
S303: carrying out digital image acquisition to all fibres in fiber dust, until obtaining that inspection weaving can be represented
Several fibre images of product overall performance；S304: the instruction of response computer input terminal, to the fibre of all categories in fibre image
Dimension is tentatively marked, and calculates the average diameter and variance for obtaining fiber of all categories simultaneously；
Wherein, the classification of fiber includes material, color, shape, technique；
Preferably, at least 300 fibers in fibre image are tentatively marked, to be subsequent computer depth
Study provides sufficient learning sample, can improve computer learning efficiency while simplifying the complexity of computer learning, have
Help the operational paradigm of subsequent image semantic segmentation；
S305: marking and combine computer deep learning and image, semantic parted pattern according to preliminary, to fibre image into
All fibres of all categories in fibre image are respectively divided and are identified, obtain semantic segmentation image by row semantic segmentation；
Specifically, computer using tentatively mark resulting unlike material, color, shape, technique fiber as study sample
This carries out semantic segmentation to all fibre images, by sample in conjunction with by computer deep learning and image, semantic parted pattern
The fiber of background and unlike material, color, shape, technique in product image is respectively divided and identifies, and obtains semantic segmentation figure
Picture；
S306: in conjunction with the average diameter and variance of fiber of all categories, to all fibres of all categories in semantic segmentation image
It is analyzed and is calculated, obtain the mixing content of fiber of all categories in inspection textile.
Specifically, using this detection method respectively to the inspection textile of certain melange yarn, it is therefore intended that, obtain this quilt
The ratio of the fiber of various colors in textile is examined, fiber mixes content and is as follows:
Example IV
Referring to figure 4., which mixes
The flow diagram of detection method of content；The present embodiment on the basis of example 1, to step " to all in fiber dust
Implement until fiber progress digital image acquisition, several fibre images until obtaining to represent textile overall performance "
Further optimization；
Specifically, this detection method may include steps of if inspection textile includes a variety of achromatic fibrils:
S401: inspection textile is disassembled into fiber dust；
Wherein, all fibres in fiber dust are single fiber state；
S402: fibre staining differentiating solvent is instilled toward textured fiber powder, the fiber of all categories in fiber dust is dyed；
Since achromatic fibrils are difficult to differentiate between, thus it is different physically or chemically according to different fibers, it is dyed,
Help rapidly and accurately to offer an explanation various fibers；
S403: carrying out digital image acquisition to all fibres in fiber dust, until obtaining that inspection weaving can be represented
Several fibre images of product overall performance；
S404: the instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and same
When calculate the average diameter and variance for obtaining fiber of all categories；
Wherein, the classification of fiber includes material, color, shape, technique；
Preferably, at least 300 fibers in fibre image are tentatively marked, to be subsequent computer depth
Study provides sufficient learning sample, can improve computer learning efficiency while simplifying the complexity of computer learning, have
Help the operational paradigm of subsequent image semantic segmentation；
S405: marking and combine computer deep learning and image, semantic parted pattern according to preliminary, to fibre image into
All fibres of all categories in fibre image are respectively divided and are identified, obtain semantic segmentation image by row semantic segmentation；
Specifically, computer will tentatively mark the fiber of resulting unlike material, color, shape, technique, as study sample
This carries out semantic segmentation to all fibre images, by sample drawing in conjunction with computer deep learning and image, semantic parted pattern
The fiber of background and unlike material, color, shape, technique as in is respectively divided and identifies, and obtains semantic segmentation image；
S406: in conjunction with the average diameter and variance of fiber of all categories, to all fibres of all categories in semantic segmentation image
It is analyzed and is calculated, obtain the mixing content of fiber of all categories in inspection textile.
Preferably, " fibre staining differentiating solvent is instilled toward fiber dust, to the fibre of all categories in fiber dust to step S402
Dimension is dyed " it advanced optimizes, it specifically includes:
S4021: the first coloring solution that fabric sample can be completely covered is instilled toward fiber dust；
S4022: 1~2 the second coloring solution of drop is instilled toward fiber dust；
Wherein, first coloring solution is 6.25*10-3The halide or polyhalide and 2.5*10 of~1.44g/ml- 3Mg/ml glycerol mixed aqueous solutions, second coloring solution be 0.200lg/ml halide or polyhalide it is water-soluble
Liquid.
Specifically, being examined respectively to certain mao of terylene fabric, linen-cotton reeled yarn, linen-cotton modal fabric using this detection method
It surveys, it is therefore intended that, the ratio of various fibers in these three tested textiles is obtained, fiber mixes content and is as follows:
Embodiment five
Referring to figure 4., which mixes
The flow diagram of detection method of content；The present embodiment on the basis of example 1, to step " to all in fiber dust
Implement until fiber progress digital image acquisition, several fibre images until obtaining to represent textile overall performance "
Further optimization；
Specifically, this detection method may include steps of if inspection fabric includes a variety of achromatic fibrils:
S401: inspection textile is disassembled into fiber dust；
Wherein, all fibres in fiber dust are single fiber state；
S402: fibre staining differentiating solvent is instilled toward textured fiber powder, the fiber of all categories in fiber dust is dyed；
Since achromatic fibrils are difficult to differentiate between, thus it is different physically or chemically according to different fibers, it is dyed,
Help rapidly and accurately to offer an explanation various fibers；
S403: carrying out digital image acquisition to all fibres in fiber dust, until obtaining that inspection weaving can be represented
Several fibre images of product overall performance；
S404: the instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and same
When calculate the average diameter and variance for obtaining fiber of all categories；
Wherein, the classification of fiber includes material, color, shape, technique；
S405: marking and combine computer deep learning and image, semantic parted pattern according to preliminary, to fibre image into
All fibres of all categories in fibre image are respectively divided and are identified, obtain semantic segmentation image by row semantic segmentation；
S406:
In conjunction with the average diameter and variance of fiber of all categories, all fibres of all categories in semantic segmentation image are divided
It analyses and calculates, obtain the mixing content of fiber of all categories in inspection textile.
Preferably, " fibre staining differentiating solvent is instilled toward fiber dust, to the fibre of all categories in fiber dust to step S402
Dimension is dyed " it advanced optimizes, it specifically includes:
S4021: the first coloring solution that fabric sample can be completely covered is instilled toward fiber dust；
S4022: 1~2 the second coloring solution of drop is instilled toward fiber dust；
Wherein, first coloring solution is 6.25*10-3The halide or polyhalide and 2.5*10 of~1.44g/ml- 3Mg/ml glycerol mixed aqueous solutions, second coloring solution be 0.200lg/ml halide or polyhalide it is water-soluble
Liquid.
Specifically, the halide or polyhalide are EXn, wherein E be potassium, lithium, sodium, magnesium, beryllium, aluminium or zinc, X be fluorine,
Chlorine, bromine or iodine, n are at least 1.
Specifically, being detected respectively to two kinds of unknown fabric products of ingredient using this detection method, it is therefore intended that, it obtains
Obtain the ratio of various fibers and color in two kinds of tested textiles, wherein both tested textiles are numbered respectively is
FX1002 and FX1003, fiber mix content and are as follows:
Embodiment six
To design by customer satisfaction textile, usually by differences such as unlike material, color, geometry, techniques
The fiber of type carries out blended or first pure spinning and interweaves again, and textile is made.In order to deeply (to every fiber), comprehensively (to a large amount of
Single fiber is counted with the yarn and fabric property of accurate evaluation entirety) above-mentioned textile is dissected, carry out reverse-engineering point
Analysis improves working efficiency, currently, this respect reverse engineering analysis method still belongs to blank, the present embodiment so as to shorten the time of drawing a design
Above-mentioned technological gap is filled up.
Referring to Fig. 5, this, which is illustrated as the embodiment of the present invention six, provides a kind of reverse engineering analysis method of textile
Method flow diagram, suitable for the design, production, detection of textile, the explanation of term identical or corresponding with the various embodiments described above
Details are not described herein, which includes:
Any fiber mixes detection method of content and detects to inspection textile in through the foregoing embodiment, is sent
Examine the mixing content of fiber of all categories in textile；
The corresponding design of mixing content configuration, the production decision obtained according to detection, produces textile.
Specifically, as shown in figure 5, the step of reverse engineering analysis method it is as follows:
S501: inspection textile is disassembled into fiber dust；
Wherein, all fibres in fiber dust are single fiber state；
S502: carrying out digital image acquisition to all fibres in fiber dust, until obtaining that inspection weaving can be represented
Several fibre images of product overall performance；
S503: the instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and same
When calculate the average diameter and variance for obtaining fiber of all categories；
Wherein, the classification of fiber includes material, color, shape, technique；
S504: marking and combine computer deep learning and image, semantic parted pattern according to preliminary, to fibre image into
All fibres of all categories in fibre image are respectively divided and are identified, obtain semantic segmentation image by row semantic segmentation；
S505: in conjunction with the average diameter and variance of fiber of all categories, to all fibres of all categories in semantic segmentation image
It is analyzed and is calculated, obtain the mixing content of fiber of all categories in inspection textile
S506: the corresponding design of mixing content configuration, the production decision obtained according to detection produces textile.
A kind of reverse engineering analysis method of textile is present embodiments provided, which passes through above-mentioned detection method
Textile is detected, further according to the corresponding design of mixing content configuration, production decision that detection obtains, is advantageously implemented and knits
Fast proofing, production and the processing of article.
Preferably, step S506 " spin by the corresponding design of mixing content configuration, the production decision obtained according to detection, production
Fabric " specifically includes:
S5061: several different classes of fibers are chosen according to the mixing content that detection obtains；
S5062: several different classes of fibers are subjected to weight proportion, mixing, obtain mixed raw material；
S5063: textile is produced using mixed raw material.
As it can be seen that understanding inspection textile using the mixing content that the reverse engineering analysis method can be obtained quickly through detection
Ingredient constitute, so that fabric manufacturer be helped to configure corresponding production decision, be conducive to fast proofing, the batch production of textile
And processing.
Embodiment seven
It is set referring to Fig. 6, a kind of fiber of the textile for being illustrated as the offer of the embodiment of the present invention seven mixes content detection
Standby structural schematic diagram mixes detection method of content, the detection for realizing fiber such as provided by any embodiment of the invention
Equipment includes textile disassembling apparatus 60, fibre image acquisition device 61, processing terminal 62 and for receiving operator command
Computer input terminal 63；
Specifically, processing terminal can be made of some relevant apparatus such as computer or cloud server, the fibre image
Acquisition device 61 can be optical micro imaging device；It is some that the computer input terminal 63 can be keyboard, mouse or touch screen etc.
For carrying out human-computer interaction device with computer (processing terminal), the memory of computer is stored with for marking type of fibers
Software or readable program；
The fibre image acquisition device 61, for carrying out digital image acquisition to all fibres in fiber dust, directly
To several fibre images for obtaining to represent inspection textile overall performance；
The processing terminal 62 includes:
First processing units 621, for the instruction of the response computer input terminal 63, to of all categories in fibre image
Fiber is tentatively marked, and calculates the average diameter and variance for obtaining fiber of all categories simultaneously；Wherein, the classification of fiber includes
Material, color, shape, technique；
The second processing unit 622, for marking and combining computer deep learning and image, semantic to divide mould according to preliminary
Type carries out semantic segmentation to fibre image, all fibres of all categories in fibre image is respectively divided and is identified, obtain semanteme
Segmented image；
Third processing unit 623, for combining the average diameter and variance of fiber of all categories, to each in semantic segmentation image
The all fibres of classification are analyzed and are calculated, and the mixing content of fiber of all categories in inspection textile is obtained.
Fibre image acquisition device 61 acquire fiber dust digital picture, processing terminal 62 perform various functions using with
And data processing, such as realize that fiber provided by the embodiment of the present invention mixes detection method of content.
The above, the above embodiments are merely illustrative of the technical solutions of the present invention, rather than its limitations；Although referring to before
Stating embodiment, invention is explained in detail, those skilled in the art should understand that: it still can be to preceding
Technical solution documented by each embodiment is stated to modify or equivalent replacement of some of the technical features；And these
It modifies or replaces, the spirit and scope for technical solution of various embodiments of the present invention that it does not separate the essence of the corresponding technical solution.
Claims (8)
1. a kind of fiber of textile mixes detection method of content characterized by comprising
Inspection textile is disassembled into fiber dust；Wherein, all fibres in fiber dust are single fiber state；
Digital image acquisition is carried out to all fibres in fiber dust, until obtaining that inspection textile overall performance can be represented
Several fibre images；
The instruction of response computer input terminal tentatively marks the fiber of all categories in fibre image, and calculates obtain simultaneously
Obtain the average diameter and variance of fiber of all categories；Wherein, the classification of fiber includes material, color, shape, technique；
Computer deep learning and image, semantic parted pattern are marked and combined according to preliminary, semantic point is carried out to fibre image
It cuts, all fibres of all categories in fibre image is respectively divided and is identified, semantic segmentation image is obtained；
In conjunction with the average diameter and variance of fiber of all categories, all fibres of all categories in semantic segmentation image are analyzed simultaneously
It calculates, obtains the mixing content of fiber of all categories in inspection textile.
2. fiber according to claim 1 mixes detection method of content, which is characterized in that the combination fiber of all categories
Average diameter and variance are analyzed and are calculated to all fibres of all categories in semantic segmentation image, obtain inspection textile
In fiber of all categories mixing content, specifically include:
Pixel shared by the longitudinal cross-section to the fiber of all categories in semantic segmentation image counts respectively, obtains of all categories
The sum of longitudinal cross-section area of fiber；
According to the following formula, fiber of all categories in inspection textile is calculated in conjunction with the average diameter and variance of fiber of all categories
Weight percent obtains the mixing content of fiber of all categories in inspection textile:
Wherein, ρ1、ρ2...ρmThe density of fiber respectively of all categories；n1、n2...nmThe longitudinal cross-section of fiber respectively of all categories
The sum of area；ξd1、ξd2...ξdmThe stochastic variable of the diameter of fiber respectively of all categories, E are mathematic expectaion, and D is variance.
3. fiber according to claim 1 mixes detection method of content, which is characterized in that described to disassemble inspection textile
At fiber dust, specifically include:
Inspection textile is disassembled into single fiber state；
All fibres are carried out to cut sample preparation, obtain fiber dust；
Wherein, each fibre length of fiber dust is equal.
4. fiber according to claim 1 mixes detection method of content, which is characterized in that if inspection textile includes colourless
Fiber, all fibres in fiber dust carry out digital image acquisition, whole until obtaining that inspection textile can be represented
Before several fibre images of body performance, further includes:
Fibre staining differentiating solvent is instilled toward fiber dust, the fiber of all categories in fiber dust is dyed.
5. fiber according to claim 4 mixes detection method of content, which is characterized in that the past fiber dust instills fine
Dimension coloring differentiating solvent, dyes the fiber of all categories in fiber dust, specifically includes:
The first coloring solution that fabric sample can be completely covered is instilled toward fiber dust；
The second coloring solution is instilled toward fiber dust；
Wherein, first coloring solution is 6.25*10-3The halide or polyhalide and 2.5*10 of~1.44g/ml-3mg/ml
Glycerol mixed aqueous solutions, second coloring solution are the halide of 0.200lg/ml or the aqueous solution of polyhalide.
6. a kind of reverse engineering analysis method of textile, design, production, detection suitable for textile, which is characterized in that packet
It includes:
Detection method of content is mixed by fiber such as described in any one of claim 1 to 5 to detect inspection textile, is obtained
The mixing content of fiber of all categories into inspection textile；
The corresponding design of mixing content configuration, the production decision obtained according to detection, produces textile.
7. reverse engineering analysis method according to claim 6, which is characterized in that the mixing obtained according to detection contains
The corresponding design of amount configuration, production decision, produce textile, specifically include:
Several different classes of fibers are chosen according to the mixing content that detection obtains；
Several different classes of fibers are subjected to weight proportion, mixing, obtain mixed raw material；
Textile is produced using mixed raw material.
8. a kind of fiber of textile mixes content detection equipment, for realizing fiber such as described in any one of claim 1 to 5
Mix detection method of content, which is characterized in that including textile disassembling apparatus, fibre image acquisition device, processing terminal and
For receiving the computer input terminal of operator command；
The textile disassembling apparatus, for disassembling inspection textile at fiber dust；Wherein, all fibres in fiber dust
Dimension is single fiber state；
The fibre image acquisition device, for carrying out digital image acquisition to all fibres in fiber dust, until obtaining
Several fibre images of inspection textile overall performance can be represented；
The processing terminal includes:
First processing units carry out the fiber of all categories in fibre image preliminary for the instruction of response computer input terminal
Mark, and the average diameter and variance for obtaining fiber of all categories are calculated simultaneously；Wherein, the classification of fiber includes material, color, shape
Shape, technique；
The second processing unit for the preliminary mark of basis and combines computer deep learning and image, semantic parted pattern, to fibre
It ties up image and carries out semantic segmentation, all fibres of all categories in fibre image are respectively divided and are identified, semantic segmentation figure is obtained
Picture；
Third processing unit, for combining the average diameter and variance of fiber of all categories, to of all categories in semantic segmentation image
All fibres are analyzed and are calculated, and the mixing content of fiber of all categories in inspection textile is obtained.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910412007.7A CN110047074A (en) | 2019-05-17 | 2019-05-17 | The fiber of textile mixes content detection, reverse engineering analysis method and equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910412007.7A CN110047074A (en) | 2019-05-17 | 2019-05-17 | The fiber of textile mixes content detection, reverse engineering analysis method and equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110047074A true CN110047074A (en) | 2019-07-23 |
Family
ID=67282422
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910412007.7A Pending CN110047074A (en) | 2019-05-17 | 2019-05-17 | The fiber of textile mixes content detection, reverse engineering analysis method and equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110047074A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112164134A (en) * | 2020-09-28 | 2021-01-01 | 华南理工大学 | Random curve modeling method based on image processing |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070248246A1 (en) * | 2002-06-24 | 2007-10-25 | Arkady Cherkassky | Electro-optical method and apparatus for evaluating protrusions of fibers from a fabric surface |
| CN101907574A (en) * | 2009-06-08 | 2010-12-08 | 贾立锋 | Method, color reagent and system for identifying type of fibers |
| CN103063584A (en) * | 2012-12-31 | 2013-04-24 | 广东工业大学 | Method for identifying fiber category |
| CN106373119A (en) * | 2016-09-05 | 2017-02-01 | 广东工业大学 | Fiber detection method and system |
| CN108038838A (en) * | 2017-11-06 | 2018-05-15 | 武汉纺织大学 | A kind of cotton fibriia species automatic testing method and system |
| CN108875777A (en) * | 2018-05-03 | 2018-11-23 | 浙江大学 | Kinds of fibers and blending rate recognition methods in textile fabric based on two-way neural network |
| CN109492544A (en) * | 2018-10-19 | 2019-03-19 | 兰波(苏州)智能科技有限公司 | A method of classified by enhancing optical microscopy to animal origin |
| CN109633132A (en) * | 2018-11-16 | 2019-04-16 | 绍兴文理学院元培学院 | Blended yarn component ratio testing method |
-
2019
- 2019-05-17 CN CN201910412007.7A patent/CN110047074A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070248246A1 (en) * | 2002-06-24 | 2007-10-25 | Arkady Cherkassky | Electro-optical method and apparatus for evaluating protrusions of fibers from a fabric surface |
| CN101907574A (en) * | 2009-06-08 | 2010-12-08 | 贾立锋 | Method, color reagent and system for identifying type of fibers |
| CN103063584A (en) * | 2012-12-31 | 2013-04-24 | 广东工业大学 | Method for identifying fiber category |
| CN106373119A (en) * | 2016-09-05 | 2017-02-01 | 广东工业大学 | Fiber detection method and system |
| CN108038838A (en) * | 2017-11-06 | 2018-05-15 | 武汉纺织大学 | A kind of cotton fibriia species automatic testing method and system |
| CN108875777A (en) * | 2018-05-03 | 2018-11-23 | 浙江大学 | Kinds of fibers and blending rate recognition methods in textile fabric based on two-way neural network |
| CN109492544A (en) * | 2018-10-19 | 2019-03-19 | 兰波(苏州)智能科技有限公司 | A method of classified by enhancing optical microscopy to animal origin |
| CN109633132A (en) * | 2018-11-16 | 2019-04-16 | 绍兴文理学院元培学院 | Blended yarn component ratio testing method |
Non-Patent Citations (2)
| Title |
|---|
| 应乐斌: "混纺织物不同纤维的图像识别与定量分析", 《中国优秀硕士论文全文数据库信息科技辑》 * |
| 贾立锋: "应用支持向量机进行棉麻纤维自动识别的研究", 《现代仪器》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112164134A (en) * | 2020-09-28 | 2021-01-01 | 华南理工大学 | Random curve modeling method based on image processing |
| CN112164134B (en) * | 2020-09-28 | 2024-03-22 | 华南理工大学 | Random curve modeling method based on image processing |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103185776B (en) | Blood cell analyzer, blood cell analyzing method, and information processing device | |
| CN209300267U (en) | A kind of cigarette appearance images processing control system and its on-line measuring device | |
| CN106525666A (en) | Blood analyzer, blood analyzing method, and information processing apparatus | |
| CN103348250B (en) | Sample analysis system, device for analyzing samples and method of sample analysis | |
| CN102662048B (en) | Method and device for analyzing urine | |
| CN105628563B (en) | Detergent is to fabric decontaminating power method for rapidly testing and its test device | |
| CN110047074A (en) | The fiber of textile mixes content detection, reverse engineering analysis method and equipment | |
| CN205581122U (en) | Eight passageway automatic coagulation analyzer | |
| CN103499301A (en) | Bast-fiber fineness automatic analyzer and application thereof | |
| CN108563569A (en) | A kind of warnlng surveillance system automation interface test method | |
| CN110349123A (en) | A kind of quantitative evaluation method of apparent mass of clear concrete | |
| Wang et al. | Objective evaluation on yarn hairiness detection based on multi-view imaging and processing method | |
| Li et al. | Analysis of conical slump shape reconstructed from stereovision images for yield stress prediction | |
| CN205580989U (en) | Six channel automatic coagulation analyzer | |
| CN108875777B (en) | Method for identifying fiber types and blending proportion in textile fabric based on double-path neural network | |
| CN102735689A (en) | Paper defect detection system based on image pattern recognition | |
| CN102435153A (en) | Profile degree testing method for polyester filament yarn | |
| CN107271345A (en) | A kind of two method combined calibrating oxygen transit dose standard films and preparation method thereof | |
| CN106442528A (en) | Dual camera microscope testing device and testing method for contents of fiber components | |
| CN102419329A (en) | Cotton fiber embedding inspection method | |
| CN109297854A (en) | A kind of stem run in item silk mixes the rapid assay methods of ratio in fact | |
| CN204854647U (en) | System for fibre winding pitch in compound pipe of on line measurement | |
| CN105806218B (en) | Tube sheet pore measuring method based on photogrammetric technology | |
| CN206057196U (en) | The dual camera microscope test device of fibrous fraction content | |
| CN108519382A (en) | The evaluation method of textile fabric recognition result |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190723 |