CN202433309U - Real-time yarn appearance digitized analytic system - Google Patents

Abstract

The utility model relates to a real-time yarn appearance digitized analytic system, which includes a camera obscura, an image acquisition system and a PC (personal computer) processing terminal, wherein the image acquisition system is electrically connected with the PC processing terminal, and at least two reflecting mirrors and one LED (light-emitting diode) backlight source are arranged in the camera obscura. The real-time yarn appearance digitized analytic system can dynamically and continuously acquire yarn images; during the feature extraction process, the related analysis and processing are performed for the same detection position on five side yarn images in each frame, the features are extracted; and finally, the yarn appearance feature information detected from all images is integrated to form feature vectors of the yarn appearance, the feature similarity DW (X, Y) between the feature vectors and a standard sample is calculated, and on the basis, the property and quality of the yarn are judged. The multiple-angle yarn three-dimensional appearance feature information can be obtained in real time, the quality of the yarn can be determined quickly and accurately, the influences on the quality evaluation by personal factors can be reduced efficiently, the analytic accuracy is improved, and the real-time yarn appearance digitized analytic system is reasonably designed and highly practical.

Description

A kind of real-time yarn appearance digital assay system

Technical field

The utility model belongs to the intelligent digitalized information analysis of machine vision field, particularly relates to a kind of real-time yarn appearance digital assay system.

Background technology

The yarn qualities direct relation the appearance property of fabric and clothes, physical property with take performance, in weaving is produced, play a part very important.The fineness of yarn, filoplume degree and bar evenness directly influence the outward appearance of lining, and also influence is powerful, stretchability, elasticity, physical and mechanical properties such as wearing quality.The appearance information of yarn (thickness variation, ovality, the uneven rate of diameter etc.) has extremely important meaning to thoroughly evaluating yarn qualities, guidance system fabrication technique, prediction appearance of fabrics effect.

Therefore, just become the key problem of a product quality monitoring from the textile industry large-scale production for the detection of yarn qualities; Detect round quality of yarn, home and abroad weaving colleague has developed a lot of yarn qualities detecting instruments.At present, the detection of the presentation quality of yarn is mainly accomplished by yarn evenness appearance and filoplume appearance.The principle of yarn evenness and method are divided into two kinds: measurement by capacitance and photoelectric type measuring.The ultimate principle of measurement by capacitance is to let yarn pass through between the plane-parallel capacitor pole plate with certain speed; Cause the media variations of electric capacity owing to quality (line density) variation of yarn unit length; And then causing the variation of electric capacity, the variation of electric capacity promptly reflects the irregular degree of yarn evenness; Photoelectric type measuring then lets yarn passing through in the photodetector system of certain speed from the two-way orthogonal directions, and photoelectric sensor is converted into electric signal with the yarn diameter signal, and the variation of electric signal promptly reflects the evenness fault degree of yarn.Yet, traditional condenser type or photo-electric yarn evenness tester, electric capacity signal or the photosignals etc. that can only obtain reflecting thickness of yarn are amount indirectly, and appearance or the color and veins etc. that can not directly obtain yarn are information more intuitively.In some cases, more can't realize sign accurately, realistically to yarn appearance.Simultaneously, this method is affected by environment bigger, the humiture difference of same yarn environment of living in will cause testing result than big-difference.In addition, utilize the correlativity of yarn qualities irregularity that the dried appearance of condenser type bar records and fabric face quality not strong, be difficult to predict reliably fabric quality.

Along with the development of PC processing terminal technology and image analysis technology, the researchist begins to use optics and digital picture sensing technology, the accurate yarn appearance that detects.Adopt PC processing terminal image digitization treatment technology can accurately measure appearance informations such as yarn diameter, ovality and irregularity, torsion resistance, the appearance ratings of yarn is assessed with realization.PC processing terminal visual analysis system is meant optics and PC processing terminal image processing techniques is combined; Utilize the knowledge of applied mathematics such as wavelet analysis, Fourier analysis; The figure warp thread that imageing sensor is collected in real time looks like to carry out Flame Image Process, obtains that fineness of yarn is irregular intuitively.

At present the research in this field mainly is placed on the test of geometric shape and characterizes, the envelop of function of test still is confined to traditional bar evenness and tests with filoplume, measurement pattern with single direction obtain the single-frame images analysis, be measured as the master.Yet because the 3-D solid structure of yarn object, single angle field of view imaging detects and has information dropout, is prone to cause the error of extracting characteristic.The utility model proposes PC processing terminal image processing techniques; Designed four direct reflection imaging devices, realized that the same visual field obtains the yarn appearance image of five different angles observations, and pass through image processing techniques; Extract different angles observation yam surface characteristic; And merge, obtain the digitalized signature parameter of representing yarn, for scientific basis being provided to yarn property evaluation and quality analysis.Along with the raising of PC processing terminal cost reduction with performance, and the development of PC processing terminal image processing techniques, PC processing terminal technology is in the development and the application of textile industry, and prospect will be more wide.

The utility model content

The purpose of the utility model provides a kind of above-mentioned deficiency that overcomes; Can obtain the yarn 3 D stereo external appearance characteristic information of multi-angle in real time; Confirm quality of yarn fast and accurately; And effectively reduced the influence of human factor to quality assessment, improve the real-time yarn appearance digital assay system of the accuracy rate analyzed.

The technical scheme that realizes the utility model purpose is: a kind of real-time yarn appearance digital assay system; Have camera bellows, image capturing system and PC processing terminal; Image capturing system is electrically connected with the PC processing terminal, is provided with at least two catoptrons and a LED-backlit light source in the camera bellows.

The image capturing system of above-mentioned real-time yarn appearance digital assay system has ccd image sensor and camera lens.

The LED-backlit light source of above-mentioned real-time yarn appearance digital assay system is arranged on the rear wall of camera bellows, and image capturing system is arranged on the antetheca of camera bellows and is relative with the LED-backlit light source, and said catoptron has four; The catoptron height is identical with the camera bellows inside wall height; Two one components are listed in LED-backlit light source both sides, be symmetrically distributed, every catoptron all with the sidewall formation angle of camera bellows; Angle remains between 45～75 degree, and minute surface points to the camera bellows middle part.

The utlity model has positive effect: the utility model can obtain the 3 D stereo external appearance characteristic information of the yarn of multi-angle in real time; Confirm quality of yarn fast and accurately; And effectively reduced the influence of human factor to quality assessment, and improve the accuracy rate of analyzing, reasonable in design; Practicality is high, in textile industry, detect industry automatically wide application is arranged.

Description of drawings

For the content that makes the utility model is expressly understood more easily, according to specific embodiment also in conjunction with the accompanying drawings, the utility model is described in further detail, wherein below

Fig. 1 is the utility model structural representation;

Embodiment

See Fig. 1, the real-time yarn appearance digital assay system of the utility model has camera bellows 1, image capturing system 2 and PC processing terminal, and image capturing system 2 is electrically connected with the PC processing terminal; Image capturing system 2 has ccd image sensor 21 and camera lens 22, and 1394 image pick-up cards of use are provided with four catoptrons 3 and a LED-backlit light source 4 in the camera bellows 1; LED-backlit light source 4 is arranged on rear wall 11 middle parts of camera bellows 1, and image capturing system 2 is arranged on the antetheca 12 of camera bellows 1 and is relative with LED-backlit light source 4, and catoptron 3 height are identical with camera bellows 1 inside wall height; Two one components are listed in LED-backlit light source 4 both sides; Be symmetrically distributed, every catoptron 3 all with the sidewall 13 formation angles of camera bellows 1, angle remain on 45～75 spend between; Minute surface points to camera bellows 1 middle part; Yarn 9 vertically passes camera bellows 1, and between image capturing system 2 and LED-backlit light source 4, four catoptrons 3 can reflex to image capturing system 2 with the image of yarn 9 sides.

Key step is:

A, yarn 9 pass in camera bellows 1 casing of sealing, after good LED-backlit light source 4 brightness of regulating and controlling, image capturing system 2 exposures, drive yarn 9 by certain speed and pass camera bellows 1;

B, ccd image sensor 21 obtain the three-dimensional different angles image of yarn 9 in real time; And be transferred to the image processing software on the PC processing terminal; The PC processing terminal utilizes image processing techniques that yarn 9 images are carried out relevant treatment according to preestablishing yarn 9 analysis length;

C, PC processing terminal carry out filtering to stitching image and remove noise, and the yarn target is extracted in processing such as difference shadow subtracting background;

D, parameter extractions such as color, thickness variation, ovality and cashmere density are arranged respectively to comprising each side figure warp thread picture in the N two field picture;

E, the characteristic parameter that extracts is carried out weight data fusion constitutive characteristic vector, representing the numerical characteristic of yam surface, and store in the database.

F, according to the characteristic parameter of given master pattern with detect the vector parameters that sample merges, (X Y) calculates, and carries out yarn qualities made as the criterion of yarn property and analyze and judge to carry out characteristic similarity DW with weighting Euler distance.

Consider that casing yarn 9 imaging is clear, employing be built-in rectangle strip back light method, and place the rear wall 11 of camera bellows 1, the bright dark degree that can regulate LED-backlit light source 4 through controller outside is conveniently to obtain figure warp thread picture the most clearly.Because except the actual side of yarn 9, also have the image of yarn 9 sides in four mirror images in the image that obtains, the yarn in the image that seek out in the mirror image is also high-visible, this to the requirement of light than higher.In imaging process, at first regulate the intensity of LED-backlit light source 4, utilize the time shutter of image processing software automatic fine tuning joint image capturing system 2 again, obtain figure warp thread picture the most clearly to reach.

After bright, the dark degree of camera bellows and aperture size mix up, start coiling apparatus, send order by the upper computer software control system through serial ports, the rotation of control step motor, thus make yarn maintain certain speed.The confirming of this speed confirms that by the picture number that the visual field of image capturing system 2 and the picture number that will handle p.s. and p.s. take concrete formula is following:

V＝L*T

$u_i=\frac{1}{3}(u_{\mathrm{sai}}+u_{\mathrm{sbi}}+u_{\mathrm{sci}}+u_{\mathrm{sdi}}+u_{\mathrm{sei}})$

$R_i=\frac{1}{3}(R_{\mathrm{sai}}+R_{\mathrm{sbi}}+R_{\mathrm{sci}}+R_{\mathrm{sdi}}+R_{\mathrm{sei}})$

$G_i=\frac{1}{3}(G_{\mathrm{sai}}+G_{\mathrm{sbi}}+G_{\mathrm{sci}}+G_{\mathrm{sdi}}+G_{\mathrm{sei}})$

$B_i=\frac{1}{3}(B_{\mathrm{sai}}+B_{\mathrm{sbi}}+B_{\mathrm{sci}}+B_{\mathrm{sdi}}+B_{\mathrm{sei}})$

Wherein V is the speed that yarn transmits, and L is the developed width that the visual field comprised of CCD camera lens, and T is the amount of images that will take p.s..

After the spiral speed that configures yarn 9; Begin to move yarn 9; And, obtain image and be sent in 1394 image pick-up cards of PC processing terminal, and pass in the software by image pick-up card and to handle by 1394 live wire serial ports by software control image capturing system 2 images acquired of PC processing terminal.

Because image capturing system 2 is when images acquired; Reason such as imaging circumstances and internal circuit all can make the figure warp thread of picked-up look like to exist the noise of stochastic distribution; It is Gaussian noise; In order to utilize each width of cloth figure warp thread picture, then must carry out pre-service to image, this step is to obtain the prerequisite of high-definition image.Because frequency domain filtering is to remove noise through the high frequency composition that filters image, it also can lose image when removing noise detailed information makes image blur, and in addition, the frequency domain transform operand is bigger, inapplicable and processing speed requirement faster.Consider the validity of image filtering and the real-time of Flame Image Process, the utility model adopts the average and the median filtering algorithm of airspace filter that image is carried out pre-service, removes the noise of image.

Mean filter is the most frequently used a kind of linear filter method; Because the deficiency of linear filtering is when reducing noise, also to have blured entire image; Particularly edge of image and details; Therefore directly use mean filter to be unfavorable for the reservation of high-frequency signal, the utility model adopts a kind of algorithm of improved mean filter that view data is handled.

In general mean filter be successively to the operation neighborhood of pixel points average and make image blurring; The value of reservation original pixel point with good conditionsi if we do not get average; Promptly have only when the territory average is greater than a certain threshold value near the value of pixel; Just get the value of neighborhood average, otherwise the value of this point remains unchanged for this point.Image just can not blur so, can remove noise again simultaneously.

Same, because medium filtering is got in its filter window the value through ordering back intermediate pixel, and do not have positive connection with the value of original pixel point.So if the object size is less in the image, this object does not have pixel to come the centre after the ordering, and this object will be corroded even disappear after the filtering so.So establish a threshold value during our underway value filtering operation earlier, the intermediate value of just getting neighborhood during greater than threshold value when the difference of the value of having only this point and neighborhood intermediate value, otherwise just keep the value of this point.Like this, if noise spot, the intermediate value of it and neighborhood will differ bigger, just can remove it.If the image information object, also have the point of object itself in the neighborhood, so the value of the intermediate value of neighborhood and this point just can not differ too big,, just can keep the value of this point as long as appropriate threshold is set.

The described image processing module of the utility model mainly comprises the processing that realizes following two aspects.The one, difference shadow method is removed redundant background information, and the 2nd, adopt special algorithm to extract the three-dimensional feature information of yarn.Difference shadow method is meant under identical environment and background, removes the image image as a setting behind the objective body, with the direct subtracting background image of gathering in real time of target image, to obtain the image that only contains the yarn target, reaches the convenient effect of handling.In the utility model, concrete method of operating is withdrawn from yarn for after when yarn is arranged, regulating the extremely normal IMAQ state of bright dark degree in the camera bellows, gathers background image with CCD, as the image that will deduct, is stored in the PC processing terminal.Carry out normal yarn Image Acquisition afterwards, whenever obtain piece image, be with this figure image subtraction in the first step of handling and be stored in the background image in the PC processing terminal originally, only obtain, again this image is handled with the image of yarn information.

The utility model adopts the image to same one thread different angles, analyzes the method for characteristic information extraction, and it is as shown in Figure 2 that it extracts characteristic.Adopt the method for Feature Fusion can react the actual information situation of yarn more really, increased result's accuracy and authenticity, satisfy under the specific environment detection requirement yarn.Before this to the same detection position of each two field picture, it carries out correlation analysis and processing simultaneously to the figure warp thread picture of five sides in the image respectively, extracts characteristic.To detect all images yarn appearance characteristic information at last and merge the proper vector that constitutes yarn appearance.

The extraction of color is to adopt clustering method to obtain R, G, the B color component of detection position, and represents the color of yarn with the average of whole yarn 9 detection position color components. $(R,G,B)=(\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{R}_i,\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{G}_i,\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{B}_i),$ The number of N check point wherein.Yarn diameter is to calculate the diameter of five side yarn averages as this detection position of yarn earlier, the diameter computing formula of k detection position: ${\mathrm{\μ}}_k=\frac{1}{5}\underset{i=1}{\overset{5}{\mathrm{\Σ}}}\sqrt{{(x_{\mathrm{Ki}1}-x_{\mathrm{Ki}2})}^2+{(y_{\mathrm{Ki}1}-y_{\mathrm{Ki}2})}^2},$ Again yarn 9 all detection position points are made even all as the diameter that detects yarn.

The diameter that the circularity of yarn 9 adopts five outboard profiles to obtain is asked for.The computing method of k detection position circularity are: μ in the formula _kBe the mean diameter of five side image k detection positions, δ _kThe mean square deviation of five side image k detection position diameters, ${\mathrm{\δ}}_k=\frac{1}{5}\underset{i=1}{\overset{5}{\mathrm{\Σ}}}\left[\right||(x_{1\mathrm{Ki}},y_{1\mathrm{Ki}})-(x_{2\mathrm{Ki}},y_{2\mathrm{Ki}})||-{\mathrm{\μ}}_k];$ Yarn evenness is:

Wherein, N is the number of yarn check point.

The processing of yarn 9 lint density is to adopt the image scanning detection mode.With the yarn check point is starting point, to left and right certain area scanning, digit's area lint number.

The evaluation of performance and quality adopts the method for the sample parameter analogy of parameter vector and standard to handle.(X Y) carries out as criterion, wherein to adopt weighting Euler distance to carry out characteristic similarity DW

X=(x ₁, x ₂..., x _n) be the proper vector that detects yarn, and Y=(y ₁, y ₂..., y _n) be the proper vector of master sample, w _iBe the weight of characteristic, the proportion of different components in the representation feature vector.

In implementation process, adopt following equipment:

(1) ccd image sensor 21 models are Basler A602fc-2, and major parameter comprises that resolution is 656*491; 1,300,000 pixels; Maximum frame number 100FPS; 8bit color figure place, manual focusing; The IEEE1394 interface; Can accept two kinds of power supply supplies of 24V AC and 12V DC.

(2) camera lens 22 models are Computar H0514-MP, 1/2 " specification; C interface; 5 (mm) focal length; Aperture (F): 1.4-16C; Visual angle (level): 65.5 °; Object image distance leaves (m) recently: 0.1; Before the effective aperture (φ mm): 27.8, back (φ mm) 14.8; The pre-filter screw thread (φ M * P=): 43.0 * 0.75; Physical dimension (diameter * dark mm): 44.5 * 45.5.

(3) LED-backlit light source 4 models are the HFL-27-27-W white light source; Accept the 24VDC power supply; Power 2.4W; Light-emitting area 30*38; Standard operation environment: temperature: 0-40 ℃, humidity: 20-85% (non-condensing) power supply driver model: YMAPS-24W6-1T; Be of a size of: 12.8cm*7cm*5.5cm

(4) image pick-up card is the T2000 capture card; Fire-wire interfaces.

(5) the controller model is UIM241040171; Power supply voltage 12V～40VDC; Output motor electric current: the every phase of peak value 4A/8A (actual current is set by user instruction, Adjustable real-time); Type of drive: perseverance flows PWM mutually; Control excitation mode: synchronizing, half step, 4 segmentations, 16 segmentations; Physical dimension: 42.3mm * 42.3mm * 13.5mm; Weight: 0.1kg.

(6) stepper motor model FL57STH76-2804-01; Fuselage is long: 7.6cm+0.2cm; Motor axial length: 2.0cm; Fuselage size: 5.6*5.6cm; Rated voltage 2.3v; Rated current 2.8A; Resistance 0.83 Europe.

(7) PC processing terminal system: processor: Pentium (R Dual-Core) CPU E5700; Dominant frequency 3.00GH; Internal memory 2.0GB; Operating system: Microsoft Windows XP SP3.

As shown in Figure 3, at first power on, start-up system is carried out initialization.Initialized process comprises sets the length that detects yarn 9, the frame per second of CCD, and the rotating speed of stepper motor, and the micron chi of the standard of employing is calibrated the imaging pattern of regulating exposure and setting to system.In the pick-up unit with the system of being installed on, regulate the exposure of imaging system once more after, by PC processing terminal software control coiling apparatus, drive the imaging region that yarn passes cabinet space by certain speed.The CCD imaging device obtains the three-dimensional different angles image of yarn in real time, and is transferred to the image processing software on the PC processing terminal, and the PC processing terminal utilizes image processing techniques that figure warp thread is looked like to carry out relevant treatment according to preestablishing yarn analysis length.The PC processing terminal carries out filtering to stitching image and removes noise, and the yarn target is extracted in processing such as difference shadow subtracting background.And the figure warp thread picture that ipsilateral not obtains there are parameter extractions such as color, thickness variation, ovality and cashmere density respectively; The characteristic that all images is proposed adopts the data fusion technology; Obtain the numerical characteristic parameter vector of representing yarn 9 surfaces, and store in the database.According to given model, (X Y) calculates, and carries out yarn qualities is made analysis and judge as the criterion of yarn property to carry out characteristic similarity DW with weighting Euler distance.

Above-described specific embodiment; Purpose, technical scheme and beneficial effect to the utility model have carried out further explain, it should be understood that the above is merely the specific embodiment of the utility model; Be not limited to the utility model; All within the spirit and principle of the utility model, any modification of being made, be equal to replacement, improvement etc., all should be included within the protection domain of the utility model.

Claims (3)

1. real-time yarn appearance digital assay system; Have camera bellows (1), image capturing system (2) and PC processing terminal; Image capturing system (2) is electrically connected with the PC processing terminal, is provided with at least two catoptrons (3) and a LED-backlit light source (4) in the camera bellows (1).

2. real-time yarn appearance digital assay according to claim 1 system, it is characterized in that: said image capturing system (2) has ccd image sensor (21) and camera lens (22).

3. real-time yarn appearance digital assay according to claim 2 system; It is characterized in that: said LED-backlit light source (4) is arranged on the rear wall (11) of camera bellows (1), and image capturing system (2) is arranged on last and relative with LED-backlit light source (4) at the antetheca (12) of camera bellows (1), and said catoptron (3) has four; Catoptron (3) height is identical with camera bellows (1) inside wall height; Two one components are listed in LED-backlit light source (4) both sides, be symmetrically distributed, every catoptron (3) all with sidewall (13) the formation angle of camera bellows (1); Angle remains between 45～75 degree, and minute surface points to camera bellows (1) middle part.

Images