CN111627053A

CN111627053A - Method and system for detecting cleaning cleanliness of universal spinneret orifice

Info

Publication number: CN111627053A
Application number: CN202010412672.9A
Authority: CN
Inventors: 徐增波
Original assignee: Shanghai University of Engineering Science
Current assignee: Shanghai University of Engineering Science
Priority date: 2020-05-15
Filing date: 2020-05-15
Publication date: 2020-09-04
Anticipated expiration: 2040-05-15
Also published as: CN111627053B

Abstract

The invention provides a method and a system for detecting the cleaning cleanliness of a universal spinneret orifice, comprising the following steps: extracting an edge contour closed curve of a standard clean hole high gradient threshold value and a curvature curve thereof, and constructing a parameterized segmented closed curve standard detection template containing control points; in the dirt detection stage, the parameterized segmented closed curve standard detection template is mapped to the outline to be detected through non-rigid registration; constructing a dirt detection curve, wherein the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template, a distance curve of the contour to be detected based on nearest neighbor search and a mixed curvature curve; and carrying out floor reference correction on the dirt detection curve to form a global threshold so as to segment and position dirt and quantify a cleanliness index.

Description

Method and system for detecting cleaning cleanliness of universal spinneret orifice

Technical Field

The invention relates to the technical field of image processing, in particular to a method and a system for detecting the cleaning cleanliness of a universal spinneret orifice.

Background

In the monitoring of chemical fiber production quality, whether spinneret holes are cleaned or not is an important factor influencing the quality of finished filaments. At present, most of the computer vision detection methods are adopted to judge whether dirt remains in holes or not by imaging through a rear projection light source and analyzing and extracting characteristic indexes such as area, perimeter and the like of hole outline areas. The method has obvious effect only on the conventional dirt, has high false judgment rate on the fine dirt, and basically fails in the judgment of indexes such as area, perimeter and the like particularly when the deformation degree of the hole is large.

Specifically, the chemical fiber is a silk thread formed by spraying molten polyester macromolecules through special-shaped spinneret orifices on a spinneret plate under the conditions of high temperature and high pressure, and then condensing and crystallizing, the shape and the size of the spinneret orifices determine the performance of the chemical fiber, and the spinneret plate can be reused after being ultrasonically cleaned and dried after a certain service cycle. The spinneret orifice quality monitoring is an important process for ensuring the spinneret quality, wherein the dirt detection is the most important link, and the abnormal spinning quality parameters can be caused after the non-clean spinneret plate is on line. In actual production, detection personnel can carry out dirt analysis, hole positioning, automatic hole blowing and manual removal through full-automatic computer vision detection equipment after a spinneret plate is cleaned and dried. In recent years, computer vision technology is widely applied in the field of textile industry, but research related to micropore defect detection is less, and the method mainly focuses on general visual detection application levels such as accurate focusing, size detection, control positioning and the like. In practical application, in a general cleanliness detection method, under the illumination of a vertical rear projection light source, a hole image is collected through image collection equipment, if dirt exists around a hole outline, the hole permeability is reduced, and the edge outline of a hole permeable area changes, so that whether the dirt remains in the hole can be judged by analyzing and extracting characteristic indexes such as the area, the perimeter and the like of the hole outline area. However, the conventional area, circumference and other index judgment method has a remarkable effect only on detecting dirt with remarkable visual difference, and has a high false judgment rate on most of fine dirt, because the fluctuation of the area and circumference indexes caused by the dispersion of the processing precision of the holes per se exceeds the fine dirt detection sensitivity range.

Disclosure of Invention

The invention aims to provide a method and a system for detecting the cleaning cleanliness of a general spinneret orifice, and aims to solve the problem that the misjudgment rate of fine dirt of the existing spinneret orifice is high.

In order to solve the technical problems, the invention provides a method and a system for detecting the cleaning cleanliness of a universal spinneret orifice, wherein the method for detecting the cleaning cleanliness of the universal spinneret orifice comprises the following steps:

extracting an edge contour closed curve of a standard clean hole high gradient threshold value and a curvature curve thereof, and constructing a parameterized segmented closed curve standard detection template containing control points;

in the dirt detection stage, mapping the parameterized segmented closed curve standard detection template to a profile to be detected through non-rigid registration;

constructing a dirt detection curve, wherein the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template, a distance curve of the to-be-detected contour based on nearest neighbor search and a mixed curvature curve;

and carrying out floor reference correction on the dirt detection curve to form a global threshold so as to segment and position dirt and quantify a cleanliness index.

Optionally, in the method for detecting the cleaning cleanliness of the universal spinneret orifice, the method for detecting the cleaning cleanliness of the universal spinneret orifice further includes:

extracting a standard hole outline and a curvature curve thereof through standard clean hole image edge analysis, and carrying out segmentation and fitting on a straight line segment and a circular arc segment of the hole outline based on a dual-threshold segmentation algorithm of the curvature curve;

according to the segmentation and fitting of the straight line segment and the circular arc segment of the standard hole profile, reconstructing a parameterized segmented closed curve standard detection template based on control points;

extracting an edge contour closed curve of a high gradient threshold of a hole to be detected, calculating the length of the edge contour closed curve of the high gradient threshold of the hole to be detected and the area enclosed by the edge contour closed curve, detecting the significant dirt, if the detection result is unqualified, detecting the next hole to be detected, otherwise, detecting the non-significant dirt;

extracting an edge contour closed curve of a hole low gradient threshold to be detected, mapping the edge contour closed curve of the standard clean hole high gradient threshold to the edge contour closed curve of the hole high gradient threshold to be detected through non-rigid registration, constructing a distance curve between the edge contour closed curve of the standard clean hole high gradient threshold and the edge contour curve of the hole low gradient threshold to be detected based on nearest neighbor search, and constructing a dirt detection curve by taking a mixed weighting curve as weight;

through floor reference calibration processing of the dirt detection curve, positioning and quantifying dirt by using global threshold segmentation, constructing a cleanliness index for dirt judgment, and judging whether hole dirt detection is qualified;

and (3) taking the round hole, the flat hole, the three-leaf hole and the cross hole as research objects, and giving result analysis and conclusion.

Optionally, in the method for detecting the cleaning cleanliness of the universal spinneret orifice, extracting an edge contour closed curve of a standard clean orifice high gradient threshold includes:

extracting a hole edge profile curve by adopting a Canny edge operator, respectively detecting strong and weak edges by adopting a dual gradient threshold method, and only keeping the weak edges connected with the strong edges;

setting a first gradient threshold σ_thr1And a second gradient threshold σ_thr2Respectively extracting strong contour edges and strong and weak combination contour edges, wherein the first gradient threshold is smaller than the second gradient threshold, and the first gradient threshold and the second gradient threshold respectively correspond to Canny detection double-gradient threshold [ sigma ]_thr1，3*σ_thr1]And [ sigma ]_thr2，3*σ_thr2]；

The extracted image edges are respectively a first edge and a second edge, wherein the first edge is used for constructing parameterized segmented closed curve standard detection and non-rigid registration and significant fouling detection, and the second edge is used for non-significant fouling detection.

Optionally, in the method for detecting the cleanliness of the washed common spinneret orifice, reconstructing a parameterized segmented closed curve standard detection template based on control points includes:

thinning and edge intersection point searching based on 8 neighborhoods, and breaking the edge intersection points;

searching curve end points and tracking the contour track, counting the length, and if the length is less than a set threshold value, reversely tracking along the end points and erasing the track;

searching the longest curve segment, marking the starting point and the end point position of the longest curve segment, searching the starting point of the nearest adjacent curve segment from the end point position, marking the searched line segment, and connecting the end point-starting point pair by a straight line;

and continuously searching the unmarked nearest neighbor curve segment from the end point of the last adjacent curve segment, continuously connecting and circularly iterating until the searched starting point is the starting point of the initial longest line segment, and ending the iteration.

Optionally, in the method for detecting the cleaning cleanliness of the universal spinneret orifice, reconstructing an edge contour closed curve of the high gradient threshold of the orifice to be detected from the first edge includes:

it is composed of

As a coordinate m of the contour point₁The number of the contour points;

meter L₁The peripheral P and filling surface S indexes are calculated, and then the indexes of qualified and clean N holes (N > 1024) with the same hole type are counted, and the average mu is calculated_p、μ_sAnd standard sigma_p、σ_sIndex, taking the optimal threshold value as tau_p＝μ_p±α*σ_p、τ_s＝μ_s±α*σ_sI.e. | P- μ_p|≤α*σ_pAnd | S- μ_s|≤α*σ_sAnd (3) when the cleanliness is qualified, detecting subsequent non-obvious dirt, and otherwise, directly detecting a next hole α which is 3.5.

the method adopts a parameterized straight-line segment and circular arc segment piecewise curve model based on control point positioning to carry out nonlinear registration of the contour to be detected, and specifically comprises the following steps: calculating a closed contour curvature curve, setting appropriate double thresholds according to curvature distribution characteristics to divide the contour curve into a plurality of straight lines and large and small circular arc sections, automatically positioning the position of each sectional control point, decomposing the closed curve into straight line sections and circular arc sections which are connected end to end, performing corresponding straight line or circular arc parameter fitting on each divided line section by adopting a least square method, calculating the intersection point of each adjacent line section after fitting as a new control point, and reconstructing the whole new parameterized segmented closed curve as a standard matching template.

extracting the edge contour closed curve L of the high gradient threshold of the hole to be detected₁Ith tracking

And a certain radius range (r) from front to back₂) Inner most distal profile

And

and performing circle fitting by using least square algorithm, and calculating c by using the radius of the fitted circle as curvature_iMeter L of₁Has an overall profile curvature distribution of C ═ C_i，i∈[1，m₁]}；

Setting a curvature division threshold c_thr1And c_thr2Segment length threshold g_thr；c_i＞c_thr1Marked as a straight line segment, c_i≤c_thr1And c is_i＞c_thr2Marking as a large arc segment, otherwise marking as a small arc segment:

the control points of each segment are respectively the starting position and the ending position of the segment, the starting position and the ending position of two adjacent segments share the same position, and meanwhile, a middle point control point which is positioned in the center of the circular arc and used for positioning the direction of the circular arc is added to the circular arc segment compared with the straight line segment;

dividing line segment after curvature divisionThe length of the length set G ═ G is obtained_i，i∈[1，f_g]Of g of_iRepresenting the length f of each segment line_gDividing the number of line segments;

search for minimum segment length g_kminE.g. g_kmin＜g_thrDiscarding the kmin segment,

the common control points of two adjacent segmentation line segments are respectively replaced by the middle points of the current segmentation line segment;

the length statistics and the minimum length judgment and fusion are carried out again, and the iteration is carried out in a circulating mode until all the segmentation line sections meet the conditions; if only one circular arc segment is obtained after the curvature division, the hole pattern is a conventional circle, and the hole pattern is divided into two connected circular arc segments; where the curvature division parameter is taken as c_thr1＝400，c_thr2＝120， g_thr＝40：

The parameterized segmented closed curve standard detection template is represented as follows:

F＝{f_i＝(f_i，x，f_i，y)，i∈[1，f_n]}，P＝{p_i＝(s，e)，i∈[1，f_l]，s∈[1，f_n]，e∈[1，f_n]，s≠e}，

Q＝{q_i＝(s，c，e)，i∈[1，f_c]，s∈[1，f_n]，e∈[1，f_n]，s≠e，s≠c，c≠e}，

f is that all the feature point sets P are straight segment index sets Q are circular segment index sets,

fitting a curvature radius set s for the arc segment as a starting characteristic point number e as an end point characteristic point number, c as an arc midpoint characteristic point number f_nFor the total number f of feature points_lIs the total number f of straight line segments_cThe total number of the arc segments.

Optionally, in the method for detecting the cleaning cleanliness of the universal spinneret orifice, the non-rigid registration includes:

the method comprises the steps of adopting a nonlinear least square parameter optimization method, controlling the position of an F by adjusting a line segment of a parameterized piecewise closed curve standard detection template, non-rigidly registering the parameterized piecewise closed curve standard detection template to an edge contour closed curve of a hole to be detected with a high gradient threshold value reconstructed from a first edge, and enabling the distance mean square deviation of each point of the contour curve to be detected projected to the template reconstructed piecewise closed curve { F, P, Q }, to be minimum;

is provided with

A projection distance function of a k point of the profile to be detected on a straight-line segment of the parameterized segmented closed curve standard detection template i,

a projection distance function of a k point of the profile to be detected in the arc section of the parameterized segmented closed curve standard detection template j is obtained; the nonlinear least square-based optimization objective function of the parameterized segmented closed curve standard detection template control F is as follows:

optimized template control point set F_τThe optimization method adopts a ceres-solvent algorithm library.

Optionally, in the method for detecting the cleaning cleanliness of the universal spinneret orifice, the edge profile closed curve of the hole to be detected with the low gradient threshold extracted from the second edge is as follows:

it is composed of

As a coordinate m of the contour point₂The number of the contour points;

from F_τThe reconstructed closed curve is:

it is composed of

As a coordinate m of the contour point₃Counting the number of points of the reconstructed curve;

calculating a reconstructed volume L₃The maximum value of the distances of all the search points (j, i) at the point of middle i is taken as the value of the distance curve of the point as follows:

the curvature weight of the straight line segment is set to be 1, the weighting function of the circular arc segment adopts a gradient interpolation method, and the weighting functions of the straight line segment and the circular arc segment are as follows:

it D_lIs L₃Middle l section straight line region point set D_cIs L₃C, middle arc area point set; the hybrid curvature weighting function is as follows:

which is a convolution operator, the Gaussian function g (i) is used to smooth the coefficient jump of the curve-weighted curvature transition region, mu₁＝0，σ₁＝3.0；

The single control point region gaussian weighting function is as follows:

f_icontrol point L for f₃A middle position index; mixing of

The resultant gaussian weighting function is as follows:

sigma of₂As a standard deviation sigma of a Gaussian function₂＝2.5；

The mixed weighting function multiplied by the mixed curvature weighting function and the single control point region gaussian weighting function is as follows:

W_t(i)＝W_t1(i)·W_t2(i) (8) dirt detection curve L after weighting of distance curve₅The following were used:

L₅＝{W(i)＝W_t(i)·D(i)，i∈[1，m₃]} (9)。

eliminating the unevenness of the floor reference by tophat transformation, and setting the global dirt threshold value to be 2.5;

let i dirty fingers P_iIs { W_i，H_i，S_i}，i∈[1，n]N is the number of local dirt, and the index of the overall cleanliness of the spinneret orifice is calculated as follows:

it h_thr＝15，w_thr＝4；

In the actual detection, a cleanliness threshold P meeting the process requirements is set_thr，P＞P_thrAnd when the hole is unqualified, automatically blowing the hole, then detecting again, and if the hole is not qualified, manually cleaning.

The invention also provides a system for detecting the cleaning cleanliness of the universal spinneret orifices, which comprises a closed curve extraction module, a template mapping module, a dirt detection curve module and a reference correction module, wherein the closed curve extraction module comprises:

the closed curve extraction module is configured to extract an edge contour closed curve of a standard clean hole high gradient threshold and a curvature curve thereof, and construct a parameterized segmented closed curve standard detection template containing control points;

in a fouling detection phase, the template mapping module is configured to map the parameterized segmented closed curve standard detection template onto a profile to be detected through non-rigid registration;

the dirt detection curve module is configured to construct a dirt detection curve, and the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template and a distance curve of the profile to be detected based on nearest neighbor searching and a mixed curvature curve;

the benchmark correction module is configured to form a global threshold by performing floor benchmark correction on the dirt detection curve to segment and locate dirt and quantify a cleanliness index.

In the method and the system for detecting the cleaning cleanliness of the universal spinneret orifice, the research of the characteristics of tiny dirt is taken as a starting point, the profile curve and the curvature curve of the cleaning orifice are extracted from the profile curve distribution characteristics of the edge of the orifice, a parameterized segmented curve template containing a control point is constructed, the template profile is mapped onto the profile to be detected through nonlinear registration, a distance curve between the registered template profile and the profile curve to be detected based on nearest neighbor search is constructed, a dirt detection curve of a normal circle and an abnormal orifice normalization is constructed by taking a mixed weighting curve as a weight, finally the dirt is segmented and positioned through curve singularity detection, and a cleanliness index for judging the dirt is constructed, so that the problem of the cleanliness detection of the universal spinneret orifice is solved.

Drawings

FIG. 1 is a graph of orifice image edge profile extraction and comparison to a globally optimal threshold segmentation in accordance with an embodiment of the present invention;

FIG. 2 is a reconstruction curve of the edge profile of the spinneret orifice according to another embodiment of the present invention;

FIG. 3 is a conventional circular qualified aperture area and perimeter index value histogram and its positive distribution fit according to another embodiment of the present invention;

FIG. 4 is a schematic diagram of a hole standard template generation according to another embodiment of the present invention;

FIG. 5 is a diagram illustrating standard template non-linear matching according to another embodiment of the present invention;

FIG. 6 is a schematic diagram of distance curve calculation according to another embodiment of the present invention;

FIG. 7 is a schematic view of a fouling detection curve according to another embodiment of the present invention;

FIG. 8 is a schematic view of a segmentation of a fouling detection curve according to another embodiment of the present invention;

FIG. 9 is a graph showing the results of conventional round hole fouling detection according to another embodiment of the present invention;

FIG. 10 is a graphical representation of flat hole fouling detection results according to another embodiment of the present invention;

FIG. 11 is a schematic representation of the results of a trilobe hole dirt test according to another embodiment of the present invention;

fig. 12 is a schematic diagram of a cross-hole dirt detection result according to another embodiment of the present invention.

Detailed Description

The following describes the cleaning cleanliness detection method and system for a universal spinneret orifice according to the present invention in further detail with reference to the accompanying drawings and specific embodiments. Advantages and features of the present invention will become apparent from the following description and from the claims. It is to be noted that the drawings are in a very simplified form and are not to precise scale, which is merely for the purpose of facilitating and distinctly claiming the embodiments of the present invention.

The core idea of the invention is to provide a method and a system for detecting the cleaning cleanliness of a universal spinneret orifice, so as to solve the problem of high false judgment rate of fine dirt of the conventional spinneret orifice.

The core idea of the invention is to provide a method for detecting the cleaning cleanliness of the universal spinneret orifice to improve the detection precision.

In order to realize the idea, the invention provides a method and a system for detecting the cleaning cleanliness of a universal spinneret orifice, wherein the method for detecting the cleaning cleanliness of the universal spinneret orifice comprises the following steps: extracting an edge contour closed curve with a high gradient threshold of a standard clean hole and a curvature curve thereof, and constructing a parameterized segmented closed curve standard detection template containing control points; in the dirt detection stage, the parameterized segmented closed curve standard detection template is mapped to the outline to be detected through non-rigid registration; constructing a dirt detection curve, wherein the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template, a distance curve of the contour to be detected based on nearest neighbor search and a mixed curvature curve; and carrying out floor reference correction on the dirt detection curve to form a global threshold so as to segment and position dirt and quantify a cleanliness index.

< example one >

The embodiment provides a method and a system for detecting the cleaning cleanliness of a universal spinneret orifice, wherein the method for detecting the cleaning cleanliness of the universal spinneret orifice comprises the following steps: extracting an edge contour closed curve of a standard clean hole high gradient threshold value and a curvature curve thereof, and constructing a parameterized segmented closed curve standard detection template containing control points; in the dirt detection stage, the parameterized segmented closed curve standard detection template is mapped to the outline to be detected through non-rigid registration; constructing a dirt detection curve, wherein the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template and a distance curve of the to-be-detected contour based on nearest neighbor search and a mixed curvature curve; and carrying out floor reference correction on the dirt detection curve to form a global threshold so as to segment and position dirt and quantify a cleanliness index.

Specifically, in the method for detecting the cleaning cleanliness of the universal spinneret orifice, the method for detecting the cleaning cleanliness of the universal spinneret orifice further includes: extracting a standard hole outline and a curvature curve thereof through standard clean hole image edge analysis, and carrying out segmentation and fitting on a straight line segment and a circular arc segment of the hole outline based on a dual-threshold segmentation algorithm of the curvature curve; according to the segmentation and fitting of the straight line segment and the circular arc segment of the standard hole profile, reconstructing a parameterized segmented closed curve standard detection template based on control points; extracting an edge contour closed curve of a high gradient threshold of a hole to be detected, calculating the length of the edge contour closed curve of the high gradient threshold of the hole to be detected and the area enclosed by the edge contour closed curve, detecting the significant dirt, if the detection result is unqualified, detecting the next hole to be detected, otherwise, detecting the non-significant dirt; extracting an edge contour closed curve of a hole to be detected with a low gradient threshold, mapping the edge contour closed curve of the standard clean hole with the high gradient threshold to the edge contour closed curve of the hole to be detected with the high gradient threshold through non-rigid registration, constructing a distance curve between the edge contour closed curve of the standard clean hole with the high gradient threshold and the edge contour curve of the hole to be detected with the low gradient threshold based on nearest neighbor search, and constructing a dirt detection curve by taking a mixed weighting curve as weight; through floor reference calibration processing of the dirt detection curve, using global threshold segmentation to position and quantify dirt, constructing a cleanliness index for dirt judgment, and judging whether hole dirt detection is qualified; and (3) taking the round hole, the flat hole, the three-leaf hole and the cross hole as research objects, and giving result analysis and conclusions.

Further, in the method for detecting the cleaning cleanliness of the universal spinneret orifice, the step of extracting an edge contour closed curve of a standard cleaning orifice high gradient threshold includes: extracting a hole edge contour curve by adopting a Canny edge operator, respectively detecting strong and weak edges by adopting a dual gradient threshold method, and only keeping the weak edges connected with the strong edges; setting a first gradient threshold σ_thr1And a second gradient threshold σ_thr2Respectively extracting strong contour edges and strong and weak combined contour edges, wherein the first gradient threshold is smaller than the second gradient threshold, and the first gradient threshold and the second gradient threshold respectively correspond to Canny detection double-gradient threshold [ sigma ]_thr1，3*σ_thr1]And [ sigma ]_thr2，3*σ_thr2](ii) a The extracted image edges are respectively a first edge and a second edge, wherein the first edge is used for constructing parameterized segmented closed curve standard detection and non-rigid registration and significant dirt detection, and the second edge is used for non-significant dirt detection. The contour curve extraction comprises contour edge extraction, closed curve reconstruction, prominent dirt hole pre-filtering and hole contour standard template generation.

Fig. 1 shows the spinneret hole image edge contour extraction and global optimal threshold segmentation ratio rollers (a) to (c) in this embodiment: cleaning round holes and images of non-significant dirt and significant dirt of the round holes; (d) (f): clean trilobal pores and their non-significant fouling, significant fouling images. Lines 1 to 4 are the original image and its optimal threshold segmentation map E₂E₁(ii) a 5 th behavior optimal threshold segmentation contour E₂The local edge distribution mixed signature of (a) is denoted by o and x, respectively.

The contour edge extraction includes: the outline of the image of the clean hole has smooth edges, and the gray scale of the dirt area in the non-clean hole changes with the size of the dirt and the distance from the orifice. It is difficult to find a suitable threshold value to meet various dirt region segmentation requirements by applying a global threshold value method, and the fluctuation of the threshold value can cause the segmented contour edge to drift locally. Compared with a global threshold method, the edge detection algorithm can optimize the position of the edge of the positioning hole profile according to the first-order gradient maximization and the second-order gradient zero crossing point characteristics. The method uses Canny edge operator to extract the hole edge contour curve, because the method adopts a dual gradient threshold value method to respectively detect strong and weak edges and only keeps the weak edge connected with the strong edge, the noise immunity is good, and the method is particularly suitable for extracting the contour edge of a dirt area. Here, two gradient thresholds sigma are set_thr1And σ_thr2To extract strong and strong + weak contour edges (sigma)_thr2＜σ_thr1) The corresponding two-ladder threshold value detected by canny is [ sigma ]_thr1，3*σ_thr1]And [ sigma ]_thr2，3*σ_thr2]Extracted image edge separation E₁And E₂Of which E₁Use as subsequent template construction and registration, significant fouling detection E₂Used as a subsequent non-significant fouling test. Figure 1 shows the contour edge location of round and tri-leaf holes of the same specification containing no fouling, no significant fouling and comparison with the optimal global threshold segmentation results. As can be seen from fig. 1(b), the global optimum threshold is not suitable for segmentation of insignificant fouling. As can be seen from the partial edge blending signature on line 5 of FIG. 1, the binary edge E of the clean hole₂The position deviation is small, but the dirt area in the dirt hole has two-value edges E₂The positional deviation is large, especially in the less soiled areas. Although the optimal threshold segmentation result is close to the actual contour edge, there is a pixel-level difference between the two, which is obviously not satisfactory for high-precision microscopic measurement applications. So the pure global optimal threshold method is not suitable for hole contour edge extraction.

FIG. 2 is a reconstruction curve of the edge profile of the spinneret orifice. Fig. 2(a) to (f) correspond to the reconstructed edge profile curves of fig. 1(a) to (f). Reconstructing the closed curve includes: in the usual case E₁The edge points in (1) may be discontinuous, or have bifurcation, or have isolated noise, and the closed contour curve cannot be extracted directly by the edge tracking algorithm. To ensure E₁The mesopore edge distribution has a contour closed characteristic, edge trimming and connection are required, and the algorithm flow comprises the following steps: (1) thinning and edge intersection point searching based on 8 neighborhoods, and breaking the edge intersection points; (2) searching curve end points and tracking the contour track, counting the length, and if the length is less than a set threshold value, reversely tracking along the end points and erasing the track; (3) searching the longest curve segment, marking the starting point and the end point position of the longest curve segment, searching the starting point of the nearest adjacent curve segment from the end point position, marking the searched line segment, and connecting the end point-starting point pair by a straight line; (3) and continuously searching the unmarked nearest neighbor curve segment from the end point of the last adjacent curve segment, continuously connecting and circularly iterating until the searched starting point is the starting point of the initial longest line segment, and ending the iteration. Fig. 2(a) to 2(f) correspond to the contour edge reconstruction closed curves of fig. 1(a) to 1(f), respectively.

FIG. 3 is a conventional circular qualified aperture area and perimeter index histogram and its positive distribution fit. Left: area index: and (3) right: perimeter index. Significant fouling pore pre-filtering includes: let E₁Reconstructed closed curve

It is composed of

As a coordinate m of the contour point₁The number of contour points. Meter L₁The peripheral P and filling surface S indexes are calculated, and then the indexes of qualified and clean N holes (N > 1024) with the same hole type are counted, and the average mu is calculated_p、μ_sAnd standard sigma_p、σ_sIndex, taking the optimal threshold value as tau_p＝μ_p±α*σ_p、τ_s＝μ_s±α*σ_sI.e. | P- μ_p|≤α*σ_pAnd | S- μ_s|≤α*σ_sIn practical application, one α is 3.5, the histogram of the detection indexes of the area and the perimeter of a conventional round qualified spinneret orifice and the positive space distribution fitting curve thereof are shown in fig. 3, the specification of the spinneret plate is PRA _0094 (outer diameter/mm) -144 (hole number) -0.18 (hole diameter/mm) x0.54 (hole depth/mm), and the fitting parameter in the diagram is mu_s＝0.025503，σ_s＝0.000305， μ_p＝0.599034，σ_p＝0.004071。

The hole outline standard template generation comprises the following steps: as can be seen from fig. 3, machining errors cause some discreteness in the shape of the spinneret hole profile. If the rigid matching based on least squares is directly performed by using the closed profile curve of fig. 2 as a template, the local matching deviation caused by the dispersion of the local shape of the normal aperture profile is easy to cause the false recognition of the subsequent dirt detection. Therefore, the invention adopts the parameterized straight-line segment and circular arc segment piecewise curve model based on control point positioning to carry out nonlinear registration of the contour to be measured, and the error is reduced to the minimum. The template construction process comprises the following steps: calculating a closed contour curvature curve, setting a proper double threshold value according to curvature distribution characteristics, dividing the contour curve into a plurality of straight lines and large and small circular arc sections, automatically positioning the position of each sectional control point, decomposing the closed curve into straight line sections and circular arc sections which are connected end to end (mostly adopting straight line and circular arc structures when a wire spraying hole is designed), performing corresponding straight line or circular arc parameter fitting on each divided line section by adopting a least square method, calculating the intersection point of each adjacent line section after fitting as a new control point, and finally reconstructing the whole new parameterized segmented closed curve as a standard matching template.

FIG. 4 is a schematic diagram of the generation of a well master template. (a) The method comprises the following steps A curvature calculation model; (b) to (e): FIG. 2(b) is a schematic diagram of a closed curve of the reconstruction template after a profile curvature distribution curve, arc segment segmentation and least square fitting of the arc segment; (e) (ii) to (i): and (e) a profile curvature distribution curve, segmentation of the straight line segment and the circular arc segment, positioning of the control points, least square fitting results of the straight line segment and the circular arc segment and a reconstructed template closed curve schematic diagram in the figure 2 (e).

The hole contour curvature curve extraction includes: extracting the closed curve L of the contour₁Ith tracking

And a constant radius range (r) from front to back₂) Inner most distal profile

And

and performing circle fitting by using least square algorithm, and calculating c by using the radius of the fitted circle as curvature_iCalculating the curvature distribution C ═ C of the whole contour_i，i∈[1，m₁]}. Fig. 4(a) is a schematic diagram of curvature calculation, and fig. 4(b) and 4(e) are profiles of curvature distribution of the closed contour curves of fig. 1(b) and 1(e), respectively.

The template generation comprises the following steps: setting a curvature division threshold c_thr1And c_thr2The segment length threshold is g_thr。c_i＞c_thr1Marked as a straight line segment, c_i≤c_thr1And c is_i＞c_thr2The mark is a large arc segment, and the mark is a small arc segment if the mark is not the large arc segment. The control points of each segment are respectively the starting position and the end position of the segment, the starting position and the end position of two adjacent segments share the same position, and meanwhile, a middle point control point which is positioned in the center of the circular arc and used for positioning the direction of the circular arc is added to the circular arc segment compared with the straight line segment. After curvature segmentation, length statistics of segmentation line segments is carried out to obtain a length set G ═ G_i，i∈[1，f_g]Of g of_iRepresenting the length of each segment, f_gIs divided into line segmentsAnd (4) counting. Search for minimum segment length g_kminE.g. g_kmin＜g_thrAnd discarding the kmin segment, and respectively replacing the control points shared by two adjacent segments with the middle point of the current segment. And then, the length statistics and the minimum length judgment and fusion are carried out again, and the iteration is carried out in a circulating mode until all the segmentation line segments meet the conditions. If only one circular arc segment is obtained after the curvature division, the hole pattern is a conventional circle, and the hole pattern is divided into two connected circular arc segments. Here the curvature division parameter c_thr1＝400， c_thr2＝120，g_thr＝40。

As can be seen from fig. 4, the corresponding straight line segment with the curvature greater than the straight line threshold (e.g., the black line segment in fig. 4(e) -4 (f)), the corresponding large circular arc segment smaller than the straight line threshold and greater than the circular arc threshold (e.g., the green line segment in fig. 4(b) -4 (c), 4(e) -4 (f)), and the corresponding small circular arc segment (e.g., the red line segment in fig. 4(e) -4 (f)) otherwise. FIG. 4(f) is a schematic diagram of the distribution of the segmentation line segments before fusion. Fig. 4(g) is a control point distribution diagram of fig. 4(f) after merging according to each straight line segment and circular arc segment. Fig. 4(h) is a graph of the results of least squares parametric fitting for each straight line and circle segment in fig. 4 (g). Fig. 4 (i) is a template closed curve graph of the solution result of the adjacent intersection control points of each fitting straight line and each arc segment in fig. 4(h) and the reconstructed template. The template parameterized model is represented as follows:

F＝{f_i＝(f_i，x，f_i，y)，i∈[1，f_n]}，P＝{p_i＝(s，e)，i∈[1，f_l]，s∈[1，f_n]，e∈[1，f_n]，s≠e}， Q＝{q_i＝(s，c，e)，i∈[1，f_c]，s∈[1，f_n]，e∈[1，f_n]，s≠e，s≠c，c≠e}，

f is a set of all feature points P which are straight line segment indexes and Q which is a circular arc segment index set

Fitting a curvature radius set s for the circular arc segment as an initial characteristic point number eThe number c of the end point feature point is the number f of the arc midpoint feature point_nFor the total number f of feature points_lIs the total number f of straight line segments_cThe total number of the arc line segments.

Further, cleanliness detection comprises non-rigid registration, distance curve construction, dirt detection curve construction, dirt segmentation and identification and cleanliness index generation. The non-rigid registration is that a non-linear least square parameter optimization method is adopted, the position of the F is controlled by adjusting the template line segment, and the template reconstruction segmented closed curve is non-rigidly registered to the E₁Reconstructed contour closed curve L to be detected₁In the above way, the distance mean square error of each point of the profile curve to be detected projected onto the template reconstruction segmentation closed curve { F, P, Q } is minimum. Order to

A projection distance function of k points of the profile to be measured on a straight line section of a template i

And (4) a projection distance function of the k point of the profile to be measured on the arc section of the template j. The nonlinear minimum two-times-based optimization objective function of the template control F is as follows:

optimized template control point set F_τ. Fig. 5 shows the non-linear registration result of the standard template and the reconstruction curve thereof in fig. 2(b) and 2(e), and the optimization method adopts a ceres-solvent algorithm library.

FIG. 5 is a standard template non-linear match. (a) (ii) to (b): the template non-linear matching of FIG. 2(b) and its reconstruction curve; (c) (ii) to (d): the template non-linear matching of FIG. 2(e) and its reconstruction curve; (e) is a partial enlarged view of (d). Blue-reconstruction curve, red-profile curve to be measured.

FIG. 6 is a schematic diagram of distance curve calculation. (a) The method comprises the following steps A local distance curve calculation process; (b) the method comprises the following steps Calculating a local distance curve; (c) and (d): distance curves of fig. 1(b) and 1 (e); (e) is L₄Schematic of conversion to 2-dimensional distance curveA drawing; (f) and (g) the smoothed 2d distance curves of (c) and (d). In the graphs (b), (e), (f) and (g), the blue curve is a template reconstruction curve, and the red curve is a distance curve.

The distance profile includes: let E₂(edge of weak gradient edge threshold extraction) extracted contour edge curve

It is composed of

As coordinates of contour points, m₂The number of contour points. Let F_τReconstructed closed curve

It is composed of

As coordinates of contour points, m₃The number of points of the reconstructed curve. FIG. 6 is a schematic diagram of distance curve calculation, wherein the blue dotted line L₃Red dotted line L₂Wherein fig. 6(a) is an enlarged view of a portion of the fouling in fig. 6 (b). When calculating the distance curve, m is determined₂×m₃Is initialized to-999, and then searches for the profile curve L₂From point j to point L₃Nearest to i and its distance

As indicated by the blue line segment in fig. 6 (a). After the search is completed, the maximum value of i columns is searched in the moment M

Such as

Watch L₃Point i has no L₂The midpoint is searched as the nearest neighbor point, and then the first i rows are searched in the front row and the back row in sequence

And

and (j1, i1) and (j2, i2), and L₃In (i) to (L)₂The shortest distance k between the middle j1 and the j2

Namely, it is

As shown by the green line segment in fig. 6 (a). Finally, calculating the reconstructed curve L₃The maximum value of the distance between all the search point pairs (j, i) at the point of middle i is taken as the value of the point distance curve, as shown in red in fig. 6(b), which is defined as follows:

FIG. 6(b) is a graph showing the search result of the local distance curve in FIG. 6(a), and it can be seen from the graph that the distance curve L is₄There are many local disturbances due to irregular dirt edge distribution, where L₄FIG. 6 shows the distance curve calculation results of FIGS. 1(b) and 1(e), where FIG. 6(e) L₄Converted into a schematic diagram of a 2-dimensional distance curve, wherein the curve is normally directed to the closed curve area of the template.

FIG. 7 is a schematic view of a fouling detection curve, and a blue curve is a template reconstruction curve. (a) And (d): the weighting curves of fig. 1(b) and 1(e), respectively; (b) and (e): 2-dimensional curve profiles of (a) and (d), respectively; (c) and (f): the fouling detection curves (red) of fig. 6(c) and 6(e), respectively.

Constructing a fouling detection curve comprises: the smaller the curvature, the larger the profile machining error in the spinneret machining process, resulting in a larger error in the reconstructed curve from the original edge profile, as shown in fig. 5 (e). In order to suppress the error as much as possible, the distance curve L is used here₄Is based onWeighting of the curvature. The curvature weight of the straight line segment is set to be 1, the weighting function of the circular arc segment adopts a gradient interpolation method, and the smaller the curvature is, the smaller the weight is. Formula 3, formula 4 show the weighted functions of straight line segment and circular arc segment, respectively, and D_lIs L₃Set of points in the middle section of the straight line region, D_cIs L₃And c, a middle arc area point set. Formula 5 is a mixed curvature weighting function, formula is a convolution operator, and the Gaussian function g (i) is used for smoothing the coefficient jump of the curvature transition region after curve weighting, and mu thereof₁＝0， σ₁3.0. Meanwhile, considering that the curvature distribution of the control point regions of adjacent line segments has certain discontinuity, in the nonlinear registration based on the error least square optimization principle, the optimal control point and the actual curvature transition point always have a matching error, if the profile to be measured has certain dirt or a large processing error, the error is represented as a singular bulge (shown in fig. 5 (e)) in a certain range of the control point in a distance curve, and the bulge is easily mistakenly judged as dirt in subsequent dirt segmentation, so that the singular bulge is restrained by adopting the weighting of a Gaussian function to the control point regions. Equation 6 is a single control point region Gaussian weighting function f_iControl point L for f₃Index of middle position, formula 7 is a Gaussian mixture weighting function, where σ₂Is the standard deviation of a Gaussian function, usually σ₂2.5. Equation 8 is a hybrid weighting function obtained by multiplying equation 5 by equation 7.

Dirt detection curve L after weighting of distance curve₅As shown in equation 9.

W_t(i)＝W_t1(i)·W_t2(i) (8)

L₅＝{W(i)＝W_t(i)·D(i)，i∈[1，m₃]} (9)

FIG. 7 is a scale detection curve calculation process of FIGS. 6(c) and 6(e), wherein the blue curve in FIGS. 7(b) and 7(e) is the template profile after non-linear registration, and the distances from the points on the curve along the normal to the red curve are their mixed weighting coefficients. As can be seen from a comparison of fig. 7(f) with fig. 6(g), the non-linear registration error in the dirt detection curve is substantially suppressed after the control points and the low curvature regions are hybrid weighted. Fig. 7(c) has substantially no difference compared to fig. 6(f) because the mixing weighting effect of the circular holes is almost negligible.

Fig. 8 is a schematic view of a segmentation of the fouling detection curve. (a) And (b): the curve floor reference estimate of FIG. 7(c) and its calibrated curve, respectively; (d) and (e): the curved floor reference estimate of FIG. 7(f) and its calibrated curve, respectively; (c) and (f): 2-dimensional fouling detection curves (red) of (b) and (e), respectively; (g) the method comprises the following steps A dirt threshold segmentation and parameter definition schematic diagram; (h) (ii) to (i): (c) and (f) a dirt segmentation and localization result map.

The dirt segmentation and identification comprises the following steps: in general, the clean hole contour edge distance curve and its dirt detection curve are theoretically nearly smooth, and the presence of local dirt appears as a singular protrusion on the detection curve, which can be segmented and located by a global threshold method. However, due to the processing error of the spinneret plate and the factors such as deformation of the plate and abrasion of the orifice after the spinneret plate is used for a plurality of times, the hole outline edge per se is locally distorted and is represented by the irregularity which is used as the floor reference on the dirt detection curve (as shown in fig. 8(a) and 8 (d)), and the global threshold method is not suitable for the partition and positioning of the dirt, so the tophat transform is adopted to eliminate the unevenness of the floor reference. Fig. 8 shows a baseline correction and segmentation positioning process of the fouling detection curve, with a global fouling threshold of 2.5. From the reference calibration results of fig. 8 (b) and 8(e), the reference unevenness of the dirt detection curve is well eliminated. Fig. 8(g) is a schematic view of the division of the fouling, wherein W, H and S-parameters represent the width, height and area of the fouling, respectively. In the dirt division and positioning mark maps of fig. 8(H) and 8(i), the dirt positioning is as indicated by the black line frame in the figures, with the local dirt parameters as indicated by (W, H, S) in the figures.

The cleanliness index generation comprises: let i dirty fingers P_iIs { W_i，H_i，S_i}，i∈[1，n]And n is the number of local dirt. In the spinning production, some sharp dirt can cause the phenomenon of filament drifting, so when the cleanliness index is generated, the weight of the sharp dirt needs to be increased. The overall cleanliness index of the spinneret orifice is calculated as follows:

formula h_thr＝15，w _thr4. In the actual detection, a cleanliness threshold value P meeting the process requirement is set_thr，P＞P_thrAnd when the hole is unqualified, automatically blowing the hole, then detecting again, and if the hole is not qualified, manually cleaning. The hole cleanliness indices of FIGS. 8(h) and 8(i) are shown as the corresponding hole center values.

The results of the fouling test were analyzed as follows:

fig. 9 is a schematic diagram showing the detection result of the conventional round hole fouling. (a) (ii) to (b): FIG. 1(a) is a graph showing a distance curve and a scale detection result thereof; (c) (ii) to (d): FIG. 1(c) is a graph showing a distance curve and a result of detecting fouling thereof; (e) the method comprises the following steps And the results of part of conventional round hole dirt detection are marked as 'V' in a qualified mode, and are marked as 'x' in an opposite mode.

FIG. 10 is a graph showing the results of flat hole fouling detection. (a) (ii) to (d): flat hole images, nonlinear registration of standard templates, dirt detection curves and dirt detection result graphs thereof; (e) the method comprises the following steps And if the partial flat hole dirt detection result is obtained, the qualified result is marked as 'V' and the opposite result is marked as 'x'.

FIG. 11 is a schematic diagram of the results of trilobe hole fouling detection. (a) (ii) to (d): FIG. 1(f) is a graph of standard template non-linear registration, fouling detection curves and their fouling detection results; (d) the method comprises the following steps And if the partial three-leaf-hole dirt detection result is qualified, the mark is √ and otherwise is marked as 'x'.

FIG. 12 is a graph showing the results of cross-hole fouling detection. (a) (ii) to (d): the cross hole image, the nonlinear registration of the standard template, the dirt detection curve and the dirt detection result graph thereof; (e) the method comprises the following steps And if the partial cross-hole dirt detection result is obtained, the qualified result is marked as 'V' and the opposite result is marked as 'x'. FIG. 9 is a schematic view showing the result of detecting fouling in a conventional round hole. For conventional circular holes, the curve L is closed due to the reconstructed template₃The difference between the curvatures of the two sections of circular arcs is small, and the curves are smooth, so that the generation of the dirt detection curves is hardly interfered by the matching error of the characteristic points, and the dirt detection curves after floor reference calibration can correctly reflect the singularity distribution characteristics of dirt. As can be seen from fig. 9, this detection can meet the dirt partitioning and positioning requirements for different levels of blockage. FIG. 9(e) P_thrFor 5, the qualified wells are marked with a √ ' in the lower right corner of the image, whereas marked with an ' x ' in the opposite direction, and the cleanliness index is given by the red value in the upper left corner of the image, as follows.

Fig. 10 to 12 are schematic diagrams showing the results of detecting dirt in flat holes, three-leaf holes and cross holes among common special-shaped holes. As can be seen from the figure, for weak and obvious dirt holes, the template nonlinear registration process can accurately reposition the control point to the target position of the contour to be measured. On the dirt detection curve, no matter the dirt is located the straight line section edge or the circular arc section edge, can both be correctly cut apart and fix a position. If critical errors occur in the pre-screening process of the significant dirt holes, the significant dirt holes flow into a subsequent detection stage, and the control point positioning optimization fails in the non-linear registration process because the edge distribution of the profile to be detected and the local difference of the template profile are too large. However, even in this case, the distance curve between the reconstructed closed curve profile and the profile to be measured still sufficiently reflects the singular projection of the actual fouling profile, so that the result only affects the magnitude of the cleanliness index quantization value and does not affect the qualified hole judgment based on the cleanliness index. Meanwhile, in a common situation, the through section of the special-shaped hole is larger than the through section of a conventional circle in size, and after ultrasonic cleaning, the probability of obviously blocking the hole is very low, so that the probability of the positioning failure in actual detection can be ignored.

The detection of the cleaning cleanliness of the spinneret orifices is an important quality monitoring link for chemical fiber production enterprises. The detection of the weakly significant dirt, particularly the weakly significant dirt of the special-shaped hole, is a technical problem in the software of the full-automatic microscopic examination instrument all the time. The practical dirt detection results of the conventional round, flat and flat filaments, the three-blade holes and the cross holes are given, and the result shows that the cleaning cleanliness detection method for the universal spinneret holes can meet the cleanliness detection requirements of the spinneret holes of chemical fiber enterprises.

In summary, the above embodiments describe in detail different configurations of the method and system for detecting the cleanliness of the spinneret holes, and it is understood that the present invention includes but is not limited to the configurations listed in the above embodiments, and any modifications based on the configurations provided in the above embodiments are within the scope of the present invention. One skilled in the art can take the contents of the above embodiments to take a counter-measure.

< example two >

The embodiment provides a general spinneret orifice washs cleanliness factor detecting system, general spinneret orifice washs cleanliness factor detecting system includes that the closed curve draws module, template mapping module, dirt detection curve module and benchmark correction module, wherein: the closed curve extraction module is configured to extract an edge contour closed curve of a standard clean hole high gradient threshold and a curvature curve thereof, and construct a parameterized segmented closed curve standard detection template containing control points; in a fouling detection phase, the template mapping module is configured to map the parameterized segmented closed curve standard detection template onto a profile to be detected through non-rigid registration; the dirt detection curve module is configured to construct a dirt detection curve, and the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template, a distance curve of the profile to be detected based on nearest neighbor search and a mixed curvature curve; the benchmark correction module is configured to form a global threshold by performing floor benchmark correction on the dirt detection curve to segment and locate dirt and quantify a cleanliness index.

The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. For the system disclosed by the embodiment, the description is relatively simple because the system corresponds to the method disclosed by the embodiment, and the relevant points can be referred to the method part for description.

The above description is only for the purpose of describing the preferred embodiments of the present invention, and is not intended to limit the scope of the present invention, and any variations and modifications made by those skilled in the art based on the above disclosure are within the scope of the appended claims.

Claims

1. A method for detecting the cleaning cleanliness of a universal spinneret orifice is characterized by comprising the following steps:

in the dirt detection stage, the parameterized segmented closed curve standard detection template is mapped to the outline to be detected through non-rigid registration;

constructing a dirt detection curve, wherein the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template, a distance curve of the contour to be detected based on nearest neighbor search and a mixed curvature curve;

2. The method for detecting the cleanliness of the washing of a common spinneret orifice according to claim 1, wherein the method for detecting the cleanliness of the washing of a common spinneret orifice further comprises:

extracting an edge contour closed curve of a hole to be detected with a low gradient threshold, mapping the edge contour closed curve of the standard clean hole with the high gradient threshold to the edge contour closed curve of the hole to be detected with the high gradient threshold through non-rigid registration, constructing a distance curve between the edge contour closed curve of the standard clean hole with the high gradient threshold and the edge contour curve of the hole to be detected with the low gradient threshold based on nearest neighbor search, and constructing a dirt detection curve by taking a mixed weighting curve as weight;

3. The method for detecting die cleanliness of spinneret holes according to claim 2, wherein extracting the edge profile closing curve for the standard die hole high gradient threshold comprises:

extracting a hole edge contour curve by adopting a Canny edge operator, respectively detecting strong and weak edges by adopting a dual gradient threshold method, and only keeping the weak edges connected with the strong edges;

setting a first gradient threshold σ_thr1And a second gradient threshold σ_thr2Respectively extracting strong contour edges and strong and weak combined contour edges, wherein the first gradient threshold is smaller than the second gradient threshold, and the first gradient threshold and the second gradient threshold respectively correspond to Canny detection double-gradient threshold [ sigma ]_thr1，3*σ_thr1]And [ sigma ]_thr2，3*σ_thr2]；

4. The method for detecting the cleanliness of the washing of the universal spinneret orifices according to claim 3, wherein reconstructing the control point-based parameterized segmented closed curve standard detection template comprises:

5. The method for detecting the cleanliness of the spinneret holes in the spinning nozzle according to claim 4, wherein the edge profile closing curve for reconstructing the high gradient threshold of the holes to be detected from the first edge is as follows:

it is composed of

As a coordinate m of the contour point₁The number of the contour points;

6. The method for detecting the cleanliness of the washing of common spinneret orifices according to claim 5, wherein the method for detecting the cleanliness of the washing of common spinneret orifices further comprises:

the method adopts a parameterized straight-line segment and circular arc segment piecewise curve model based on control point positioning to carry out nonlinear registration of the contour to be detected, and specifically comprises the following steps: calculating a closed contour curvature curve, setting a proper double threshold value according to curvature distribution characteristics, dividing the contour curve into a plurality of straight lines and large and small circular arc sections, automatically positioning the position of each sectional control point, decomposing the closed curve into straight line sections and circular arc sections which are connected end to end, performing corresponding straight line or circular arc parameter fitting on each divided line section by adopting a least square method, calculating the intersection point of each adjacent line section after fitting as a new control point, and reconstructing the whole new parameterized segmented closed curve as a standard matching template.

7. The method for detecting the cleanliness of the washing of the universal spinneret orifices according to claim 6, wherein the method for detecting the cleanliness of the washing of the universal spinneret orifices further comprises:

And a certain radius range (r) from front to back₂) Inner most distal profile

And

Setting a curvature division threshold c_thr1And c_thr2Segment length threshold g_thr；c_i＞c_thr1Marked as a straight line segment, c_i≤c_thr1And c is_i＞c_thr2Marking as a large arc segment, otherwise marking as a small arc segment;

the control points of each segment are respectively the starting position and the end position of the segment, the starting position and the end position of two adjacent segments share the same position, and meanwhile, a middle point control point which is positioned in the center of the circular arc and used for positioning the direction of the circular arc is added to the circular arc segment compared with the straight line segment;

after curvature segmentation, length statistics of segment segments is carried out to obtain a length set G ═ G_i，i∈[1，f_g]Of g of_iRepresenting the length f of each segment line_gDividing the number of line segments;

search for minimum segment length g_kminE.g. g_kmin＜g_thrDiscarding the kmin segmentation line segment, and respectively replacing the common control point of two adjacent segmentation line segments with the midpoint of the current segmentation line segment;

performing length statistics and minimum length judgment and fusion again, and performing loop iteration until all the segmentation line segments meet the conditions; if only one circular arc segment is obtained after the curvature division, the hole pattern is a conventional circle, and the hole pattern is divided into two connected circular arc segments; where the curvature division parameter is taken as c_thr1＝400，c_thr2＝120，g_thr＝40；

fitting a curvature radius set s for the arc segment as a starting characteristic point number e as an end point characteristic point number, c as an arc midpoint characteristic point number f_nFor the total number f of feature points_lIs the total number f of straight line segments_cIs a segment of a circular arcAnd (4) counting.

8. The method for detecting the cleanliness of the spinnerets in claim 7, wherein the non-rigid registration comprises:

the method comprises the steps of adopting a nonlinear least square parameter optimization method, controlling the position of an F by adjusting a line segment of a parameterized piecewise closed curve standard detection template, non-rigidly registering the parameterized piecewise closed curve standard detection template to an edge contour closed curve of a hole to be detected high gradient threshold value reconstructed from a first edge, and enabling the distance mean square error of each point of the contour curve to be detected projected to a template reconstructed piecewise closed curve { F, P, Q } to be minimum;

is provided with

a projection distance function of a k point of the profile to be detected on the arc section of the parameterized segmented closed curve standard detection template j is obtained; the nonlinear least square-based optimization objective function of the parameterized piecewise closed curve standard detection template control F is as follows:

9. The method for detecting the cleanliness of the spinneret holes in the universal spinneret according to claim 8, wherein the edge profile closing curve of the low gradient threshold of the holes to be detected extracted from the second edge is:

it is composed of

As a coordinate m of the contour point₂The number of the contour points; from F_τ

The reconstructed closed curve is:

it is composed of

calculating a reconstructed volume L₃The maximum value of the distances of all the search points (j, i) at the point of midpoint i is taken as the value of the distance curve of the point as follows:

it D_lIs L₃Middle l section straight line region point set D_cIs L₃C, middle arc area point set;

the hybrid curvature weighting function is as follows:

which is a convolution operator, the Gaussian function g (i) is used for smoothing the curve-weighted curvature transition regionCoefficient jump of the field, mu₁＝0，σ₁＝3.0；

The single control point region gaussian weighting function is as follows:

f_icontrol point L for f₃A middle position index; mixing of

The resultant gaussian weighting function is as follows:

sigma of₂As a standard deviation sigma of a Gaussian function₂＝2.5；

W_t(i)＝W_t1(i)·W_t2(i) (8)

dirt detection curve L after weighting of distance curve₅The following were used:

L₅＝{W(i)＝W_t(i)·D(i)，i∈[1，m₃]} (9)。

the method for detecting the cleaning cleanliness of the universal spinneret orifice further comprises the following steps:

let i dirty fingers P_iIs { W_i，H_i，S_i}，i∈[1，n]And n is the number of local dirt, and the overall cleanliness index of the spinneret orifice is calculated as follows:

it h_thr＝15，w_thr＝4；

In the actual detection, a cleanliness threshold P meeting the process requirements is set_thr，P＞P_thrThen, the hole is judgedUnqualified, at this moment, carry out automatic blow hole earlier, then retest, if still unqualified, carry out manual cleaning.

10. The utility model provides a general spinneret orifice washs cleanliness factor detecting system which characterized in that, general spinneret orifice washs cleanliness factor detecting system includes that closed curve draws module, template mapping module, dirt detection curve module and benchmark correction module, wherein:

the dirt detection curve module is configured to construct a dirt detection curve, and the dirt detection curve is formed by weighting the parameterized segmented closed curve standard detection template, a distance curve of the profile to be detected based on nearest neighbor search and a mixed curvature curve;