EP1606612A1

EP1606612A1 - Method for inspecting the quality criteria of flat textile structures embodied in a multilayer form according to a contour

Info

Publication number: EP1606612A1
Application number: EP04723573A
Authority: EP
Inventors: Robert Daul; Jörg RUSCHULTE
Original assignee: Global Safety Textiles GmbH; Mahlo GmbH and Co KG
Current assignee: Global Safety Textiles GmbH; Mahlo GmbH and Co KG
Priority date: 2003-03-27
Filing date: 2004-03-26
Publication date: 2005-12-21
Also published as: US20070115462A1; JP2006521544A; CA2520444A1; US7375806B2; WO2004086013A1

Abstract

The invention relates to a method for inspecting quality criteria of flat textile structures embodied in a multilayer form according to a contour, in particular to woven, stitched, knitted, sewed or non-woven structure, preferably provided with, in particular cut areas or holes, separated or forming a strip, in particular for used for producing airbags. The inventive method is carried out with the aid of image forming inspection means, in particular optical inspection means, preferably a linear array camera or CCD-array camera. When a relative motion between the inspected structures and the camera is produced, the structure is arranged at least in an area and at a defined distance from the image forming inspection means preferably on the approximately flat surface of an control table or on an inspection strip. The texture of the structure is analysed according to a segmentation method. The characteristics like a centre of gravity, the surface, main axes etc are calculated for the different coherent segments of the same texture. An univocal system of co-ordinates for the structure and corresponding structures of the same type is defined on the basis of said characteristics, said system is non-variable with respect to the torsion, reflection, stretching/compression and the deformation of the system and makes it possible to define measuring points. In order to determine said system of co-ordinates, in addition to said segment characteristics, the position and the direction of identification threads which are intentionally introduced into the structure is taken into account. The respect of concerned distances is controlled and a quality report is produced on the basis of the measuring points which are recorded by the system on the basis of the quality requirements of a producer or consumer preferably in distant and marginal critical areas.

Description

Procedure for checking the quality criteria of flat, multilayered contour-related textile structures

description

The invention relates to a method for checking the quality criteria of flat, multilayered contour-related textile structures, in particular for the use of these structures for airbags, with the aid of imaging, in particular optical inspection means, preferably a surface or CCD line camera, with a relative movement between the Checking structure and the camera is generated and the structure is presented at least in regions at a defined distance to the imaging inspection means, preferably on a substantially flat surface of an inspection table or inspection belt.

From DE 36 39 636 AI an automatic inspection of textile webs is already known. There the flat material web is inspected with the aid of a parallel arrangement of color area cameras, the inspection being based on a color error detection carried out in real time and a local structural error detection carried out in real time.

The structural error analysis uses a cyclically described transient image memory for a more precise two-dimensional gray value evaluation in the event of errors detected in the local area which are uncertain. The amount of data for the inspection carried out in this way is considerable, so that it is not readily possible under industrial conditions with very short cycle times to check structures with regard to their quality features which have cutouts or edges which are related to the use and which have a certain minimum distance from a woven, knitted or sewn Do not fall below the edge.

The method and the associated device for correcting warpage according to EP 0 816 554 AI is to check whether webs printed with a pattern or otherwise provided with optically recognizable patterns are warped, in order to subsequently carry out pattern-correcting, if necessary , For this purpose, it is proposed that the pattern recognizable on the web be recorded by an image capturing device. Generated image signals are then fed to an image processing device, a distortion detection being carried out solely from the image signals of one or more recordings of the material web, for example by evaluating line elements, edges and / or color boundaries, and without the prior input of sample data.

With such a solution, a warping of a longer material web can be detected with sufficient accuracy in order to then control a straightening machine, but it is not possible to check an individual object which has already been subjected to various processing steps with regard to certain quality criteria as a result of the processing steps, since the individual objects, based on their position on an inspection table, cannot be compared with a more or less continuously running web.

During the manufacturing process, multi-layer, contour-based woven air bags are checked several times, ie the areal expansion of the incorporated contours is recorded at defined points and checked for compliance with the respective tolerance specification. According to the state of the art, this test is only carried out manually and is therefore very time-consuming, personnel-intensive and costly. In addition, the reliability of a visual, personal control is subject to fluctuations due to the given, changing physical or psychological constitution of the person who is responsible for the control. Such textile products, e.g. B. for the production of airbags, may have fluctuations in the dimensions, with the result that an error-free cut can be disturbed. This is because known control programs for cutting the air bags are not based on the contours, but on markings that are correlated with the contour. There is a risk here that tolerance specifications will not be met and that the products obtained in this way cannot be used as intended. From the foregoing, it is therefore the object of the invention to provide a further developed method for checking quality criteria, in particular of flat, knitted, stitched, sewn or multi-layer contour-related, knitted or sewn textile structures, the method particularly including a Allow analysis of individual structures that have a completely different contour and that are essentially only roughly aligned on an inspection table, without that in an otherwise complex manner, eg. B. must be created by manually creating the structure to be checked on an edge of the inspection table a defined starting position.

The object of the invention is achieved with a method as defined in claim 1, the subclaims representing at least expedient refinements and developments.

In the method according to the invention for checking the quality criteria of flat, multilayered contour-related woven or sewn, cut-outs or holes having textile structures, in particular individual structures, these in turn for use in airbags in particular, imaging inspection means known per se, e.g. B. one or more area or CCD line scan cameras. As imaging inspection means such. B. ultrasonic sensors, sonars or radar devices are used, with respect to optical

Inspection equipment Radiation in the visible but also in the invisible area, e.g. B. X-rays can be used.

Furthermore, a relative movement is generated between the structure or structures to be checked and the camera, the structure being presented at a defined distance on an essentially flat surface of an inspection table or inspection belt or being guided past the image field of the camera by a roller at a defined distance. In a first step, an image of the z. B. textile airbags and there is a storage of the obtained image data. Depending on the textile material, the picture can be taken with different lighting variants, for example in reflected light or transmitted light. The image capture device is constructed accordingly.

In a next step, image data correction can expediently be carried out such that the ratio of the resolution in the X direction to the resolution in the Y direction is 1.

The preprocessed image data are based on their different texture, which is also represented differently in the image, e.g. by different brightness, segmented and for each segment, e.g. Overall structures, hole and so on, features such as Area, center of gravity, circumscribing rectangle, main axes and the like are determined. A segment is understood here to mean a coherent area with a uniform texture. i.

The position of the centers of gravity of the segments relative to one another and the other features obtained is used to make the rotational position, mirror direction, stretching or compression, distortion or the like of the textile structure on the basis of standard comparison data which were previously obtained from the image of a textile target structure taken in the same way determined.

The determination of the distortion of the textile structure can additionally be ensured by the position, direction and angle detection of individual warp and weft threads that are particularly marked in color.

Obtained from these features entail a clear for the textile structure coordinate system can be determined, for example, with the origin in the center of gravity of the cut textile structure, X-axis, Y-axis g in weft or first major axis direction in warp or second Hauptachsenrichtun ^', wherein the coordinate system then used for the further measurement or control tasks. Subsequently, measurement points are defined preferably in critical distance and edge areas based on manufacturer or user quality specifications. Such measurement points can also be found and updated as part of a learning algorithm.

In particular, if it turns out that in certain cut-outs or marginal areas the risk of falling short or falling short of dimensions is vacant, a measurement point cloud can be defined in order to improve the certainty of the statement in relation to the quality in such critical areas.

All conceivable distances between the individual texture or in the image then segment boundaries can be measured, which are determined by the coordinate system described above. Tolerance limits can be defined for each measurement point. The measurement data itself can also be recorded in a quality protocol and / or used to extract the textile structure from further processing or delivery by means of a general good / bad statement.

According to the invention, there is the possibility, when recognizing an undefined position of the textile on the inspection table, in particular a stretching or compression of the textile structure in certain areas, to check whether and to what extent measurement points for determining critical distances are in the stretching and / or compression range. If this is the case, predefined measurement points can be rejected and it is possible to determine alternative measurement points. If it turns out that there are also alternative measurement points in the extension and / or compression area, then a new image is taken after the textile structure has been laid out again on the inspection table.

The image is captured using either a transmitted light or a reflected light method. In connection with this is the The inspection table or the inspection belt is designed as a fluoroscopic device or the surface of the inspection table or the inspection belt forms a contrast-generating background for the textile structure. This creates a sufficient contrast between the structure and the background or surface. Both types of lighting can be provided constructively, so that a changeover is not necessary.

In an advantageous method variant, the image is captured using a scanning process in such a way that the segmentation algorithm is already carried out on the existing partial image while the image data is being captured by the scanner.

A check of the dimensional accuracy of the textile structure can be carried out in such a way that different structure structures are imaged to produce contrast and at least one measuring line passing through the structure structures is inserted into the structure structure shown, at least a distance between at least two lying on the measurement line and by a contrast envelope defined measuring points is measured. The tissue structures are thus checked for their position and their position in relation to one another at suitable locations.

A contour determination on the textile structure can be carried out, for example, in such a way that at least one measuring line arranged on a determined contrast structure is set at least one measuring point describing a contour and a set of measuring lines with the assigned measuring points form a contour line. This contour line can then be used to control the movements of a further processing device, in particular a cutting device or a sewing device.

The invention will be explained in more detail below using an exemplary embodiment and with the aid of figures. The same reference numerals are used for the same or equivalent parts or method steps. It shows: 1 shows an exemplary view of a scanning device for obtaining the image of the textile structure,

FIG. 1 a shows an exemplary overall view of a curtain side airbag with a 5. adjusted resolution ratio in the X and Y directions,

2 shows an exemplary overall view of an image of the curtain side airbag segmented from a background,

3 shows an exemplary overall view of the segmented image of the curtain side airbag with exemplary centers of gravity of holes with respect to the position of the center of gravity of the overall image,

4a-c exemplary representations of a scan process in the context of the segmentation process,

Fig. 5a exemplary representations of different filter modes in comparison with an original detail,

5b shows an exemplary filter image of a weave contour of the curtain airbag as a closed surface,

Fig. 6 shows an exemplary filter image of a weave contour of the curtain airbag as an edge image,

7a shows an example of a principle of a characteristic thread detection,

Fig. 7b shows an exemplary representation of a center of gravity calculation and a determination of an object coordinate system of the curtain airbag,

7c exemplary representations of the mode of operation of a series of measurement filters, 8 shows an exemplary explanation of the principle of dimension determination on structures of the curtain airbag,

Fig. 9a, b exemplary illustrations a contour examination on ^{'ausgeschnitte-} NEN airbag,

10 shows an exemplary contour determination for the cutting of an airbag.

The exemplary embodiment according to FIGS. 1-10 illustrates an airbag final inspection.

In the example shown, the textile structure located on an inspection table is a curtain side airbag of defined dimensions with technically determined edge structures and cutouts in the form of elongated holes on one side of the airbag. In the following description, the terms “curtain side airbag” and “airbag” are used interchangeably for the sake of clarity, it being clear to the person skilled in the art that the method presented does not affect the manufacturing process of curtain side airbags or the manufacturing process of airbags is generally limited.

The object of this exemplary embodiment is to check whether the cutouts or holes have a defined position and whether the distances and the position of the holes from the knitted or sewn edge of the airbag material have a specific value. The dimensional accuracy of the airbag cut must also be checked.

According to FIG. 1, the airbag is first placed on a table IT in such a way that it can be detected by a camera system. It is illuminated from below by transmitted light or from above by incident light. In this way, the different optical appearances of the weaving or cutting contours in contrast to the incidence of light or passage of light can be displayed and recorded in a defined manner. For this purpose, the table IT expediently has a glass plate as a support surface in an exemplary embodiment.

At this table IT, a scanning device SE is fixed, consisting of a CCD line scan camera and a switchable lighting device, which is arranged above the table top for incident light and below the table top for transmitted light. This scan submission SE executes a relative movement to the table IT. In Fig. 1 this movement is indicated by a white double arrow. The movement of the scanning device SE takes place in analogy to the known flat bed scanning devices via the airbag to be checked, an image of the airbag being recorded in connection therewith. Following this scanning process, the image can be available as a gray value or color image.

The image determined in this way and shown by way of example in FIG. 1 a is then filtered in different ways for further processing and thus optimally processed for further use. Essentially two basic filter processes can be used.

A first filtering method is shown in FIG. 2. The cut airbag appears as a black segment on a white background. With this method, the segmentation threshold of binarization must be set so that the entire area is recognized as a coherent structure. The black pixels represent a geometrical area for which various features can be calculated, such as. B. area (number of black pixels), center of gravity S, main axis HA and others according to FIG. 2

Within the black segment, cutouts or holes are contained as white segments, which can also be seen as coherent geometric surfaces, with the same characteristics of surface area, center of gravity, main axis and so on (see FIG. 3). From the position of the centers of gravity and the main axes, ie the determined characteristics of the individual segments in general and to one another (FIGS. 2 and 3), the direction of rotation and mirroring as well as any existing compression or stretching and the distortion of the textile structure can be calculated.

In order to make a selection and to differentiate between relevant and irrelevant segments, limit values, such as B. minimum and maximum area, minimum and maximum X extent or the like, enclosed ..

The first evaluation step, i. H. As shown in FIGS. 4a to 4c, the segmentation takes place already during the image acquisition.

In order to arrive at a coordinate system that is unique for the airbag type, the center of gravity and the main axes (FIG. 2) - the second main axis is perpendicular to the first and is therefore not shown in the figure - of the airbag segment are used, so that the center of gravity of the airbag segment the coordinate origin, the two main axes correspond to the coordinate axes. If the centers of gravity of the holes L (see FIG. 3) are shown in this coordinate system, the mirror direction can easily be determined.

If the maximum dimensions of the airbag segment are determined in the direction of these coordinate axes, a clear scale based on the airbag can also be defined.

From the exact position of the hole center of gravity within the airbag segment, the coordinate system can still be adjusted with regard to distortion.

As can be seen from the section-by-section illustration in FIG. 5 a, the original image O shown on the left can be converted both into the segment image I and into an edge image K as part of the described and further segmentation processes or also by means of suitable image filters. 5b shows once again the overall image belonging to segment image I, FIG. 6 shows an edge image in the form of an exemplary overall view.

One possibility of determining a warp in the textile surface of the structure, in particular the airbag, in which the warp and weft threads in particular are not at a right angle to one another, is described below with reference to FIG. 7a. Here, the actual angle between warp and weft threads is determined, for example, in order to be able to carry out subsequent work on the weaving structure in coordination with this warping.

In this exemplary embodiment, differently colored, contrast-producing threads are detected which have been woven into the textile fabric of the structure. These appear in the illustration in FIG. 7a as a line grid LR made of characteristic threads. These characteristic threads then have the same warpage as the warp and weft threads of the entire material. An edge detection filter is expediently used in the detection of the characteristic thread.

In a next step, the center of gravity of the structure is calculated according to FIG. 7b and this is defined as the coordinate origin in the reference system of the structure. According to the warp determined in the previous step, the coordinate system is warped or distorted. The X axis of the reference system of the structure or the airbag is then aligned along the weft and the Y axis along the warp threads. The real position of the holes in this coordinate system defines the position of the structure, in particular the airbag. Work or actions that are tailored to the asymmetrical shape of the structure or the asymmetrical position of the holes are then always carried out at the appropriate position, regardless of the position and orientation of the structure.

The structure found can now be processed using further filter steps. Examples of this are shown in Fig. 7c. According to FIG. 8, a dimension determination on the airbag in a web under the circumstances described is carried out, for example, as follows. First, a single airbag is determined in the web as described above. The measurement filters prepare the image of the airbag in such a way that a two-layer area 10 is shown in white and a single-layer area 1, which can be a seam area, for example, in black. Measuring lines 11 and 12 with end points 2, 4, 6, 8, which lie in the two-layer region 10, are placed over this illustrated contour. Now it is checked along the measuring lines 11 and 12, starting from the end points 2, 4, 6, 8 and progressing towards the center, when a color change from white to black takes place. The seam area 1 is thus localized at these points. In the example from FIG. 8, this is the case at points 3, 5, 7 and 9. Now the distance along the measuring lines 11 and 12 between points 7 and 9 or 3 and 5 is determined.

Furthermore, reference points 13 to 3.6 can be assigned to the structure of the airbag from the start, from which the dimensions for the airbag dimensions can be found. By processing the image with the measuring filters, it is possible that they are no longer visible. For this purpose, the measuring lines 11 and 12 can also be used to determine distances on nearby contours, for example on 1. The measure relating to the reference points is then determined using correction values, factors and the like parameters. In principle, the color change can also be recorded from black to white in this example.

An exemplary contour check on a cut-out airbag will now be described with reference to FIG. 9a in conjunction with FIG. 9b. The contour of the airbag is determined using the steps described above. The measurement filters now process the image of the airbag obtained in such a way that the transitions, for example between a section line 103a to the background of the airbag and the section line 106 between single-layer and two-layer region 1, are represented as lines 101, 102, 103. Then the transitions are shown lines 101, 102 and 103, measurement lines 105, 110 are laid and distances along the measurement lines 105 and 110 between points 108 and 109 and 110a are determined using the same method described above.

These values can now be checked for their dimensional accuracy and their tolerance ranges. Airbags whose contour or dimensional dimensions do not meet the requirements can thus be identified and discarded.

In connection with FIG. 10, a contour determination for the cutting of airbags is now described by way of example. The respective individual airbag is detected in the web as in the steps described above. The measurement filters now prepare the image of the airbag in such a way that the transitions, for example between the single-layer and two-layer regions 206, are shown as lines 101, 102. Now, measurement lines 209 are placed over these lines 101 and 102 representing the transitions. Using the same method as described above, a point 208 is now determined on one side along the measuring line 209. Now a point 207 is determined at a fixed distance from point 208 along measuring line 209. A path 203 of a cutting device for cutting the airbags is defined on the basis of all points 207 on all measuring lines.

LIST OF REFERENCE NUMBERS

HA main axis

S focus

L holes O original cutout

I island detail

K edge image section

LR line grid made of characteristic threads

1 single layer Measuring line end point Measuring line end point Color change point Measuring line end point Color change point Measuring line end point two-layer area Measuring line Measuring line Reference point Reference point Transition line Transition line Transition line a Cutting line Measuring line Cutting line Measuring point Measuring point Measuring line a Measuring point Two-layer area Point Point Measuring line

Claims

claims

1. Method for checking the quality criteria of flat, multilayered contour-related, in particular woven, knitted, knitted, sewn or formed as a fleece, preferably having cutouts or holes, loose or in a web, in particular for the use of these structures for airbags, with imaging, in particular optical inspection means, preferably a CCD line scan camera or a surface camera, a relative movement being generated between the structure to be checked and the camera and the structure at least in certain areas at a defined distance from the imaging inspection means, preferably on an im the main flat surface of an inspection table or inspection belt or is guided over the roller at a defined distance past the image field of the camera, with the following steps:

Pick up the structure by means of. the inspection means, in particular the camera, and storing or temporarily storing the image data obtained; Segmentation of the image data obtained on the basis of the texture differences recognizable in the image;

Determination of segment features for the individual image segments, such as segment center of gravity - segment surface - segment main axis and / or surrounding rectangle or the like, on the basis of which a coordinate system which is unique for the structure and corresponding structures of the same type can be determined and which is suitable for rotation, reflection, stretching, compression and distortion of the structure is invariant, with the coordinate system being used to define measuring points.

2. The method according to claim 1, characterized in that the definition or specification of measuring points is preferably carried out in critical distance and edge areas on the basis of manufacturer or user quality specifications.

3. The method according to claim 1 or 2, characterized in that a check of the actual dimensional accuracy and predetermined distances, in particular from cut areas to seams or from seams to the outer edge of the structure is carried out.

4. The method according to any one of the preceding claims, characterized in that a quality report is created on the basis of the verification data determined.

5. The method according to any one of the preceding claims, characterized in that an optical detection of the position and / or direction of woven-in characteristic threads is used to determine the coordinate system which is unique for the structure to be checked witf.d.

6. The method according to any one of the preceding claims, characterized in that, in order to check the actual dimensional accuracy of the distances in addition to the segment boundaries (edge image K), image processing edge scanning algorithms are used.

7. The method according to any one of the preceding claims, characterized in that upon detection of an undefined position, in particular a region-specific stretching or compression of the textile structure, it is checked whether and to what extent measurement points for determining critical distances are in the stretching and / or compression range then discard selected measurement points, determine alternative measurement points or have the image of the textile structure in question repeated.

8. The method according to any one of the preceding claims, characterized in that the recording of the image is carried out using a transmitted light or incident light method, the inspection table or the inspection belt being designed as a transmitted light illumination device or the surface of the inspection table or the inspection belt forms a contrasting background for the textile structure.

9. The method according to any one of the preceding claims, characterized in that the textile structures are separated from the web on the basis of a contour line determined according to the aforementioned method.

10. Flat, multi-layer contour-based manufactured, in particular woven, knitted, knitted, sewn or formed as a nonwoven textile structure, in particular for use in airbags, which are separated from the web by a method according to claim 9.