JP2012193473A - Fabric cutting device and fabric cutting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a fabric cutting device capable of cutting a fabric corresponding to a pattern displayed in the fabric with a cutting line being corrected and to provide a fabric cutting method.SOLUTION: A fabric cutting device comprises: a matching process part 103 for extracting respectively a feature point of parts image data corresponding to each part to be cut and a feature point of captured image data of fabric captured by a capturing part 12 and extracting a matching point corresponding to the feature point of parts image data and the feature point of captured image data; a weighting process part 104 for weighing the extracted matching point; and a correction process part 105 for calculating a correction quantity of a cutting position based on the weighed matching point and correcting cutting line data based on the calculated correction quantity; and a cut controlling part 106 for controlling cutting means for cutting fabric based on the corrected cutting line data.

Description

  The present invention relates to a cloth cutting apparatus and method for cutting a cloth on which a pattern is displayed along a preset cutting line.

  When producing clothing and other products, the fabric is finished through a process of cutting fabrics, knitted fabrics, non-woven fabrics, etc. to create the original parts of the product, and joining the created parts by sewing etc. It is common. Further, in the case where a pattern is displayed in advance on the fabric before cutting, in general, a part is created by cutting according to the displayed pattern. Conventionally, the pattern displayed on the fabric has been formed by a dyeing process. However, in recent years, it has also been proposed to form a pattern using an ink jet recording apparatus.

  As means for cutting a fabric on which a pattern is created, cutting by an automatic cutting machine is known in addition to conventional manual cutting. When cutting with an automatic cutting machine, the cutting is usually performed according to predetermined cutting data set by an image processing apparatus such as CAD.

  However, when actually cutting the fabric on which the pattern is displayed, the fabric is easily distorted when the fabric is set on the cutting stage, and the position of the fabric is different from the position of the pattern corresponding to the set cutting data. A pattern may be displayed. In this case, it is necessary to correct the cutting data according to the actual pattern position.

  In order to deal with this, for example, in Patent Document 1, in a pattern matching method in an automatic cutting machine for cutting a pattern matching pattern from a sheet material with a pattern in which a pattern matching pattern is arranged with respect to a theoretical pattern, Calculate the deviation between the easily identifiable feature point position data selected for the pattern arrangement of the pattern and the feature point position data detected on the stretched sheet material, and the pattern matching pattern is calculated by the calculated deviation. There is described a pattern matching method for correcting. Further, in Patent Document 2, a position correction point mark associated with position information of each part is printed in advance on a cloth, and after being loaded on a cutting stage, the point mark is detected by a camera, and the amount of positional deviation from the design value is determined. A cutting position correction method is described in which position correction is performed by calculation.

Japanese Patent No. 2538514 Japanese Patent No. 4054556

  In the method described in Patent Document 1, correction is performed based on the deviation of position data using the design value and the feature point on the actual cutting object. However, the correction is performed by focusing on only the feature point, not the entire pattern. Therefore, it is not sufficient for correcting the positional deviation of the entire pattern. For example, in the case of a highly stretchable fabric, the fabric may be locally distorted when set, and even if the distortion is small in the area near the feature point, the distortion generated in other areas should be corrected. Becomes difficult. Further, in the method described in Patent Document 2, it is necessary to provide a space for printing point marks in the vicinity of each part, which is restricted when setting the shape of the part. And since the cutting position is corrected only by the amount of deviation of the point mark, it is difficult to correct the distortion of the printed pattern and the local distortion of the fabric as in the method described in Patent Document 1.

  In clothing, there are cases where a design that displays a specific pattern at a position corresponding to the chest or waist is used, but if such a design displays a specific pattern at a position shifted from the position of the chest or waist, The meaning as a design disappears. For example, in a design that displays a pattern pattern that makes a pocket image in the position of the chest and waist, it becomes meaningless as a design unless the pattern is accurately aligned with the position of the chest and waist.

  Such pattern alignment is important when responding closely to customer requests, for example, alignment corresponding to the customer's physique or alignment corresponding to the customer's selected pattern. come.

  Therefore, an object of the present invention is to provide a fabric cutting apparatus and method capable of correcting and cutting a cutting line corresponding to a pattern displayed on a fabric.

  The cloth cutting device according to the present invention includes a cloth setting means for setting a cloth on which a pattern is displayed on a cutting stage, a cutting means for cutting the set cloth into a shape of a part, and a display corresponding to the pattern. Storage means for storing image data and cutting line data corresponding to the shape of the part, photographing means for photographing the pattern displayed on the set fabric, and a pattern of the pattern photographed by the photographing means Features of photographing processing means for creating photographed image data, image extracting means for extracting part image data corresponding to each part based on the display image data and the cutting line data, characteristics of the photographed image data and the part image data Feature point extraction means for extracting each point, and the feature points of the captured image data and the feature points of the part image data are associated with each other A matching processing unit that extracts a matching point and performs weighting processing, and a correction processing that calculates a correction amount of the cutting position based on the weighted matching point and corrects the cutting line data based on the calculated correction amount. And cutting control means for controlling the cutting means so as to cut the fabric based on the corrected cutting line data. Furthermore, a projection unit that projects a cutting line onto the set fabric and a projection processing unit that performs processing so as to project the cutting line onto the fabric based on the corrected cutting line data.

  The cloth cutting method according to the present invention is a cloth cutting method in which a cloth on which a pattern is displayed is set on a cutting stage and the cloth is cut into a shape of a part, and the pattern displayed on the set cloth is displayed. Photographing is performed to create photographed image data of the pattern, and part image data corresponding to each part is extracted based on display image data corresponding to the pattern and cutting line data corresponding to the shape of the part. Extracting feature points of the image data and the part image data, respectively, extracting and weighting matching points that associate the feature points of the captured image data and the feature points of the part image data, and weighted matching A correction amount of the cutting position is calculated based on the point, and the cutting line data is corrected based on the calculated correction amount, and corrected. Cutting the fabric on the basis of the cutting line data.

  By having the above-described configuration, the present invention can perform cutting by correcting the cutting line corresponding to the pattern displayed on the fabric.

It is a schematic block diagram regarding the fabric cutting apparatus which concerns on this invention. It is a block block diagram regarding the fabric cutting device which concerns on this invention. It is a processing flow regarding the cutting process of a fabric. It is a schematic diagram regarding the picked-up image data of a fabric. It is a schematic diagram regarding the set cutting line and pattern. It is a schematic diagram regarding part image data. It is a schematic diagram which shows the feature point extracted in parts image data and picked-up image data. It is a schematic diagram which shows a matching point. It is a schematic diagram in the case of setting weighting to the pattern. It is a schematic diagram in the case of setting weighting to the pattern. It is a schematic diagram which shows the corrected cutting line.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. The embodiments described below are preferable specific examples for carrying out the present invention, and thus various technical limitations are made. However, the present invention is particularly limited in the following description. Unless otherwise specified, the present invention is not limited to these forms.

  FIG. 1 is a schematic configuration diagram relating to a fabric cutting device according to the present invention. In the cloth cutting device, the cloth F on which the pattern M is displayed is conveyed and set by the feed roller 2 on the cutting stage 1 and is cut into a predetermined part shape by the cutting head 3. The cutting head 3 is mounted on the carriage 4 so as to be movable in the width direction of the fabric F, and the carriage 4 is slidably attached to the guide rail 5. Then, the cutting head 3 and the carriage 4 are moved in the width direction and the longitudinal direction of the fabric F by a drive motor (not shown), respectively, so that the cutting head 3 performs a cutting operation while positioning it at a predetermined position on the fabric F. It can be cut into a predetermined part shape.

  The cutting head 3 is irradiated with laser from the laser irradiation device 7 through the mirror 6 so that the laser beam is irradiated from the cutting head 3 toward the fabric F so that the fabric F is thermally melted. Become.

  The cutting stage 1 has a large number of honeycomb-shaped through holes 1a formed on the entire surface of the fabric F, and an air intake device (not shown) disposed on the lower surface of the cutting stage 1 forms the fabric on the surface. The fabric F is set by the F being sucked and brought into close contact therewith.

  Above the placement surface of the cutting stage 1, a photographing device 8 for photographing the pattern M displayed on the fabric F and a projection device 9 for projecting and displaying a cutting line on the fabric F are installed.

  In this example, a laser is used as the cutting means, but a cutting means using a cutting blade may be used. The fabric F is a pattern material M displayed on a sheet material such as a woven fabric, a knitted fabric, or a nonwoven fabric. Anything displayed by the above method may be used. The pattern M may be displayed by being woven or knitted into the fabric F using a plurality of different colored yarns, or may be displayed by adhering a seal or the like to the fabric F.

  FIG. 2 is a block configuration diagram relating to the fabric cutting device. The fabric cutting device includes a control unit 10, a projection unit 11, an imaging unit 12, an input unit 13, a storage unit 14, a laser irradiation unit 15, a cutting head moving unit 16, a fabric feeding unit 17, and a fabric setting unit 18.

  The control unit 10 controls the operation of the entire apparatus, and the projection unit 11 operates the projection apparatus 9 to project and display a cutting line on the surface of the fabric F. The photographing unit 12 photographs the entire fabric F set on the placement surface of the cutting stage 1 by the photographing device 8 and creates photographed image data related to the pattern M displayed on the fabric F. The input unit 13 performs operation input such as parameters set when performing the cutting operation, or inputs correction data when the projected cutting line is confirmed and correction is necessary.

  The storage unit 14 stores data necessary for cutting processing such as pattern pattern image data, cutting line data, and photographed image data used when printing on the fabric F. I also remember the data. The laser irradiation unit 15 operates the laser irradiation device 7 to irradiate the laser toward the cutting head 3. The cutting head moving unit 16 operates a drive motor that moves the cutting head 3 and the carriage 4 to move the cutting head 3 to a predetermined position on the fabric F and position it. The fabric feed unit 17 operates the drive motor that drives the feed roller 2 to transport the fabric F to the placement surface of the cutting stage. The fabric setting unit 18 operates the air intake device disposed on the lower surface of the cutting stage to adsorb and set the fabric F on the placement surface.

  The control unit 10, the input unit 13, and the storage unit 14 can be configured using a known information processing apparatus such as a computer. A CPU is used as the control unit 10, an input device such as a keyboard and a mouse is used as the input unit 13, an internal hard disk and an external hard disk are used as the storage unit 14, and a program stored in the storage unit 14 is read to the CPU and processed. Thus, functions necessary for the cutting operation can be realized. In addition, the projection unit 11, the imaging unit 12, the laser irradiation unit 15, the cutting head moving unit 16, the fabric feeding unit 17, and the fabric setting unit 18 function by connecting corresponding devices to the information processing device. Can be realized.

  The control unit 10 includes an imaging processing unit 100, a projection processing unit 101, an image extraction unit 102, a matching processing unit 103, a weighting processing unit 104, a correction processing unit 105, and a cutting control unit 106. The imaging processing unit 100 processes the captured image of the fabric F input from the imaging unit 12, and generates captured image data. As will be described later, the projection processing unit 101 controls the cutting line data corrected by the correction processing unit 105 so as to be projected from the projection unit 11 onto the surface of the fabric F and displayed.

  Based on the pattern pattern image data and cutting line data stored in the storage unit 14, the image extraction unit 102 obtains pattern pattern image data included in an area corresponding to the shape of each part defined by the cutting line data. Extract and generate part image data for each part.

  The matching processing unit 103 extracts a plurality of feature points from the part image data and the photographed image data, respectively, and performs feature matching between the feature points of the part image data and the feature points of the photographed image data. A pair of feature points of both close image data is extracted as a matching point.

  A known matching processing method is used for extracting feature points of image data. For example, a local feature amount such as SURF (Speed-Up Robust Features) or SIFT (Scale-Invariant Feature Transform) may be calculated based on the image data. In SURF, a region where the matrix value of the Hessian matrix becomes maximum is extracted as a feature point based on data for each pixel of image data. Then, a circular area centered on the selected feature point is set, and a gradient direction (direction of change in density data of the pixel) for each pixel in the circular area is obtained by Haar-wavelet conversion. By expressing the obtained gradient direction as a vector, the sum of the vectors is set as the gradient direction of the selected feature point. Then, the circular region set by the scale value is divided into 16 blocks of 4 × 4, and a four-dimensional feature vector is obtained based on the sum of gradient directions of feature points selected as samples in each block, and 16 × Four 64-dimensional local feature quantities are set.

  The set local feature amount is compared between the part image data and the captured image data to perform corresponding point search. In the corresponding point search, for example, the Euclidean distance is compared between the local feature amount of one feature point of one image data and the local feature amount of all feature points of the other image data, and the feature having the smallest distance is obtained. What is necessary is just to make points correspond.

  Since the feature points set as the corresponding points may be set erroneously, a known method such as RANSAC (RANdom SAmple Consensus) is used to create a pair of feature points that eliminates the feature points set in error as much as possible. Extract as matching points. By performing such processing, it is possible to obtain a stable processing result that is not easily affected by instability factors such as noise, shadow, and brightness fluctuation in image data.

  In addition to the method described above, a known method can be used as a general matching process for image data.

The weighting processing unit 104 preferentially uses a matching point of a region corresponding to a part of the pattern pattern or cutting line portion that is important in pattern matching in the part image data in the correction processing unit described later. Then, a weighting process is performed on the extracted matching points.
For example, the weighting process may be performed so that the matching point having the largest feature amount or the region where the matching points are concentrated is dominant.
In addition, for example, a weighting process can be performed on a pattern desired to be accurately positioned. Examples of the pattern desired to be accurately positioned include a pattern displayed at one point on the chest and waist. In addition, pattern patterns (so-called fake patterns) that image pockets, necklaces, ties, belts, vests, buttons, etc. give the impression that they are worn by displaying them at specific positions, so positioning is accurate Need to be done. In addition, it is preferable to perform the weighting process so that the matching points extracted from the pattern pattern near the sewing position at the time of sewing and the pattern pattern straddling the sewing position are dominant.
By performing the weighting process, when the correction amount is calculated by the correction unit 105 described later, the correction amount is calculated by preferentially using the matching points of the pattern pattern that has been gained by the weighting process, and positioning is performed. The pattern matching of the pattern desired to be performed accurately is performed preferentially.
In addition, by performing weighting processing, even if the number of matching points extracted from the pattern that you want to accurately position is small compared to other patterns, it greatly affects the matching points of other patterns. Therefore, it is possible to preferentially perform the pattern matching of the pattern desired to be accurately positioned.

As a method of setting a region for pattern matching preferentially by weighting processing, it may be automatically set according to the distribution of the feature amount. It can also be created and set. Alternatively, the area to be weighted may be set while displaying the captured image.
In addition, as a method of performing weighting processing on each matching point, for example, a large coefficient is applied to the feature amount of the matching point in the region where pattern matching is preferentially performed, and the matching point in the region where the region is not set is small. When correction is performed by a correction processing unit, which will be described later, such as by applying a coefficient, the feature amount of the matching point in the region where pattern matching is preferentially larger than the feature amount of the matching point in the region where the region is not set There are ways to operate. The coefficient may be set as appropriate, or a value (for example, a color density) associated with the weighting map may be used.
In addition, after performing the weighting process, it is possible to select from all the matching points the one with a large feature amount by setting a threshold value or the like, and the correction processing unit can calculate the correction amount at the selected matching point. It is.

  The correction processing unit 105 calculates the correction amount of the cutting position using the matching points extracted by the matching processing unit 103. At the plurality of extracted matching points, the difference in coordinate position between the matching points is not constant due to the fabric distortion or the like, and thus it is not possible to simply obtain the deviation value and use it as the correction amount. Therefore, the projection transformation matrix is estimated and calculated from a plurality of matching points, and the parallel movement amount and the rotation angle are determined based on the obtained projection transformation matrix to obtain the correction amount of the cutting position. In this case, the projective transformation matrix is used, but the correction amount may be calculated by other known methods (for example, a lookup table, a neural network, etc.). Then, the cutting line data is corrected based on the obtained cutting position correction amount.

  The cutting control unit 106 controls the laser irradiation unit 15 and the cutting head moving unit 16 so as to irradiate the laser while moving and positioning the cutting head 3 based on the corrected cutting line data, so that the fabric F is irradiated with the laser. Cut and cut out each part.

  FIG. 3 is a processing flow relating to a fabric cutting process. First, a pattern image displayed on the entire fabric is photographed from above with respect to the fabric set on the cutting stage, and photographed image data is obtained (S101). FIG. 4 is a schematic diagram relating to the pattern of the entire fabric F taken. In this example, two pattern patterns M1 to M8 are displayed at one point for each of the four parts P1 to P4. In FIG. 4, for easy understanding, hatching is displayed corresponding to the shapes of the parts P <b> 1 to P <b> 4, but only a pattern is displayed on an actual fabric.

  Next, part image data corresponding to one part is extracted based on the image data relating to the pattern and the cutting line data (S102). FIG. 5 is a schematic diagram (FIG. 5A) regarding the cutting line and a schematic diagram regarding the pattern (FIG. 5B). The cutting line data p1 to p4 are set as cutting line data corresponding to the shapes of the four parts as shown in FIG. 5A, and the image data m1 to m8 related to the pattern are shown in FIG. As shown, the position of each image is set in accordance with the shape of the part. In FIG. 5B, in order to facilitate understanding, hatching is displayed corresponding to the lines of the cutting line data p1 to p4, but the image data is only a pattern.

  When extracting the part image data, the part area surrounded by the cutting line data is overlapped with the image data related to the pattern and the image data corresponding to the part area is extracted to obtain the part image data. FIG. 6 is a schematic diagram relating to part image data relating to the part P1. The cutting line data p1 shown in FIG. 5A and the image data related to the pattern are overlapped to extract the image data m1 and m2 related to the pattern included in the region of the part P1, thereby generating part image data.

Next, feature points are extracted from the obtained photographed image data and one part image data (S103). Here, the feature points are extracted by calculating local feature amounts such as the above-described SURF, for example. FIG. 7 is a schematic diagram showing the feature points extracted in the part image data and the photographed image data, and the feature points are indicated by white circles. In this example, a plurality of feature points s _i (i = 1,..., N) corresponding to the pattern are extracted from the part image data shown in FIG. Feature points S _j (j = 1,..., M) are extracted.

Then, the feature amount is compared between the feature point s _i of the part image data and the feature point S _j of the photographed image data, the feature points are matched to extract a pair of feature points, and an incorrect correspondence relationship is obtained. The matching points are extracted by excluding those (S104). When comparing the feature values of the feature points of both image data, as described above, the feature points of the combination that minimizes the Euclidean distance of the feature values are extracted as a pair, and the extracted feature point pairs are erroneously included. If there is an object that has a corresponding relationship, the above-described method such as RANSAC is used to eliminate the erroneous correspondence relationship, and feature point pairs that are not excluded are extracted as matching points. FIG. 8 is a schematic diagram showing matching points. In this example, the feature points s _{1 to} s ₃ related to the image data m1 of the part image data are extracted as matching points associated with the feature points S _{1 to} S ₃ related to the image data M1 of the captured image data, and the image data m2 The feature points s ₄ and s ₅ are extracted as matching points associated with the feature points S ₄ and S ₅ related to the image data M2.

  Next, a weighting process is performed on the extracted matching points (S105). Only one matching point to be weighted or a plurality of matching points may be selected. For example, as shown in FIG. 9, the selection of the matching point that is dominant by weighting is performed by selecting an important pattern for pattern matching in advance in each part (in this example, indicated by a dotted circle W). The weighting process may be performed so that the matching points regarding the image data regarding the pattern are dominant. In addition, as shown in FIG. 10, image data related to a pattern may be displayed, and a region for pattern matching may be preferentially selected by a coloring process method such as a brush tool (in this example, hatching). Displayed with the displayed circle W '). In this example, the image pattern-related image data is used to preferentially select the pattern matching area, but the captured image data may be displayed and selected. As described above, if an area where pattern matching is preferentially performed in the coloring process is selected, not only a designer but also a general consumer can easily select the pattern while viewing the image data displayed on the screen.

  Next, the correction amount of the cutting position is calculated based on the weighted matching points (S106). As described above, the correction amount of the cutting position is calculated as a correction amount by calculating a projection transformation matrix from the extracted matching points and determining a parallel movement amount and a rotation angle based on the calculated projection transformation matrix.

  After the correction amount of the cutting position is obtained for one part, it is checked whether there is cutting line data corresponding to the next part (S107). If there is a next part, the process returns to step S101 to photograph the fabric. After that, part image data corresponding to the next part is extracted, and the same processing as the above-described extraction of feature points and matching points, weighting processing, and correction processing is performed. After the correction amount of the cutting position is obtained for all parts, the process proceeds to step S108, and the cutting line data is corrected so that the cutting line of each part is translated or rotated based on the obtained correction amount (S108). FIG. 11 is a schematic diagram when the corrected cutting line data p <b> 1 ′ to p <b> 4 ′ are superimposed on the photographed image data of the fabric F. Each cutting line data is weighted and corrected with respect to the pattern surrounded by the one-dot chain line circle W, and therefore the cutting line is set so that the weighted pattern pattern is accurately aligned. .

  Then, the cutting line is actually projected and displayed on the surface of the fabric F based on the corrected cutting line data (S109), and finally the operator confirms the cutting line with the naked eye, and then along the cutting line. The cutting process is performed by moving the cutting head and irradiating the laser (S110).

  As described above, when the fabric is set in a locally distorted state when the fabric is carried into the cutting stage or when the pattern is formed on the fabric, the pattern itself is distorted due to the stretchability of the fabric. Even in this case, a plurality of feature points are matched based on the part image data and the photographed image data, a plurality of matching points are set, and a correction amount of the cutting position is set from the matching points, so that the cutting line is actually set. The position can be corrected to an optimum position according to the pattern of the fabric.

  Also, if there is a pattern that is important for pattern matching, such as a specific pattern placed on the chest or waist, or if you want to accurately align the pattern in relation to the sewing location By weighting the matching points in the region, it can be reflected in the correction amount of the cutting position, and pattern matching can be performed more strictly.

  Even when cutting fabrics with different patterns according to the cutting lines individually corresponding to the customer's physique or the patterns selected according to the customer's preference, cutting according to the pattern of each fabric Parts can be cut by finely correcting the line, and a fabric cutting apparatus capable of handling a variety of products in small quantities can be obtained.

DESCRIPTION OF SYMBOLS 1 ... Cutting stage 2 ... Feed roller 3 ... Cutting head 4 ... Carriage 5 ... Guide rail 6 ... Mirror 7 ... Laser irradiation apparatus 8 ... Imaging apparatus 9 ...・ Projector

Claims (3)

  Corresponding to the cloth setting means for setting the cloth on which the pattern is displayed on the cutting stage, the cutting means for cutting the set cloth into the shape of the part, the display image data corresponding to the pattern and the shape of the part Storage means for storing cutting line data, photographing means for photographing the pattern displayed on the set fabric, and photographing processing means for creating photographed image data of the pattern photographed by the photographing means Image extracting means for extracting part image data corresponding to each part based on the display image data and the cutting line data, and extracting feature points of the photographed image data and the part image data, respectively, and the photographed image Matchon that extracts matching points that associate the feature points of the data and the feature points of the part image data A processing unit; a weighting processing unit that weights the extracted matching points; and a correction amount of a cutting position is calculated based on the weighted matching points, and the cutting line data is calculated based on the calculated correction amount. A fabric cutting apparatus comprising: correction processing means for correcting; and cutting control means for controlling the cutting means so as to cut the fabric based on the corrected cutting line data.   The projection means which projects a cutting line on the set fabric, and the projection processing means which processes so as to project the cutting line on the fabric based on the corrected cutting line data. The fabric cutting apparatus according to the description.   A cloth cutting method for setting a cloth on which a pattern is displayed on a cutting stage and cutting the cloth into a shape of a part, and photographing the pattern displayed on the set cloth Creating data, extracting part image data corresponding to each part based on display image data corresponding to the pattern and cutting line data corresponding to the shape of the part, and features of the photographed image data and the part image data Each point is extracted, a matching point that associates the feature point of the photographed image data and the feature point of the part image data is extracted, the extracted matching point is weighted, and the weighted matching point is extracted. Based on the calculated correction amount, the cutting line data is corrected based on the calculated correction amount. Fabric cutting method for cutting the fabric on the basis of the cutting line data.

Images