A method for automatically locating defects in coupons of flexible materials with non-homogeneous properties.

Abstract

The invention relates to a method for automatically locating defects in a coupon of flexible material having non-homogeneous properties intended to be cut into pieces, the method comprising: i the step of determining a rotation value to apply to one of the two contours to minimize the total surface area of ​​the non-overlapping zones (S33); the step of applying the rotation value to the location of each defect in the image of the initial coupon to pre-position each defect (S34); the step of determining a geometric transformation that locally minimizes the surface area of ​​the zones of the two non-intersecting contours (S35); and the step of applying the geometric transformation to the location of each pre-positioned defect according to its location within the contour to accurately position it in the image of the coupon in the ready-to-cut state (S36).

Description

[Technical Field]

[0001] The present invention relates to cutting pieces from coupons of flexible materials having non-homogeneous properties, in particular leather and natural leather.

[0002] The field of application of the invention is the manufacture of articles, in particular leather, which require the assembly of pieces cut from such coupons. The industries concerned are in particular the furniture, saddlery, leather, footwear and clothing industries. [Background technology]

[0003] The automated cutting of pieces from leather or natural leather for the purpose of manufacturing articles generally involves several main steps: first, digitizing the contours and defects of the leather, then arranging as many pieces as possible on the digitized leather, then, third, cutting the pieces from the leather according to a pre-defined arrangement, and finally, fourth, removing the cut pieces.

[0004] How these key steps are chained together results in different cleavage methods.

[0005] Thus, in what is called an "on-line" cutting method, the four main steps described above are carried out one after the other on the same cutting table.

[0006] The advantages of this approach are simplicity of configuration for the user and versatility and flexibility (in terms of configuration planning), while the main disadvantage of such an approach is the difficulty of properly synchronizing and balancing all the major steps.

[0007] According to another method, known as the "offline" method, the steps of digitizing the leather, placing the pieces, then cutting the pieces, and removing the pieces are performed separately on separate devices, with the time intervals between these steps being at the discretion of the user.

[0008] The advantages and disadvantages of this method are the opposite of those of the "online" method. In particular, separating the major steps makes it possible to manage steps that take more time than others (for example, by adapting the number of digitizers to the number of cutters and / or by adapting the computation time dedicated to the placement of the pieces). Conversely, this "offline" method requires organization on the part of the user and adds handling (and even storage) of the leather between steps.

[0009] Another drawback of this method, compared to the "online" method, is the introduction of a new step, which consists of repositioning the previously digitized hide on the cutting machine. In fact, repositioning the hide in the same way is not ergonomically easy. In addition, this process always introduces additional imprecision that must be taken into account during the placement of the pieces, by leaving a sufficient amount of unused space at the hide's borders, which reduces the efficiency of the placement.

[0010] There is also a method called "semi-offline", which is an intermediate method between the "online" and "offline" methods mentioned above. In this method, only the step of digitizing the leather contour and defects is transferred to a separate device and separated from the other main steps. In fact, digitizing the leather defects is often the most time-consuming step, and the quality of this digitization (i.e., taking into account all defects and identifying their exact location without enlarging them) determines the efficiency of placement and the reduction rate of rejects.

[0011] The disadvantage of this "semi-offline" method (similar to the "offline" method) is that it is difficult for the operator to position the leather. On the other hand, with this method, the placement of the pieces on the leather is not yet done, but inside the actual contours that are again digitized after the positioning step. Unlike the "offline" method, there is no need to leave any margins at the borders of the leather, which helps to make the placement work more efficient.

[0012] The difficulty with such a "semi-offline" method lies on the one hand in the ease of the leather positioning step, and on the other hand in the (automatic) accurate positioning of the defects within the new digitized contour.

[0013] In fact, leather is a flexible material and it will deform more or less depending on how an operator positions the leather on a table during the first digitization and how another operator positions the same leather on a different table during the second digitization. However, the way the leather deforms has a direct impact on the positioning of defects.

[0014] In addition to these deformations due to differences in how the operator placed the leather on the table, the leather may have been stored for weeks or even months between the two digitization steps. However, the storage conditions and differences in humidity and temperature between the two digitization steps also affect the possibility of deformation of the leather and therefore the location of defects.

[0015] Furthermore, while the leather positioning in the initial digitizing step is typically done on a digitizer with a relatively smooth polyurethane conveyor, the leather defect repositioning in the new digitizing step is done on a cutting machine with a felt conveyor that grips the leather tightly, which can still pose additional challenges in successfully performing the leather defect repositioning step.

[0016] The multiple deformations that the leather undergoes as described above are unfortunately not uniform and therefore unpredictable, and the defect location step must be able to identify the contours of the leather and the location of all defects as accurately as possible.

[0017] To solve this problem of repositioning defects, it is known to use a video projector that projects a pre-digitized image of the entire leather (or part of it) placed on the cutting machine, and the operator then repositions the entire leather and its defects bit by bit.

[0018] However, this method is still quite weak in terms of ergonomics and precision: the operator must actually pull the leather edge to correct its position while relying on the projected contour (which is not very accurate), which can create strong tension near the edge of the leather. Summary of the Invention

[0019] Therefore, the main objective of the present invention is to overcome these drawbacks by proposing a simple and ergonomic positioning method that automatically calculates the defect location.

[0020] According to the invention, this object is achieved by a method for automatically locating defects in a coupon of flexible material with non-homogeneous properties intended to be cut into pieces, said method comprising the steps of: - obtaining a digital image of the outline of the coupon in its initial state and the location of any defects; - acquiring a new digital image of the coupon profile after repositioning the coupon ready to cut; - superimposing digital images of the outline of the coupon in its initial and ready-to-cut states; - determining a rotation value to apply to at least one of the two contours in order to minimize the total surface area of ​​the zone bounded by the two non-overlapping contours; - applying a rotation value to the location of each defect in the digital image of the pristine coupon to pre-position each defect in the digital image of the ready-to-cut coupon; - determining a number of geometric transformations to locally minimize the surface area of ​​the zones of the two non-overlapping contours; - to accurately relocate defects in a digital image of the coupon ready to be cut; applying one of the geometric transformations to the location of each pre-located defect in the digital image of the ready-to-cut coupon according to the location of the defect in the outline of the initial coupon; Contains, in order.

[0021] It is worth noting that the method according to the present invention includes a defect location algorithm that can apply a specific geometric transformation to the location of each defect depending on the defect's location within the coupon contour. In other words, the geometric transformation applied to each defect is not the same for all defects. This method makes it possible to automatically and with high precision locate all defects on the coupon.

[0022] Furthermore, the method according to the invention only requires the input of the conveyor cutter to be equipped with a simple linear scanner (or one or more matrix cameras) similar to that used in the digitization step, so that the coupons are placed on the scanner in the position and manner desired by the operator. In particular, this solution allows users of cutters adapted for the "online" method to implement the "semi-offline" method without any changes to the cutter hardware.

[0023] Preferably, the rotation of the digital image of the coupon contour is relative to the burr centers of the two contours after superposition.

[0024] Each geometric transformation may consist of a rotation component and a similarity component.

[0025] In this case, a discrete field of angular sectors covering the two contours is advantageously constructed in a polar coordinate system with the origin at the respective varicenters of the two contours, and a geometric transformation, the components of which are rotational and similar, is associated with each angular sector in such a way as to locally minimize the surface area of ​​the non-overlapping zone of the two contours.

[0026] Preferably, the rotation and similarity ratio components of each geometric transformation are determined by a bisection method to find the rotation and similarity ratio values ​​that minimize the surface area of ​​the non-overlapping zone of the two contours by the angle values ​​associated with the geometric transformation.

[0027] The step of applying one of the geometric transformations to each pre-located defect is applied to each vertex of a polygon that encloses the defect's outline.

[0028] In this case, for each vertex of each polygon enclosing the defect contour, two angles that geometrically enclose this vertex are identified, and a combination of the rotation and similarity ratio values ​​of two geometric transformations associated with the two corresponding angle values ​​is applied to the coordinates of that vertex.

[0029] The present invention also relates to a method of cutting pieces from a coupon of flexible material having non-homogeneous properties, said method comprising: - digitizing the initial contour of the coupon and the location of defects; -For each coupon, a new step of digitizing the coupon contour on the digitizing-cutting table; - automatically locating defects on the coupon according to the method defined above; - positioning the pieces to be cut from the coupon; - cutting the pieces; Equipped with.

[0030] The invention also relates to a computer program comprising instructions for carrying out the steps of the method for automatically locating defects in coupons of flexible material with non-homogeneous properties as defined above.

[0031] The present invention also relates to a computer readable storage medium having recorded thereon a computer program comprising instructions for carrying out the steps of the method for automatically locating defects in coupons of flexible material having non-homogeneous properties as defined above. [Brief explanation of the drawings]

[0032] [Figure 1] 1 is a flow chart showing the main steps of a "semi-offline" piece cutting method according to the present invention.

[0033] [Figure 2] 4 is another flow chart illustrating the main steps of the defect location method according to the present invention;

[0034] [Figure 3] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 4] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 5] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 6] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 7] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 8] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 9] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 10] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; [Figure 11] 3 shows an exemplary implementation of the different steps of the defect location method according to the invention; DETAILED DESCRIPTION OF THE INVENTION

[0035] The invention applies in particular to the cutting of pieces from coupons of flexible material with non-homogeneous properties, such as leather or natural leather, for the purpose of manufacturing articles.

[0036] More specifically, the present invention is embodied in a cutting method called "semi-offline" cutting, the main steps of which are set out in the flow chart of FIG.

[0037] In the initial step S10 of the method, all coupons C1, ..., C i , …C nIt is planned to digitize the contours of the coupons and accurately determine the location of defects within the contours in these coupons. Digitizing the location of defects can be done automatically using a scanner or by operator.

[0038] This initial step of digitizing the coupons is performed on a digitizing table equipped with a scanner and is asynchronous compared to the other steps of the cutting method. i , …C n The digital data is stored and the digitized coupon can then be stored in a repository.

[0039] Then, each coupon C i is removed from its storage location to be laid flat on a cutting table equipped with a scanner. i undergoes a new digitization step (step S20) of its contours.

[0040] In the next step, the coupon C is ready for cutting. i 3. The defects are automatically repositioned within their contours according to the method of the present invention, which means repositioning from the data saved during step S10, but does not mean positioning these defects anew. This step S30 will be explained in more detail later.

[0041] In the next step S40, the pieces to be cut are divided into coupons C i Usually, the placement of the pieces to be cut depends on the geometric shape of these pieces, the possible links between them, and the coupon C i Furthermore, the layout is optimized to minimize material waste.

[0042] From this configuration a cutting program is created, which program is the result of converting the configuration into instructions for moving the cutting tools of the cutting table.

[0043] Coupon Ci The coupon is then transferred to the cutting zone of the table and cut according to the cutting program (step S50). The cut coupon is then transported out (step S60), and the cutting method is the same as that of the new coupon C. i+1 Then, the process is restarted from step S20.

[0044] 2 to 11, the coupon C according to the present invention i The main steps of the method for automatically locating defects are described below (corresponding to step S30 of the method described above).

[0045] In the first step S31, coupon C i The contours are again digitized using the scanner on the cutting table (the coupon is now ready to be cut).

[0046] By the program, coupon C i It is also possible to superimpose two digital images of the outline of the coupon, i.e., the image of the initial coupon I0 acquired in step S10 and the image of the ready-to-cut coupon I1 acquired in step S31 (step S32).

[0047] As shown in FIG. 3, this step is obtained by superimposing the burr centers B0, B1 of the two images I0, I1 of the coupon, respectively.

[0048] Once superimposed, a calculation algorithm makes it possible to determine a rotation value to be applied to the digital image I0, I1 of at least one of the two coupon contours in order to minimize the total surface area of ​​the zones of the two contours that do not overlap (i.e. do not intersect). This rotation is performed on the digital image of the coupon contour relative to the respective barycenters B0, B1 of the two contours (step S33).

[0049] For this purpose, the algorithm calculates the surface area of ​​the two contour images I0 and I1 and finds a rotation value to apply to one of them so that the value of "(contour image I0 / contour image I1)U(contour image I1 / contour image I0)" is as small as possible.

[0050] In the example of Figure 3, the surface area minimizing zone is the hatched zone.

[0051] In the next step (step S34), the rotation value determined in the previous step is applied to the position of each defect in the initial coupon digital image I0 in order to pre-position each defect in the coupon digital image I1 in the ready-to-cut state.

[0052] In the next step (step S35), a number of geometric transformations are calculated that locally minimize the surface area of ​​the zones of the two non-overlapping (or non-intersecting) contour images I0, I1.

[0053] One of the geometric transformations pre-calculated according to the position within the contour is applied to the position of the pre-located defect in the digital image I1 of the coupon ready to be cut, to accurately locate the position within the digital image I1 of the coupon ready to be cut (step S36).

[0054] The algorithm for performing these last two steps S35 and S36 is described in more detail below.

[0055] In particular, the algorithm for calculating the geometric transformation to be applied to each defect is performed in a polar coordinate system with origin O at the barycenters B0, B1 of the two contour images I0, I1, respectively.

[0056] In this reference frame, n angular sectors Δ1, ... Δ2 that cover the images I0, I1 of the two contours are i , …Δ nA discrete field of is constructed, and a geometric transformation with rotation and similarity factors determined to locally minimize the surface area of ​​the zones of two non-overlapping (or non-intersecting) contours is applied to each angular sector “Δ i " is associated with

[0057] For example, the angular discretization of images I0, I1 is chosen to be 1 degree, which is the angular sector of 360 Δ1, Δ2, ..., Δ i , …Δ 360 We construct a field with 360 different geometric transformations (angle sectors “Δ i Of course, it is possible to choose a different angle discretization.

[0058] Each angle sector “Δ i ” the calculation algorithm is to calculate the angle rotation “R i ” and similarity ratio “H i " to associate a geometric transformation consisting of

[0059] Each angle sector “Δ i ” the angular rotation component of the associated geometric transformation “R i " is calculated by the calculation algorithm in the following way:

[0060] As shown in Figure 4, “Δ i A 10° wide angular sector "A" centered on a " is considered, and this angular sector "A a " is weighted with a value of 2. a " is assigned.

[0061] Furthermore, “Δ i " and another angular sector "A" with a width of 30°. b " is also considered, and this other angle sector "A b " is weighted with a value of 1. b " is assigned.

[0062] The principle adopted here is that if the contour is quite "cut" (due to high weights), the first narrower angular sector ("A") is used to prioritize the search in the narrower angular sector. a ) with high weight “P a ” to select a wider second angular sector (“A b ”) with low weight “P b ". The search is expanded if the contour is relatively straight (because of the large weights). (In this case, the results in the narrow angular sectors are almost constant, and the results in the wide angular sectors dominate.)

[0063] The values ​​of the angle sectors and weights are shown here as an example, and of course other values ​​are conceivable, for example coupon types exhibiting particular geometric properties.

[0064] Each angular sector “Δ i ” for the initial coupon image I0, the angle rotation “R i Conversion by "I 0-R is then considered (see Figure 5).

[0065] From these data, the surface area is calculated by the following formula (shown in Figure 6): Ra is defined by

[0066]

number

[0067] Similarly, from these data, the surface area is given by the following equation (shown in Figure 7): Rb is defined by

[0068]

number

[0069] The calculation algorithm is (S Ra P a +S Rb Pb ) to minimize the angle rotation “R i The value of " is found by binary search.

[0070] Furthermore, each angular sector “Δ i ” for the similarity component of the related geometric transformation “H i " is calculated by the calculation algorithm in the following way:

[0071] Each angle sector “Δ i ” for the initial coupon image I0, the similarity ratio “H i Conversion by "I 0-H is taken into consideration (see Figure 8).

[0072] From these data, the surface area is calculated by the following formula: Ha (shown in Figure 9).

[0073]

number

[0074] Similarly, from these data, the surface area is given by S Hb (shown in Figure 10).

[0075]

number

[0076] The calculation algorithm is (S Ha P a +S Hb P b ) to minimize the similarity ratio “H i The value of " is found by binary search.

[0077] Angular sector “Δ i "For all geometric transformations, angle rotation" R i " and the similarity ratio "H iOnce the value of " is calculated, the calculation algorithm applies a geometric transformation to each pre-located defect in the digital image I1 of the coupon according to its polar coordinates.

[0078] More precisely, for each pre-located defect, a geometric transformation is applied to each vertex of a polygon that encloses the defect's contour.

[0079] So, for each vertex of the pre-located defect, the method performs a linear interpolation between the closest values ​​in the previously calculated discrete field.

[0080] FIG. 11 shows an example of such linear interpolation when the contour of a pre-located defect Z is enclosed by a polygon ABCD.

[0081] If Δ A , Δ B , Δ C , Δ D If A, B, C, and D are the angular coordinates of the vertices A, B, C, and D of the polygon enclosing the contour of the pre-positioned defect, then the calculation algorithm determines the angular rotation R of the geometric transformation to be applied. A , R B , R C , R D and the similarity ratio H A , H B , H C , H D Determine.

[0082] For each vertex of the polygon, "Δ1" and "Δ2" refer to successive angular coordinates in the previously calculated discretized field that surround the angular coordinate of that vertex. In the example angular discretization of images I0 and I1, for each degree, Δ2 - Δ1 = 1°.

[0083] Furthermore, for a vertex A of the polygon enclosing the contour of the defect Z, the following equation holds due to the angular discretization of images I0 and I1: Δ A =α1Δ1+α2Δ2 is Δ1 and Δ A is the angle between α and Δ Aand Δ2. Of course, similar equations can be written for the other vertices B, C, and D of the polygon.

[0084] More generally (i.e., the angle discretization is not necessarily in whole degrees), α1 and α2 are coefficients whose sum is equal to 1 (and correspond to the value of the corresponding angle divided by the value of the angle Δ2-Δ1).

[0085] R refers to the values ​​of the angular rotation and similarity ratio of the geometric transformation calculated for the angular coordinates Δ1 and Δ2 surrounding the angular coordinates of the vertices A, B, C, and D of the polygon, respectively. Δ1 , H Δ1 and R Δ2 , H Δ2 , the algorithm gives the value of the geometric transformation applied to vertex A as follows: R A =α1R Δ1 +α2R Δ2 H A =α2H Δ1 +α2H Δ2

[0086] Of course, similar equations are determined for the other vertices B, C, and D of the polygon.

[0087] Applying these equations to all vertices A, B, C, D of a polygon enclosing the contours of pre-located defect Z results in the polygon A'B'C'D' shown in Figure 11. This polygon encloses defect Z', which has been precisely located in the digital image of the ready-to-cut coupon.

[0088] This process is repeated for all defects that have pre-positioned the ready-to-cut coupon in digital image I1.

Claims

1. A method for automatically positioning defects in a coupon of a flexible material having heterogeneous properties that is intended to be cut into pieces, - Step (S10) to obtain a digital image (I0) of the contour of the coupon (Ci) in its initial state and the location of the defect (Z), - After repositioning the coupon to a cutting-ready state, a new digital image (I1) of the coupon contour is acquired (S31), - Step (S32) of superimposing digital images (I0, I1) of the coupon's contour in its initial state and cut-ready state, - The step (S33) of determining a rotation value to apply to at least one of the two contours in order to minimize the total surface area of ​​the zone enclosed by two non-overlapping contours, - Step (S34) of applying rotation values ​​to the positions of each defect in the digital image (I0) of the coupon in the initial state in order to pre-position each defect in the digital image (I1) of the coupon in the cutting preparation state, - The step (S35) of determining a set of geometric transformations to locally minimize the surface area of ​​two non-overlapping contour zones, - A step to precisely reposition defects in the digital image of the coupon in the cutting preparation state by applying one of the geometric transformations to the position of each pre-positioned defect in the digital image of the coupon in the cutting preparation state, depending on the position of the defect in the contour of the coupon in the initial state. (S36) and, A method that includes the following in order.

2. The method according to claim 1, wherein the rotation of the digital images (I0, I1) of the coupon contours is performed with respect to the respective varicenters (B0, B1) of the two contours after they have been superimposed.

3. The method according to claim 1, wherein the geometric transformation is composed of a rotation component (RA-RD) and a similarity ratio component (HA-HD).

4. In a polar coordinate system with the origin (O) at the respective varicenters of the two contours, a discrete field of angular sectors (Δ1, ... Δi, ... Δn) covering the two contours is constructed, and rotation The method according to claim 3, wherein a geometric transformation with determined components of (Ri) and similarity ratio (Hi) is associated with each angular sector (Δi) such that it locally minimizes the surface area of ​​the non-overlapping zones of the two contours.

5. The method according to claim 4, wherein the rotation and similarity ratio elements of each geometric transformation are determined by a bisection method to find the rotation and similarity ratio values ​​that minimize the surface area of ​​the non-overlapping zones of the two contours by the angular values ​​associated with the geometric transformations.

6. The method according to claim 1, wherein the step of applying one of the geometric transformations to each of the pre-positioned defects (S36) is applied to each vertex (A, B, C, D) of the polygon surrounding the contour of the defect (Z).

7. The method according to claim 6, wherein for each vertex (A, B, C, D) of each polygon enclosing the contour of a defect, two angles (Δ1, Δ2) geometrically enclosing this vertex are identified, and a combination of rotation and similarity ratio values ​​of two geometric transformations associated with the two corresponding angle values ​​is applied to the coordinates of that vertex.

8. A method for cutting pieces from a coupon of a flexible material having heterogeneous properties, - Step (S10) of digitizing the initial contour and defect location (Z) of the coupon (Ci); - For each coupon (Ci), a new step (S20) is taken to digitize the outline of the coupon on the digitization and cutting table; - Steps for automatically positioning defects in a coupon according to any one of claims 1 to 7. (S30) and; - The step of arranging the pieces to be cut from the coupon (S40),; - Step of cutting the piece (S50), A method that includes [a certain feature].

9. A computer program comprising instructions for performing a step of the method according to any one of claims 1 to 7.

10. A computer-readable recording medium on which a computer program comprising instructions for performing the steps of the method according to any one of claims 1 to 7 is recorded.

Images