Shape recognition device and sewing machine

Abstract

Shape recognition device (4), comprising an imaging means (41) for imaging a first object and a second object from above, wherein the two objects are arranged so that they vertically overlap each other, and which detects an interface between the first object and the second object from acquired image data obtained via the imaging means (41), wherein the imaging means (41) focuses an upper surface of either the first object or the second object at a focal distance, such that one of the objects first object and second object is imaged at the focal distance and the other of the objects first object and second object is imaged at a non-focal distance, wherein the shape recognition device (4) comprises: an image recognition control unit (43) which, when luminance values ​​of respective pixels of captured image data, which are captured by the imaging means (41), are greater than luminance values ​​of surrounding pixels, performs sharpening processing by converting the luminance values ​​of the respective pixels of the captured image data into larger luminance values ​​according to difference values ​​between the luminance values ​​and, after sharpening processing, performs binarization processing of the captured image data and edge extraction to detect an interface between the first object and the second object.

Description

TECHNICAL FIELD

[0001] The present invention relates to a shape recognition device which recognizes the shape of a sewing object to be sewn to a sewing material, and to a sewing machine comprising the same. BACKGROUND OF THE INVENTION

[0002] Prior art sewing devices include a fabric-feeding lower plate and a fabric-feeding outer frame, which are designed to operate within a horizontal plane. The procedure for sewing a sewing object to a sewing material using these sewing devices is as follows. First, the sewing material is clamped by the fabric-feeding lower plate and the fabric-feeding outer frame. Next, the sewing object is placed on the sewing material. A sewing process is then performed, wherein a structure obtained by placing the sewing object on the sewing material is moved along a preset sewing path by means of a program.

[0003] To position and hold the sewing object, in such prior art sewing devices, the outer fabric transport frame is equipped with a fabric holding frame in which openings, such as grooves or holes, are formed that correspond to sewing locations of the sewing object. The sewing object is then sewn along the openings formed in the fabric holding frame. Furthermore, sewing can be performed along the outer edge of the sewing object having a predetermined shape while maintaining a given distance from an outer edge of the sewing object.

[0004] However, during the cutting process of the sewing object to a predetermined shape, variations in size or shape from the intended shape may occur due to differences in the stretchability of the sewing objects to be sewn, differences in the skills of the workers, or the like. In this case, it is not possible to perform a sewing operation on such a sewing object while maintaining a given distance from an actual outer edge.

[0005] As a sewing machine that solves such a problem, JP H07-28 962 B2 discloses an electronic sewing machine in which "a line position detecting device comprises a scanning camera suspended on a rotary bearing, the bearing being arranged on a sewing machine head, and the camera is rotated coaxially with the axis of a needle, and a line touching an imaginary circle around a needle insertion point within a sewing area is scanned."That is, JP H07-28 962 B2 describes that an outer peripheral edge of a sewing object, which is attached to a sewing material, is highlighted with respect to the sewing material by means of an illuminating device arranged on the scanning camera, that the outer peripheral edge of the sewing object is detected by the scanning camera, and that a sewing operation is performed according to a preprogrammed sewing pattern at sewing edge positions with a given width from the outer peripheral edge of the sewing object, while a sewing path is corrected according to detected actual positions of the sewing object.

[0006] Furthermore, as a sewing machine that solves the above-mentioned problem, JP H05-225315 A discloses a sewing device that performs a sewing operation as follows. That is, in this sewing device, external shape recognition of a sewing object is performed by external shape recognition means as a step preceding a sewing step; an outer peripheral edge of the sewing object, which is attached to a sewing material, is highlighted with respect to the sewing material by means of an illuminator provided on an outer peripheral edge side of the sewing object; the outer peripheral edge of the sewing object is detected by a CCD camera; the sewing object is transferred from the external shape recognition means to an electronic sewing machine; and sewing is performed at sewing edge positions a given width from the outer peripheral edge of the sewing object based on a detected external shape of the sewing object.

[0007] In both the sewing devices disclosed in JP H07-28962 B2 and JP H05-225315 A, the outer peripheral edge of the sewing object with respect to the sewing material is highlighted using the illumination device, and the outer peripheral edge of the sewing object is detected using the camera. Therefore, depending on the color or material of a sewing object, an outer peripheral edge thereof could not be sufficiently highlighted, and the outer peripheral edge could not be detected. In particular, when the color of a sewing object and the color of a sewing material are the same color (black), in many cases an outer peripheral edge of the sewing object could not be sufficiently highlighted, and detection of the outer peripheral edge became difficult.

[0008] JP 2007-263 852 A discloses a device, a method, and a processing program for detecting a defect. JP 2014-091 008 A discloses a sewing device. JP H06-98 331 A discloses a recognition device for fabric patterns and shapes. DE 30 43 852 A1 discloses a method for detecting signal deviations using an integrating differential amplifier. DE 41 01 346 C1 discloses a device for optically scanning the fabric in a sewing machine. SUMMARY OF THE INVENTION

[0009] The object of the invention is therefore to provide excellent detection of an outer peripheral end of a sewing object regardless of color or material of the sewing object with respect to the above-mentioned problems, the invention comprising the elements (1) to (8).

[0010] (1) In a shape recognition device comprising an imaging means for imaging a first object and a second object from above, the two objects being arranged to be vertically overlapped on each other, and for recognizing an interface between the first object and the second object from acquired image data obtained via the imaging means, the imaging means focuses an upper surface of either the first object or the second object at a focal distance so that one of the first object and the second object is imaged at the focal distance and the other of the first object and the second object is imaged at a non-focal distance.The shape recognition device comprises an image recognition control unit which, when luminance values ​​of respective pixels of acquired image data acquired by the imaging means are larger than luminance values ​​of surrounding pixels, performs sharpness processing of converting the luminance values ​​of the respective pixels of the acquired image data into larger luminance values ​​according to difference values ​​between the luminance values, and after the sharpness processing, performs binarization processing of the acquired image data and performs edge extraction to detect a boundary between the first object and the second object.

[0011] (2) In the shape recognition device comprising the above-mentioned element (1), the image recognition control unit performs smoothing processing of the acquired image data after the sharpness processing and before the binarization processing.

[0012] (3) The shape recognition device comprising the above-mentioned elements (1) or (2) further comprises a height adjusting means for raising or lowering the imaging means, and the image recognition control unit controls the height adjusting means such that the imaging means is at a preset height.

[0013] (4) The shape recognition device comprising any one of the above-mentioned elements (1) to (3) further comprises an illumination device capable of illuminating the first object and the second object from a plurality of illumination positions, and capable of adjusting the illumination intensity of each of the plurality of illumination positions.

[0014] (5) In a sewing machine comprising the shape recognizing device having any one of the above-mentioned elements (1) to (4), an external shape of a sewing object is recognized as the first object and the second object, and a sewing operation is carried out.

[0015] (6) In the sewing machine comprising the above-mentioned element (5), sewing pattern data is generated on the basis of the external shape of the sewing object obtained by the shape recognition device, and a sewing operation is carried out according to the sewing pattern data.

[0016] (7) In the sewing machine comprising the above-mentioned element (5), existing sewing pattern data is corrected on the basis of the external shape of the sewing object obtained by means of the shape recognition device, and sewing is carried out according to the sewing pattern data.

[0017] (8) In the sewing machine comprising the above-mentioned element (7), the external shape of the sewing object obtained by the shape recognition device and the existing sewing pattern data are compared with each other and the presence / absence of sewing operation execution is made on the basis of comparison results.

[0018] According to the invention, when focusing and imaging the upper surfaces of the first object and the second object, which are arranged to be vertically overlapped, an image of an object located within a focal distance becomes more vivid, while an image of an object located outside the focal distance becomes blurred. Therefore, when sharpness processing is performed on acquired image data of an object, the distributions of luminance values ​​on the upper surface of one object and the upper surface of the other object can be highly separated.

[0019] For this reason, an edge, which represents an interface region between the top surfaces of one object and the other object, can be easily and clearly extracted.

[0020] Accordingly, excellent detection of an object's interface area is possible regardless of the object's color or material. BRIEF DESCRIPTION OF THE CHARACTERS Fig. 1 is a perspective view of a sewing device related to an embodiment of the invention. Fig. Figure 2 is a front view of the sewing device according to the embodiment of the invention. A camera is shown separately. Fig. 3 is a control block diagram of the sewing device related to the embodiment of the invention. Fig. 4 is a perspective view of a fabric transport unit of the sewing device related to the embodiment of the invention. Fig. 5 is a flowchart showing the control and processing performed by an image recognition control unit in order. Fig. 6 shows a captured image of a sewing object. Fig. 7 shows a captured image of the sewing object after sharpening processing. Fig. 8 shows a captured image of the sewing object after smoothing processing. Fig. 9 is a histogram showing the frequency distribution of luminance values ​​based on the luminance values ​​of all pixels of acquired image data subjected to smoothing processing. Fig. 10 shows a captured image of the sewing object after binarization processing. Fig. 11(A) to 11(C) are explanatory diagrams showing examples of non-coincidence of sewing patterns. Fig. 12(A) shows an extraction of characteristic points and Fig. Fig. 12(B) is an explanatory diagram showing a state in which the positions of corresponding characteristic points are compared. Fig. 13(A) shows a state where the deviation amounts of sewing patterns are obtained, and Fig. Fig. 13(B) is an explanatory diagram showing a state where translation of the sewing patterns is performed based on the deviation amounts. Fig. Fig. 14(A) shows a detected image in a state where a work body is in focus, and Fig. 14(B) shows a detected image in a state where an area between the work material and the work body is in focus. Fig. 15 is a histogram showing the frequency distribution of luminance values ​​based on the luminance values ​​of all pixels of acquired image data subjected to smoothing processing when two sheet materials of sewing material are overlapped on a sewing material body. DETAILED DESCRIPTION

[0021] An embodiment of the invention will be described below with reference to the drawings. Furthermore, the embodiment to be described below does not limit the technical scope of the invention. [Sewing machine outline]

[0022] As in the Fig. 1 to 3, a sewing machine 1 according to the present invention comprises a sewing machine body 2, a fabric feeding unit 3, a shape recognition unit 4 serving as a shape recognition device, a sewing machine control unit 5, a sewing pattern database 51, and a display device 52.

[0023] The sewing machine body 2 includes a needle vertical movement mechanism (not shown) installed in the sewing machine arm part 2a to perform a vertical movement operation of a sewing needle, a shuttle mechanism installed in a sewing machine bed 2b to intertwine a needle thread and a bobbin thread, and a sewing machine table 2c on which a sewing object is placed and performs a sewing operation on the sewing object.

[0024] The shape recognition unit 4 includes a camera 41 serving as an imaging means for imaging a sewing object serving as a first object and a second object, an illumination device 42 for illuminating the imaged sewing object, an image recognition control unit 43 for performing various types of processing (to be described below) on image data acquired by the camera 41, a setting input unit 44 for setting the shape recognition of a sewing object and operating the sewing machine body, and a data storage unit 45.

[0025] The image recognition control unit 43 controls the camera 41 and the illumination device 42 to acquire captured image data and performs predetermined image processing (to be described later) to perform shape recognition of a sewing object.

[0026] The image recognition control unit 43, the sewing machine control unit 5, the setting input unit 44, the data storage unit 45, and the sewing pattern database 51 constitute an information processing system composed of information processing means, information input means, and information storage means. Furthermore, there are no particular restrictions on whether the image recognition control unit 43 and the sewing machine control unit 5 are operated using the same IC chip or separate IC chips, whether the data storage unit 45 and the sewing pattern database 51 are constructed from the same hard disk or the same IC chip or separate hard disks or IC chips, and whether the image recognition control unit 43 and the sewing machine control unit 5 are provided within a casing of the sewing machine body 2 or as an external device.

[0027] The display device 52 optionally performs an audio display in addition to an image display.

[0028] The illumination device 42 comprises four (or more) light sources arranged horizontally on four sides with the camera 41 as the center. The respective light sources are configured to illuminate an underlying sewing object from the aforementioned four illumination positions and to individually adjust the illumination intensity according to the input from the setting input unit 44.

[0029] That is, as described below, when a sewing object is formed by a work piece C2 and a work piece C1 placed on the work piece, a shadow may be formed along an outer edge of the work piece C1. Since this shadow may impair the image processing described below, the shadow can be thinned by adjusting the illumination intensity of the light sources on four sides. Since the shadow is created on a side opposite the light sources, adjustments such as lowering the illumination intensity on the side opposite the one on which the shadow is created and increasing the illumination intensity on the shadow side are performed.

[0030] The camera 41 consists mainly of an imaging element and an optical system, which may be, for example, a CCD (Charged Coupled Device) and a CMOS (Complementary Metal Oxide Semiconductor) image sensor.

[0031] The optical system of the camera 41 uses an optical system that has a single focal point and a sufficiently shallow depth of field. The depth of field is, for example, such that a captured image is blurred and out of focus even if a deviation corresponding to the thickness of the sewing material C1, which is a sewing object, has occurred in the direction of the optical axis.

[0032] Furthermore, a height adjustment means 46 is provided with the camera 41 for selectively raising and lowering the camera 41 along a vertical direction (Z direction), which represents the direction of the optical axis, by means of a motor drive. The image recognition control unit 43 controls the height adjustment means 46 to perform height adjustment such that the camera 41 reaches a preset height.

[0033] As in Fig. 4, the fabric feeding unit 3 is configured to include a fabric feeding lower plate 7, a fabric feeding outer frame 8, a fabric feeding base 10, a fabric feeding arm 11, and an XY drive part (installed in the sewing machine bed 2b).

[0034] The fabric feed outer frame 8 is pivotally mounted on the fabric feed arm 11 about an X-axis, and the fabric feed arm 11 is vertically movable on the fabric feed base 10. Furthermore, the fabric feed lower plate 7 is fixed to the fabric feed base 10 below the fabric feed outer frame 8.

[0035] The fabric feed base 10 is configured to be operable within a horizontal plane (an XY plane in the drawing) by means of an XY drive part including a motor, a timing belt, a timing pulley, and the like, all not shown. The fabric feed lower plate 7 and the fabric feed outer frame 8 are integrated together with the fabric feed base 10, can be moved and positioned arbitrarily within the range of movement within the horizontal plane. Sewing positions are defined by two-dimensional XY coordinates related to the control of this fabric feed unit 3. The sewing positions of each sewing object are specified by sewing pattern data associated with each sewing object. Each sewing pattern data is stored in the sewing pattern database 51.

[0036] A sewing object is configured such that the work cloth C1, which is a first object having a predetermined outer peripheral shape, and the work cloth body C2, which is a second object to which the work cloth C1 is sewn from above, are arranged to be vertically overlapped with each other, and the work cloth conveying outer frame 8 holds only the work cloth body C2 between the work cloth conveying outer frame and the lower plate 7 during a lowering operation. In the sewing pattern data, patterns for forming a seam inside the work cloth C1 while maintaining a given distance from an outer peripheral edge of the work cloth C1 are recorded. That is, patterns for forming a seam in a similar shape whose size is slightly smaller than the outer peripheral shape of the work cloth C1 are recorded.

[0037] Since the work fabric C1 is not held by the work fabric feed outer frame 8, the work fabric is temporarily fixed to the work fabric body C2 by gluing. The temporary fixing to the work fabric body C2 can also be performed by other methods in which the outer periphery of the work fabric C1, as viewed from above, is not covered. For example, the work fabric C1 may be held by a magnet located within a seam formation position on an upper surface of the work fabric C1 (or it may be held by a magnetic body from above using the lower plate 81 as a magnet). [Sewing procedure]

[0038] The sewing operation control performed by the sewing machine control unit 5 will now be described.

[0039] When a worker sets the fabric C1 and the fabric body C2 between the fabric feeding lower plate 7 and the fabric feeding outer frame 8 and the sewing start is input from the setting input unit 44, the sewing machine control unit 5 lowers the fabric feeding outer frame 8 to hold the fabric body C2 and performs an operation of reading out sewing pattern data from the sewing pattern database.

[0040] Furthermore, while the sewing start is inputted and a sewing operation is actually started in the respective parts (units) of the sewing machine body 2, the shape recognition unit 4 acquires captured image data of the sewing material C1, performs shape recognition, and performs quality determination of the shape of the sewing material C1 or comparison collation, correction processing, or the like of the sewing pattern data.

[0041] The sewing machine control unit 5 then starts a vertical movement operation of a sewing needle, synchronously controls the XY drive section according to the sewing pattern data, moves the fabric feed lower plate 7 and the fabric feed outer frame 8 to predetermined target positions per stitch, and performs needle movement on the fabric C1 in accordance with a sewing pattern according to an outer peripheral shape. Then, sewing is performed along the outer peripheral shape of the fabric C1 by sewing at all sewing positions specified in the sewing pattern data. [Processing performed by the shape recognition unit]

[0042] Next, a description will be given of the various types of processing performed by the image recognition control unit 43 of the shape recognition unit 4 after the sewing start input has been made. Fig. Fig. 5 is a flowchart showing the control and processing performed by the image recognition control unit 43 in order.

[0043] Furthermore, the fabric feed lower plate 7 and the fabric feed outer frame 8, which holds a sewing object, are moved in advance to a prescribed imaging position via the control of the sewing machine control unit 5.

[0044] After this movement, the image recognition control unit 43 controls the height adjustment means 46 to adjust the camera 41 to a proper height (step S1). The proper height of the camera 41 is a height at which the focal point of the optical system is aligned with the upper surface of the work fabric C1, and its value is input in advance by the setting input unit 44.

[0045] The image recognition control unit 43 then controls the camera 41 to image the entire sewing material C1 from above (step S3). The image recognition control unit 43 generates captured image data consisting of regions of the sewing material C1 and the sewing material body C2 using the imaging of the camera 41 and stores the generated data in the data storage unit 45.

[0046] Furthermore, Fig. 6 shows a captured image, where reference symbol t1 represents an upper surface of the work fabric C1 and reference symbol t2 represents an upper surface of the work fabric body C2. Because imaging is performed by the camera 41 having an optical system with a shallow depth of field, it can be understood that the upper surface of the work fabric C1 located at a focal distance is vividly imaged, while the upper surface of the work fabric body C2 imaged at a non-focal distance deviating from the focal point is blurred.

[0047] Next, the image recognition control unit 43 performs sharpness processing on the acquired captured image data (step S5). That is, the image recognition control unit 43 functions as a "sharpness processing means." This sharpness processing uses an eight-way Laplacian filter. Accordingly, the luminance values ​​of the respective pixels constituting the captured image data are converted such that the luminance value differences become larger with respect to pixels with larger luminance values ​​than the surrounding pixels (e.g., neighboring pixels).

[0048] So if the sharpness processing is carried out as in Fig. As shown in Figure 7, pixels with smaller luminance values ​​in a low-sharpness region t2 of the fabric body C2 increase in size and become black as a whole. Furthermore, pixels with small luminance values ​​in a high-sharpness region t1 of the fabric C1 decrease in size.

[0049] Furthermore, the image recognition control unit 43 performs smoothing processing on the acquired image data that has undergone sharpness processing (step S7). A 15x15 mean filter is used for this smoothing processing.

[0050] Accordingly, as in Fig. 8, noise in the entire image is removed, and in the area t1 of the sewing material C1 where numerous pixels with large luminance values ​​are included, the luminance values ​​of pixels with small luminance values ​​that appear dotted are also raised, and the luminance values ​​of the entire area t1 are raised.

[0051] Next, the image recognition control unit 43 performs edge extraction processing to obtain an interface between the work cloth C1 and the work cloth body C2 (step S9).

[0052] In this edge extraction processing, due to the need to clearly distinguish the area t1 of the sewing material C1 and the area t2 of the sewing material body C2, a threshold value of the luminance values ​​for this distinction must first be determined.

[0053] For this reason, the image recognition control unit 43 obtains the frequency distribution of the luminance values ​​based on the luminance values ​​of all pixels of the acquired image data that have undergone smoothing processing. As a result, as shown in the histogram of Fig. 9, the distribution b1 of the luminance values ​​of respective pixels of the area t1 and the distribution b2 of the luminance values ​​of respective pixels of the area t2 are generated. With respect to the respective distributions b1 and b2, luminance values ​​at which the frequency distribution reaches peak values ​​are obtained, and an intermediate value k1 of these luminance values ​​becomes the threshold value of the luminance values.

[0054] Furthermore, as in Fig. As illustrated in Fig. 10, the image recognition control unit 43 performs binarization processing of the acquired image data which has undergone the smoothing processing based on the above-mentioned threshold value.

[0055] Next, edge detection of the area t1 of the work fabric C1 and the area t2 of the work fabric C2 is performed using a well-known edge detection operator, and all pixels acting as the boundary between the work fabric C1 and the work fabric C2 are specified. Accordingly, the outer peripheral edge (edge ​​e) of the work fabric C1 is detected, and the image recognition control unit 43 functions as a "detection unit."

[0056] Next, the image recognition control unit 43 generates a sewing pattern from all edge points constituting the outer peripheral edge of the sewing material C1 acquired from the captured image data.

[0057] That is, the image recognition control unit 43 generates a spline curve from all edge points (some edge points may also be thinned out), obtains the outer shape of the fabric C1, and generates a pattern serving as a sewing position trace formed within the fabric with a given sewing margin to the outer shape of the fabric. Then, a sewing pattern consisting of all sewing positions formed at a given sewing pitch along the pattern is generated, which serves as the sewing position trace (referred to as an "image-based sewing pattern").

[0058] Then, the image recognition control unit 43 compares this image-based sewing pattern with the sewing pattern defined in the sewing pattern data (step S11).

[0059] For the comparison between the sewing patterns, the image recognition control unit 43 extracts characteristic points P1 and P2 with respect to the image-based sewing pattern and a sewing pattern defined in the sewing pattern data ( Fig. 12(A)) and compares the positions of the corresponding characteristic points P1 and P2 ( Fig. 12(B)).

[0060] When the respective characteristic points are approached to each other within a range in which the characteristic points are considered to coincide with each other, the shape patterns are determined to coincide with each other, the image recognition control unit 43 notifies the sewing machine control unit 5 of permission to start sewing, and the sewing machine control unit 5 performs a sewing operation based on the sewing pattern data of the data storage unit 45.

[0061] On the other hand, if the shape patterns are determined not to coincide with each other, the image recognition control unit 43 corrects the sewing pattern data of the data storage unit 45 (step S13). Fig. Figure 11 shows examples of non-coincidence of suture patterns; imaging-based suture patterns are shown with dashed lines and suture patterns defined in the suture pattern data are shown with solid lines.

[0062] In Fig. 11(A) the shapes coincide with each other, but a positional deviation occurs as a whole. In Fig. 11B there is non-coincidence in the shapes of the sewing patterns themselves. In Fig. 11(C) the shapes coincide with each other, but one of them has a tilt.

[0063] For example, if there is a deviation in the overall positions of a sewing pattern, as in Fig. 11(A), the total amount of positional deviation is obtained as shown in Fig. 13(A), and the translation of the entire sewing pattern defined in the sewing pattern data is corrected by the amount of positional deviation as shown in Fig. 13(B).

[0064] Further: if a sewing pattern extends long in a given direction, as shown in Fig. 11B, its elongation is obtained, and coordinate conversion is performed so that the sewing pattern based on the sewing pattern data is elongated in the given direction. Otherwise, the sewing pattern defined in the sewing pattern data can be changed to the image-based sewing pattern.

[0065] Further: if a sewing pattern has a slope, as shown in Fig. As shown in Fig. 11(C), the inclination angle of the sewing pattern is obtained, and coordinate conversion is performed such that the sewing pattern based on the sewing pattern data is rotated by the inclination angle.

[0066] Furthermore, in cases of Fig. 11(A) and Fig. 11(C) the sewing patterns defined in the sewing pattern data are changed to the imaging-based sewing patterns.

[0067] Then, when the correction of the sewing pattern data of the data storage unit 45 is completed, the image recognition control unit 43 notifies the sewing machine control unit 5 of permission to start sewing, and the sewing machine control unit 5 performs a sewing operation based on the sewing pattern data of the data storage unit 45. [Technical effects of the embodiment of the invention]

[0068] When one of the work pieces, work piece C1 and work piece body C2, having different heights (for example, work piece C1) is focused by the camera 41 and either the work piece or the work piece body is imaged, the image of the upper surface of work piece C1 becomes more vivid and the upper surface of work piece body C2 having a further focal distance becomes blurred.

[0069] Therefore, the image recognition control unit 43 of the sewing machine 1 can very well separate the distribution of the luminance values ​​on the upper surface of the sewing material C1 and on the upper surface of the sewing material body C2 by performing sharpness processing of these surfaces.

[0070] For this reason, the interface between the fabric C1 and the fabric body C2 can be easily and clearly extracted as an edge.

[0071] Accordingly, the image recognition control unit 43 of the sewing machine 1 is capable of excellently recognizing the outer peripheral end of the sewing material C1, regardless of the color or material of the outer peripheral edge of the sewing material C1.

[0072] Further, accordingly, it is possible to ensure a given sewing margin on the outer peripheral edge of an actual sewing material C1 and to perform a sewing operation with more precision, and it is possible to improve the sewing quality.

[0073] Furthermore, the image recognition control unit 43 of the sewing machine 1 performs smoothing processing between the sharpness processing and the edge extraction processing (binarization processing). Therefore, by averaging the luminance values ​​of the pixels with those of the surrounding pixels, the influence of pixels acting as noise can be reduced, and the fabric area C1 and the fabric body area C2 can be separated from each other with greater precision.

[0074] Furthermore, since the sewing machine 1 is capable of illuminating the work material C1 from a plurality of illumination positions and includes the illumination device 42 capable of adjusting the illumination intensity of each light source at each illumination position, the adjustment can be made so as to suppress the formation of a shadow by the outer peripheral edge of the work material C1, and it is possible to suppress the above-mentioned influence to detect the outer peripheral edge of the work material C1 with more precision.

[0075] Furthermore, the sewing machine 1 corrects existing sewing pattern data based on the shape of the outer peripheral edge of the fabric C1 obtained by the shape recognition unit 4, and the sewing machine control unit 5 performs control to perform a sewing operation according to the sewing pattern data.

[0076] For this reason, even if a deviation from the sewing patterns of the existing sewing pattern data occurs due to the individual differences in cutting the sewing material C1 or the individual differences in attaching the sewing material C1, sewing can be performed with correspondingly more precision, and it is possible to improve the sewing quality. [Other]

[0077] A case was illustrated in which the height of the camera 41 is adjusted such that the sewing material C1 is focused in the above-mentioned imaging of the sewing material C1 and the sewing material body C2. However, the invention is not limited to this, as long as a discernible difference in sharpness is obtained between the sewing material C1 and the sewing material body C2, which forms the background of the sewing material.

[0078] For example, the upper surface of the workpiece body C2 can be focused, as shown in Fig. 14(A).

[0079] Further: if the focal distance of the camera 41 is too long, as shown in Fig. As illustrated in Figure 14(B), a discernible difference in sharpness does not simply result from a difference in the thickness of the fabric C1. Thus, it is better if the focal distance of the camera 41 is shorter.

[0080] Further, when the portion of the work material C1 is reflected, it is desirable to adjust the illumination intensity of the respective light sources of the illumination device 42 or to focus a position slightly above the upper surface of the work material C1 so that no shadow of the portion is formed.

[0081] Further, when a difference is caused predominantly in the contrast of a captured image due to a difference in color or material between the fabric C1 and the fabric body C2, it is desirable to focus on either the fabric or the fabric body where the contrast tends to become larger.

[0082] Furthermore, a case was presented in which the eight-directional Laplacian filter is used in the sharpening processing mentioned above. However, other sharpening filters or edge enhancement filters can also be used. Furthermore, Laplacian filters with fewer or more directions can be used.

[0083] Furthermore, a case was presented in which a 15x15 mean filter is used in the smoothing processing; however, other mean filters can also be used. Furthermore, the averaging range can be enlarged or reduced as appropriate. When averaging is performed, weighting can also be performed, for example, where pixels are given greater weight.

[0084] Furthermore, the shape recognition unit 4 is capable of comparing the shape of the outer peripheral edge of the sewing material C1 to be acquired by imaging with the existing sewing pattern data, and the presence / absence of a sewing operation execution can be determined based on the comparison results.

[0085] For example, the image recognition control unit 43 can perform error handling based on a mutual non-coincidence situation when comparing the image-based sewing pattern with the sewing patterns defined in the sewing pattern data. In particular, if the sizes of the respective sewing patterns differ significantly from each other or if the number of characteristic points does not match, a sewing prohibition can be reported to the sewing machine control unit 5. The notification can be made in such a way that a worker is prompted to confirm the same.

[0086] Accordingly, the worker can be encouraged to confirm when notified of the occurrence of a certain abnormality. Therefore, it is possible to suppress the oversight of the occurrence of a certain abnormality.

[0087] Furthermore, the above embodiment assumes that the number of times the sewing material C1 is imaged by the camera 41 is one. However, it is possible to perform imaging a plurality of times by changing a focal position from the sewing material C1 to the sewing material body C2, and it is possible to acquire a plurality of pieces of acquired image data. Furthermore, it is possible to acquire sewing patterns for each acquired image data and synthesize them into a unit by averaging their coordinate values ​​to improve accuracy, and to perform comparison with the sewing patterns of the sewing pattern data in the data storage unit 45.

[0088] Furthermore, although in the above embodiment, the sewing object is shown to be composed of two sheet materials, namely an upper sheet material and a lower sheet material, by overlapping the sewing material C1 on the sewing material body C2, the number of sheet materials to be overlapped can be further increased.

[0089] For example, in a sewing pattern relating to a sewing object consisting of three surface materials, which are upper and lower surface materials, by additionally overlapping a smaller sewing material on the sewing material C1, upper surfaces of the three respective surface materials can be identified by focusing on the upper surface of the lowest or the uppermost sewing material part. In this case, three concentration locations of the distribution are generated in a histogram, as shown in Fig.15. Thus, luminance values ​​at which the respective frequency distributions reach peak values ​​are obtained, and interfaces between the respective workpieces are extracted as edges, with intermediate values ​​of the luminance values ​​as thresholds k1 and k2. Accordingly, it is possible to acquire sewing patterns based on the outer peripheral edges of two upper workpieces.

[0090] Furthermore, at the shape recognition unit 4, a sewing pattern defined in the sewing pattern data is corrected based on an image-based sewing pattern. However, the sewing machine control unit 5 can control the respective parts (units) of the sewing machine to generate sewing pattern data directly from the image-based sewing pattern, and can perform sewing according to the sewing pattern data without performing such update processing. In this case, by eliminating the need for correction processing of the existing sewing pattern data, it is possible to achieve simplification and acceleration of the processing.

[0091] Furthermore, although the above embodiment shows a case where the shape recognition device (shape recognition unit 4) is applied to the sewing machine 1, the shape recognition device can also be applied to devices in any field where the shape of objects needs to be recognized, without being limited thereto.

Claims

[1] Shape recognition device (4), comprising an imaging means (41) for imaging a first object and a second object from above, the two objects being arranged so as to be vertically overlapped on one another, and which detects a boundary surface between the first object and the second object from acquired image data obtained via the imaging means (41), wherein the imaging means (41) focuses an upper surface of either the first object or the second object at a focal distance such that one of the first object and the second object is imaged at the focal distance and the other of the first object and the second object is imaged at a non-focal distance, wherein the shape recognition device (4) comprises: an image recognition control unit (43) which, when luminance values ​​of respective pixels of acquired image data acquired by the imaging means (41) are larger than luminance values ​​of surrounding pixels, performs sharpness processing of converting the luminance values ​​of the respective pixels of the acquired image data into larger luminance values ​​according to difference values ​​between the luminance values, and after the sharpness processing, performs binarization processing of the acquired image data and performs edge extraction to detect a boundary between the first object and the second object. [2] A shape recognition device according to claim 1, wherein the image recognition control unit (43) performs smoothing processing of the acquired image data after the sharpness processing and before the binarization processing. [3] Shape recognition device according to claim 1 or 2, further comprising: a height adjustment means (46) for raising or lowering the imaging means (41), wherein the image recognition control unit (43) controls the height adjustment means (46) such that the imaging means (41) is at a preset height. [4] Shape recognition device according to one of claims 1 to 3, further comprising: an illumination device (42) capable of illuminating the first object and the second object from a plurality of illumination positions, and capable of adjusting the illumination intensity of each of the plurality of illumination positions. [5] Sewing machine (1), comprising: the shape recognition device (4) according to one of claims 1 to 4, and Recognizing an external shape of a sewing object as the first object and the second object and performing a sewing operation. [6] A sewing machine according to claim 5, wherein sewing pattern data is generated on the basis of the external shape of the sewing object obtained by the shape recognition device (4), and a sewing operation is performed according to the sewing pattern data. [7] A sewing machine according to claim 5, wherein existing sewing pattern data is corrected on the basis of the external shape of the sewing object obtained by means of the shape recognition device (4), and a sewing operation is performed according to the sewing pattern data. [8] A sewing machine according to claim 7, wherein the outer shape of the sewing object obtained by the shape recognition device (4) and the existing sewing pattern data are compared with each other and the presence / absence of sewing operation execution is determined on the basis of comparison results.

Images