JP2016063894A - Shape recognition device and sewing machine - Google Patents

Abstract

PROBLEM TO BE SOLVED: To excellently recognize the outer peripheral end part of an object to be sewn regardless of the color and the material of the object to be sewn.SOLUTION: A shape recognition device 4 is provided for, from captured image data by image capturing means 41 which captures, from above, the image of a first object and a second object vertically overlapped, recognizing a border between the objects. In the shape recognition device 4, the image capturing means captures the image while positioning the upper surface of the first object or the second object at a focal distance. In addition, the shape recognition device 4 includes an image recognition control section 43 for performing: sharpening processing for the captured image data captured by the image capturing means in which, with respect to each pixel of the captured image data, when the pixel has a brightness larger than the brightness of surrounding pixels, the brightness value of the pixel is changed to a larger value according to the difference value therebetween; binarization processing for the captured image data after the sharpening processing; and edge extraction to recognize the border between the first object and the second object.SELECTED DRAWING: Figure 6

Description

  The present invention relates to a shape recognition device for recognizing the shape of a sewing object and a sewing machine including the shape recognition device.

Conventionally, the sewing machine body is sandwiched between a cloth feed lower plate and a cloth feed outer frame which are configured to be operable in a horizontal plane, and the work to be sewn is installed on the sewing machine body, and this is followed along a pre-programmed sewing path. 2. Description of the Related Art Sewing apparatuses that sew a workpiece to be sewn by sewing while moving the sewing machine are used.
In such a sewing apparatus, in order to position and hold the sewing product, the cloth feed outer frame is equipped with a cloth holding frame in which openings such as grooves and holes corresponding to the sewing locations of the sewing product are formed. . Then, the sewing object is sewn along an opening formed in the cloth holding frame. Further, there is a case where sewing is performed while maintaining a constant distance from the outer edge along the outer edge of the sewing object having a predetermined shape.
However, at the stage of cutting the sewing product into a predetermined shape, the size and shape may vary from the target shape due to the difference in stretchability of the sewing product and the skill of the operator. For such a material to be sewn, it was not possible to perform sewing while maintaining a constant interval with respect to the actual outer edge.

  In order to solve such a problem, Patent Document 1 discloses that “the edge position sensing device has a scanning camera mounted on a rotary bearing, the bearing is mounted on a sewing head, and the camera is coaxial with the axis of the needle. An electronic sewing machine is disclosed that scans a line that touches a virtual circle around the needle penetration point in the sewing surface. That is, in this patent document 1, the outer peripheral edge of the sewing product is emphasized with respect to the sewing product body by the illumination device mounted on the scanning camera, and the sewing product is fixed by the scanning camera. Detecting the outer peripheral edge of the sewing product and correcting a sewing path in accordance with the detected actual position of the sewing product with respect to a pre-programmed sewing pattern, a sewing margin having a certain width from the outer peripheral edge of the sewing product. A sewing device for sewing positions is described.

  Further, in order to solve this problem, Patent Document 2 discloses that the outer shape of the sewing product is recognized by the outer shape recognition device as a pre-process of the sewing process, and the sewing product fixed to the sewing product body is the outer periphery of the sewing product. The outer peripheral edge of the sewing product is emphasized with respect to the sewing product body by the illumination device provided in the edge measuring unit, the outer peripheral edge of the sewing product is detected by the CCD camera, and the electronic sewing machine detects the outer peripheral edge of the sewing product. A sewing device is described in which a sewing product is transferred to the sewing machine and a sewing margin position having a predetermined width is sewn from the outer peripheral edge of the sewing product based on the detected outer shape of the sewing product.

Japanese Patent Publication No. 7-28962 JP 05-225315 A

However, in any of the sewing devices described in Patent Documents 1 and 2 described above, the illumination device emphasizes the outer peripheral edge of the sewing product with respect to the sewing product body, and the camera recognizes the outer peripheral edge of the sewing product. For this reason, depending on the color and material of the sewing object, the outer peripheral edge may not be sufficiently emphasized, which may make recognition difficult. In particular, when the color of the sewing object and the sewing body is the same color (black), the outer peripheral edge tends not to be sufficiently emphasized.
As a result, it has been difficult to accurately recognize the outer peripheral edge of the workpiece.

  Then, this invention makes it a subject to recognize the outer peripheral edge part favorably irrespective of the color and material of to-be-sewn material.

The invention described in claim 1
An imaging means for imaging the first object and the second object that are arranged one above the other from above and below,
In a shape recognition apparatus for recognizing a boundary between the first object and the second object from imaged image data obtained by the imaging means,
The imaging means images one object at a focal distance and the other object at a non-focal distance by adjusting the upper surface of one of the first object and the second object to a focal distance,
A sharpening process for greatly converting the luminance value of the pixel according to the difference value when the luminance of each pixel of the captured image data is larger than the luminance of surrounding pixels with respect to the captured image data captured by the imaging unit; An image recognition control unit that performs binarization processing on the captured image data after the sharpening processing, and further performs edge extraction to recognize a boundary between the first object and the second object; It is characterized by providing.

The invention according to claim 2 is the shape recognition apparatus according to claim 1,
The image recognition control unit performs a smoothing process on captured image data after the sharpening process and before the binarization process.

The invention according to claim 3 is the shape recognition apparatus according to claim 1 or 2,
A height adjusting means for raising and lowering the imaging means;
The image recognition control unit controls the height adjusting unit so that the imaging unit has a preset height.

The invention according to claim 4 is the shape recognition device according to any one of claims 1 to 3,
An illumination device is provided that can irradiate the object from a plurality of irradiation positions and can adjust the illumination intensity for each of the plurality of irradiation positions.

According to a fifth aspect of the present invention, in a sewing machine comprising the shape recognition device according to any one of the first to fourth aspects,
Sewing is performed by recognizing the outer shape of the sewing object as the object.

The invention according to claim 6 is the sewing machine according to claim 5,
Generate sewing pattern data based on the outer shape of the sewing product obtained by the shape recognition device,
The needle movement is performed according to the sewing pattern data.

The invention according to claim 7 is the sewing machine according to claim 5,
Correcting the existing sewing pattern data based on the outer shape of the sewing product obtained by the shape recognition device;
The needle movement is performed according to the sewing pattern data.

The invention according to claim 8 is the sewing machine according to claim 7,
Compare the outline of the sewing product obtained by the shape recognition device with existing sewing pattern data,
It is characterized in that the presence or absence of execution of sewing is determined based on the comparison result.

According to the present invention, when the upper surface of one of the first and second objects that are arranged one above the other is imaged in accordance with the focal length, an image with a clearer focus is obtained with respect to the target with the focal length. Therefore, if the sharpening process is performed on the captured image data of the target object, the luminance value between the upper surface of one target object and the upper surface of the other target object is unclear. Distribution can be largely separated.
For this reason, the edge which is a boundary part of the upper surface of one target object and the other target object can be extracted easily and clearly.
Thereby, it becomes possible to recognize the boundary part well regardless of the color or material of the object.

It is a perspective view of the sewing apparatus which concerns on one Embodiment of this invention. It is a front view of the sewing device concerning one embodiment of the present invention. However, the camera is shown separately. It is a control block diagram of the sewing apparatus which concerns on one Embodiment of this invention. It is a perspective view of the cloth feed part of the sewing apparatus which concerns on one Embodiment of this invention. It is the flowchart which showed the control and process which an image recognition control part performs in order. It is a picked-up image of a sewing product. It is a picked-up image of the to-be-sewn product after a sharpening process. It is the picked-up image of the to-be-sewn product after the smoothing process. It is the histogram which showed frequency distribution of the luminance value based on the luminance value of all the pixels of the picked-up image data which performed the smoothing process. It is the picked-up image of the to-be-sewn product after a binarization process. FIG. 11A to FIG. 11C are explanatory diagrams showing examples of mismatch of the sewing patterns. FIG. 12A shows extraction of feature points, and FIG. 12B is an explanatory diagram showing a state in which the positions of corresponding feature points are compared. FIG. 13A shows a state in which the amount of deviation of the sewing pattern is obtained, and FIG. 13B is an explanatory diagram showing a state in which the sewing pattern is translated based on the amount of deviation. FIG. 14A shows a captured image in a state in which the cloth body is focused, and FIG. 14B shows a captured image in a state in which the focus is between the fabric and the fabric body. It is the histogram which showed the frequency distribution of the luminance value based on the luminance value of all the pixels of the picked-up image data which performed the smoothing process at the time of superimposing two sheets of cloth on the cloth main body.

  An embodiment of the present invention will be described below with reference to the drawings. The following is one embodiment of the present invention and does not limit the present invention.

[Outline of sewing machine]
As shown in FIGS. 1 to 3, the sewing machine 1 according to the present embodiment includes a sewing machine body 2, a cloth feeding unit 3, a shape recognition unit 4 as a shape recognition device, a sewing machine control unit 5, a sewing pattern database 51, and And a display device 52.
The sewing machine main body 2 includes a needle up / down movement mechanism (not shown) that is built in the sewing machine arm portion 2a and moves the sewing needle up and down, a hook mechanism that is built in the sewing machine bed portion 2b and entangles the upper thread and lower thread, and a workpiece to be sewn. The sewing machine table 2c is mounted, and the sewing product is sewn.

  The shape recognizing unit 4 has a camera 41 as an imaging unit that captures an image of a sewing object as an object, an illuminating device 42 that irradiates the image of the sewing object to be imaged, and image data captured by the camera 41. An image recognition control unit 43 that performs various processes to be described later, a setting input unit 44 that sets the shape recognition of the sewing object and the operation of the sewing machine body, and a data storage unit 45 are provided.

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 described later to recognize the shape of the sewing product.
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 are an information processing system including an information processing device, an information input device, and an information storage device. The recognition control unit 43 and the sewing machine control unit 5 operate on the same or different IC chips, the data storage unit 45 and the sewing pattern database 51 are constructed on the same or different hard disks or IC chips, or the sewing machine body 2 Any hardware configuration may be used, such as whether it is configured within the housing or as an external device.
The display device 52 performs audio display arbitrarily in addition to image display.

The illuminating device 42 includes four light sources arranged horizontally in four directions with the camera 41 as the center (more at most). Each light source irradiates the lower sewing product from the four irradiation positions, and can individually adjust the illumination intensity according to the input from the setting input unit 44.
That is, as will be described later, in the case where the article to be sewn is composed of the fabric main body C2 and the fabric C1 placed thereon, a shadow may occur along the outer edge of the fabric C1. Since this shadow may affect the image processing described later, the shadow can be made thinner by adjusting the illumination intensity of the four light sources. Since the shadow is generated on the side opposite to the light source, adjustments such as reducing the illumination intensity on the side opposite to the direction in which the shadow occurs and increasing the illumination intensity on the shadow side are performed.

The camera 41 is mainly composed of an image sensor such as a CCD (Charge Coupled Device) or a CMOS (Complementary Metal Oxide Semiconductor Image Sensor) and an optical system.
The optical system of the camera 41 has a single focal point and a sufficiently shallow depth of field. For example, the depth of field is such that the picked-up image is blurred and out of focus only by the deviation of the thickness of the fabric C1, which is the sewing product, in the optical axis direction.
Further, the camera 41 is provided with a height adjusting means 46 for arbitrarily raising and lowering the camera 41 along the vertical direction (Z direction) that is the optical axis direction by driving the motor. The image recognition control unit 43 controls the height adjusting unit 46 to perform height adjustment so that the camera 41 has a preset height.

As shown in FIG. 4, the cloth feed section 3 includes a cloth feed lower plate 7, a cloth feed outer frame 8, a cloth feed base 10, a cloth feed arm 11, an XY drive section (built in the sewing machine bed section 2b), It is configured with.
The cloth feed outer frame 8 is supported by the cloth feed arm 11 so as to be swingable around the X axis, and the cloth feed arm 11 is supported by the cloth feed base 10 so as to be movable up and down. The cloth feed lower plate 7 is fixed to the cloth feed base 10 below the cloth feed outer frame 8.
The cloth feed base 10 is configured to be operable in a horizontal plane (XY plane in the drawing) by an XY drive unit including a motor, a timing belt, a timing pulley, and the like (not shown). The cloth feed lower plate 7 and the cloth feed outer frame 8 are integrally moved and can be arbitrarily positioned by moving in the horizontal plane. The sewing position is defined by a two-dimensional coordinate XY related to the control of the cloth feeding unit 3. The sewing position of each sewing product is specified by sewing pattern data corresponding to each sewing product. Each sewing pattern data is stored in the sewing pattern database 51.

The to-be-sewn product has a configuration in which a fabric C1 which is a first object having a predetermined outer peripheral shape and a fabric body C2 which is a second object to which the cloth C1 is sewn from above are arranged in an up-and-down manner. Thus, the cloth feed outer frame 8 holds only the cloth main body C2 with the lower plate 7 by the lowering operation. In the sewing pattern data, a pattern for forming a seam is recorded on the inside while maintaining a constant distance from the outer peripheral edge of the fabric C1. That is, a pattern for forming a seam in a similar shape that is somewhat smaller than the outer peripheral shape of the fabric C1 is recorded.
Since the fabric C1 is not held by the fabric feed outer frame 8, it is temporarily fixed to the fabric main body C2 by gluing. Note that the cloth body C2 may be temporarily fixed by another method in which the outer periphery of the cloth C1 is not shielded when viewed from above. For example, it may be held by a magnet or the like placed on the upper surface of the fabric C1 and inside the seam formation position (or the lower plate 81 is held as a magnet from above).

[Flow of sewing operation]
The sewing operation control performed by the sewing machine controller 5 will be described.
When the operator sets the cloth C1 and the cloth body C2 between the cloth feed lower plate 7 and the cloth feed outer frame 8 and inputs the start of sewing from the setting input section 44, the sewing machine control section 5 The frame 8 is lowered to hold the fabric body C2, and the sewing pattern data is read from the sewing pattern database.
The shape recognition unit 4 acquires captured image data of the fabric C1, recognizes the shape, and performs the fabric recognition C1 until the sewing operation is actually started at each part of the sewing machine body 2 after the start of sewing is input. The shape is judged to be good or bad, sewing pattern data is compared and verified, correction processing, and the like are executed.

  The sewing machine control unit 5 starts the vertical movement operation of the sewing needle and controls the XY driving unit in accordance with the sewing pattern data in synchronization with this, and the cloth feed lower plate 7 and the cloth feed outer frame 8 for each stitch. Is moved to a predetermined target position, and the needle is moved according to the sewing pattern corresponding to the outer peripheral shape of the fabric C1. Then, sewing is performed along the outer peripheral shape of the fabric C1 by performing needle dropping at all the needle dropping positions determined in the sewing pattern data.

[Processing by shape recognition unit]
Next, various processes performed by the image recognition control unit 43 of the shape recognition unit 4 after sewing start input will be described. FIG. 5 is a flowchart showing the control and processing performed by the image recognition control unit 43 in order.
Note that the cloth feed lower plate 7 and the cloth feed outer frame 8 that hold the workpiece are moved in advance to a prescribed imaging position under the control of the sewing machine control unit 5.
After this movement, the image recognition control unit 43 controls the height adjusting means 46 to adjust the camera 41 to an appropriate height (step S1). The appropriate height of the camera 41 is a height at which the focal point of the optical system matches the upper surface of the fabric C1, and the value is input in advance by the setting input unit 44.
And the image recognition control part 43 controls the camera 41, and images the cloth C1 whole from upper direction (step S3). The image recognition control unit 43 generates captured image data including part of the fabric C <b> 1 and the fabric main body C <b> 2 by capturing with the camera 41 and stores the captured image data in the data storage unit 45.
Note that FIG. 6 is a captured image, where reference numeral t1 represents the upper surface of the fabric C1, and t2 represents the upper surface of the fabric body C2. Since the image is captured by the camera 41 having an optical system with a shallow depth of field, the upper surface of the fabric C1 that is the focal length is clearly imaged, and the upper surface of the fabric main body C2 that is imaged at a non-focal length shifted from the focal point is blurred. You can see it happening.

Next, the image recognition control unit 43 performs a sharpening process on the acquired captured image data (step S5). That is, the image recognition control unit 43 functions as “sharpening processing means”.
In this sharpening process, an 8-direction Laplacian filter is used. As a result, the luminance value of each pixel constituting the captured image data is converted so that the luminance value difference is larger for pixels having a larger luminance value than surrounding pixels (for example, adjacent pixels).
Therefore, as shown in FIG. 7, when the sharpening process is performed, in the region t2 of the fabric body C2 having a low sharpness, a large number of pixels having a small luminance value occupy a black color as a whole. In addition, in the region t1 of the fabric C1 having a high definition, the number of pixels having a small luminance value is reduced.

Further, the image recognition control unit 43 performs a smoothing process on the captured image data that has been subjected to the sharpening process (step S7). In this smoothing process, a 15 × 15 averaging filter is used.
As a result, as shown in FIG. 8, noise is removed from the entire image, and in the region t1 of the fabric C1 where many pixels having a large luminance value are included, the luminance value of pixels with small luminance values is also increased. As a result, the luminance value is increased in the entire area.

Next, the image recognition control unit 43 performs an edge extraction process for obtaining the boundary between the fabric C1 and the fabric body C2 (step S9).
In this edge extraction process, it is necessary to clearly separate the region t1 of the fabric C1 and the region t2 of the fabric body C2, and therefore, first, it is necessary to determine a threshold value of the luminance value for the separation.
For this reason, the image recognition control unit 43 obtains a frequency distribution of luminance values based on the luminance values of all the pixels of the captured image data subjected to the smoothing process. As a result, as shown in the histogram of FIG. 9, a luminance value distribution b1 of each pixel in the region t1 and a luminance value distribution b2 of each pixel in the region t2 are generated. For each of the distributions b1 and b2, the luminance value having the frequency distribution as a vertex is obtained, and the intermediate value k1 is set as the threshold value of the luminance value.
Further, the image recognition control unit 43 performs binarization processing of the captured image data that has been subjected to smoothing processing, as shown in FIG. 10, based on the threshold value.
Then, the edge detection of the region t1 of the fabric C1 and the region t2 of the fabric C2 is performed using a known edge detection operator, and all pixels serving as the boundary between the fabric C1 and the fabric C2 are specified. As a result, the outer peripheral edge (edge e) of the fabric C1 is recognized, and the image recognition control unit 43 functions as a “recognition unit”.

Next, the image recognition control part 43 produces | generates a sewing pattern from all the edge points which comprise the outer periphery part of the fabric C1 acquired from captured image data.
That is, the image recognition control unit 43 generates a spline curve from all edge points (or a part of the edge points may be thinned out), obtains the outer shape of the fabric C1, and further inward from the outer shape with a certain seam allowance. A pattern serving as a locus of the sewing position to be formed is generated. Then, a sewing pattern including all needle drop positions formed at a constant sewing pitch along the pattern serving as the locus of the sewing position is generated (referred to as “sewing pattern based on imaging”).
Then, the image recognition control unit 43 compares the sewing pattern based on this imaging with the sewing pattern defined in the sewing pattern data (step S11).

In order to compare the sewing patterns, the image recognition control unit 43 extracts feature points P1 and P2 for each of the sewing pattern based on the imaging and the sewing pattern defined in the sewing pattern data (FIG. 12A). The positions of the corresponding feature points P1 and P2 are compared (FIG. 12B).
When the feature points are close to each other within a range that can be regarded as matching, it is determined that the shape patterns match, and the image recognition control unit 43 notifies the sewing machine control unit 5 of permission to start sewing. The sewing machine control unit 5 executes a sewing operation based on the sewing pattern data stored in the data storage unit 45.

On the other hand, if it is determined that the shape patterns do not match, the image recognition control unit 43 corrects the sewing pattern data in the data storage unit 45 (step S13).
FIG. 11 shows an example of mismatching of the sewing patterns, in which the sewing pattern based on the imaging is indicated by a broken line, and the sewing pattern determined in the sewing pattern data is indicated by a solid line. In FIG. 11A, the shapes are identical to each other, but there is a positional shift as a whole. In FIG. 11B, there is a discrepancy in the shape of the sewing pattern itself. In FIG. 11C, the shapes coincide with each other, but one is inclined.
For example, as shown in FIG. 11A, when there is a deviation in the overall position of the sewing pattern, the entire positional deviation amount is obtained as shown in FIG. As shown in FIG. 4, correction is performed to translate the entire sewing pattern determined in the sewing pattern data by the amount of positional deviation.

Further, as shown in FIG. 11B, when the sewing pattern extends long in a certain direction, the elongation rate is obtained, and the sewing pattern based on the sewing pattern data is elongated along the certain direction. Perform coordinate transformation. Alternatively, the sewing pattern defined in the sewing pattern data may be updated to a sewing pattern based on imaging.
If the sewing pattern is inclined as shown in FIG. 11C, the inclination angle is obtained, and coordinate conversion is performed so that the sewing pattern based on the sewing pattern data is rotated by the inclination angle.
11A and 11C, the sewing pattern defined in the sewing pattern data may be updated to a sewing pattern based on imaging.

  When the correction of the sewing pattern data in 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 corrects the sewing pattern after the data storage unit 45 has been corrected. The sewing operation is executed based on the data.

[Technical effects of the embodiment of the invention]
When the camera 42 is focused on one of the fabric C1 and the fabric body C2 (for example, the fabric C1) having different heights, the image on the upper surface of the fabric C1 becomes clearer and the focal distance is long. The upper surface of the fabric body C2 is unclear.
Therefore, the image recognition control unit 43 of the sewing machine 1 can greatly separate the distribution of luminance values between the upper surface of the fabric C1 and the upper surface of the fabric main body C2 by sharpening them.
For this reason, the boundary between the fabric C1 and the fabric body C2 can be easily and clearly extracted as an edge.
Thereby, the image recognition control part 43 of the sewing machine 1 can recognize the outer peripheral edge part satisfactorily regardless of the color and material of the outer peripheral edge part of the fabric C1.
In addition, this makes it possible to ensure a certain sewing margin at the outer peripheral edge of the actual fabric C1 and perform more accurate hand movement, and to improve the sewing quality.

  Further, the image recognition control unit 43 of the sewing machine 1 performs a smoothing process between the sharpening process and the edge extraction process (binarization process). For this reason, it is possible to reduce the influence by averaging the luminance value of the pixel that becomes noise with the surroundings, and the region of the fabric C1 and the region of the fabric body C2 can be separated more accurately.

  Further, since the sewing machine 1 includes the illuminating device 42 that can irradiate the cloth C1 from a plurality of irradiation positions and can adjust the illumination intensity for each light source at each irradiation position, generation of a shadow by the outer peripheral edge of the cloth C1 It is possible to adjust the outer peripheral edge of the fabric C1 more accurately while suppressing the influence.

Further, the sewing machine 1 corrects the 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 performs control so that the sewing machine control unit 3 performs the needle movement according to the sewing pattern data. Is going.
For this reason, even if there is a divergence from the sewing pattern of the existing sewing pattern data due to individual differences in cutting or attachment of the fabric C1, sewing can be performed more accurately in response to this, and the sewing quality can be improved. It is possible to improve.

[Others]
In the imaging of the cloth C1 and the cloth main body C2, the case where the height of the camera 41 is adjusted so that the cloth C1 is focused is exemplified. However, the cloth C1 and the cloth main body C2 which is the background are distinguished clearly. If possible difference arises, it will not be restricted to this.
For example, as shown in FIG. 14A, the upper surface of the fabric body C2 may be focused.
If the focal length of the camera 41 is too long, the difference in the thickness of the fabric C1 does not cause an identifiable difference in the sharpness as shown in FIG. A shorter distance is better.
Further, when the cross section of the fabric C1 is reflected, the illumination intensity of each light source of the lighting device 42 is adjusted so as not to cause a shadow of the cross section, or the focus is slightly higher than the upper surface of the fabric C1. It is desirable to match.
In the case where there is a large difference in the contrast of the captured image due to the difference in color or material between the fabric C1 and the fabric body C2, it is desirable to focus on the one where the contrast is more likely to appear.

Further, in the sharpening process described above, the case where the 8-direction Laplacian filter is used has been exemplified, but other sharpening filters or edge enhancement filters may be used. Also, a Laplacian filter with fewer or more directions may be used.
Further, in the smoothing process, the case where the 15 × 15 averaging filter is used is illustrated, but other smoothing filters may be used. Moreover, the range to average can be increased / decreased suitably. Further, in the case of averaging, for example, weighting such as increasing the target pixel may be performed.

Further, the shape recognition unit 4 may compare the shape of the outer peripheral edge of the fabric C1 acquired by imaging with existing sewing pattern data, and determine whether or not the sewing is executed based on the comparison result.
For example, the image recognition control unit 43 may perform error processing according to the situation of disagreement in the comparison between the sewing pattern based on the imaging and the sewing pattern defined in the sewing pattern data. Specifically, if the size of each sewing pattern is significantly different or the number of feature points does not match, the sewing machine control unit 5 is notified of the prohibition of sewing and a notification is made to prompt the operator for confirmation. May be.
As a result, when any abnormality occurs, it is possible to notify the operator and prompt confirmation. For this reason, it is possible to suppress oversight of any abnormality.

  In the above-described embodiment, it is assumed that the number of times the fabric C1 is imaged by the camera 41 is one. However, the focal position is changed from the fabric C1 to the fabric body C2, and imaging is performed a plurality of times. Image data may be acquired. Then, a sewing pattern is acquired for each captured image data, and these coordinate values are combined into one by averaging, for example, to improve accuracy, and compared with the sewing pattern of the sewing pattern data in the data storage unit 45. May be.

Moreover, in the said embodiment, although the fabric C1 was piled up on the fabric main body C2, and the to-be-sewn material which consists of two upper and lower sides was illustrated, you may increase the number of sheets to overlap more.
For example, when a sewing pattern is obtained for the upper and lower three pieces of fabric by overlapping smaller fabrics on the fabric C1, the top surface of the bottom or top fabric is focused. , The top surface of each of the three sheets can be identified. In this case, the histogram has three locations where the distribution is concentrated as shown in FIG. 15. Therefore, the luminance values having the respective frequency distributions as vertices are obtained, and the intermediate luminance values are set as threshold values k1 and k2, and the respective fabrics are obtained. The boundary of is extracted as an edge. As a result, it is possible to acquire a sewing pattern based on the outer peripheral edge portions of the two upper fabrics.

  In addition, the shape recognition unit 4 corrects the sewing pattern defined in the sewing pattern data based on the sewing pattern based on the imaging, but does not perform such an update process, and directly performs sewing from the sewing pattern based on the imaging. The sewing machine control unit 5 may control each part of the sewing machine so as to generate pattern data and move the needle according to the sewing pattern data. In this case, since the existing sewing pattern data correction process is not required, the process can be simplified and speeded up.

  Moreover, although the case where the shape recognition apparatus (shape recognition part 4) was applied to the sewing machine 1 was illustrated in the above-described embodiment, the shape recognition is not limited to this and is applicable to apparatuses in various fields that need to recognize the shape of an object. It is possible to apply the device.

1 sewing machine 4 shape recognition unit (shape recognition device)
41 Camera (imaging means)
42 Illuminating device 43 Image recognition control unit 46 Height adjusting means C1 Cloth (first object)
C2 Fabric body (second object)

Claims (8)

An imaging means for imaging the first object and the second object that are arranged one above the other from above and below,
In a shape recognition apparatus for recognizing a boundary between the first object and the second object from imaged image data obtained by the imaging means,
The imaging means images one object at a focal distance and the other object at a non-focal distance by adjusting the upper surface of one of the first object and the second object to a focal distance,
A sharpening process for greatly converting the luminance value of the pixel according to the difference value when the luminance of each pixel of the captured image data is larger than the luminance of surrounding pixels with respect to the captured image data captured by the imaging unit; An image recognition control unit that performs binarization processing on the captured image data after the sharpening processing, and further performs edge extraction to recognize a boundary between the first object and the second object; A shape recognition device comprising:   The shape recognition apparatus according to claim 1, wherein the image recognition control unit performs a smoothing process on the captured image data after the sharpening process and before the binarization process. A height adjusting means for raising and lowering the imaging means;
The shape recognition apparatus according to claim 1, wherein the image recognition control unit controls the height adjusting unit so that the imaging unit has a preset height.   The shape according to any one of claims 1 to 3, further comprising an illuminating device capable of irradiating the object from a plurality of irradiation positions and capable of adjusting an illumination intensity for each of the plurality of irradiation positions. Recognition device. A shape recognition device according to any one of claims 1 to 4, comprising:
A sewing machine that performs sewing by recognizing an outer shape of a sewing object as the object. Generate sewing pattern data based on the outer shape of the sewing product obtained by the shape recognition device,
The sewing machine according to claim 5, wherein the needle is moved according to the sewing pattern data. Correcting the existing sewing pattern data based on the outer shape of the sewing product obtained by the shape recognition device;
The sewing machine according to claim 5, wherein the needle is moved according to the sewing pattern data. Compare the outline of the sewing product obtained by the shape recognition device with existing sewing pattern data,
The sewing machine according to claim 7, wherein the presence or absence of execution of sewing is determined based on the comparison result.