A machine vision-based garment cutting system and method

Abstract

This invention relates to the field of computer vision recognition technology, and more particularly to a machine vision-based garment cutting system and method. The system includes: an image acquisition unit for capturing fabric images; an illumination unit for detecting the light intensity on the workbench surface and adjusting the number of workbench lighting devices turned on; a planning unit for recording the pattern feature points of each fabric area; a reservation unit for determining the reserved seam area of ​​each fabric area based on shrinkage and stretchability; a layout unit for determining the position of each fabric area on the fabric and generating layout information; a cutting unit for determining the flatness of the fabric; and a quality inspection unit for adjusting the illumination angle of the workbench lighting devices. By utilizing the above-mentioned devices, this invention achieves precise control of lighting, reduces ineffective utilization due to fabric unevenness, and improves fabric utilization while further enhancing the accuracy and reliability of the machine vision-based garment cutting system.

Description

Technical Field

[0001] This invention relates to the field of computer vision recognition technology, and in particular to a clothing cutting system and method based on machine vision. Background Technology

[0002] The current trend in garment cutting is characterized by rapid style updates and shortened production cycles, aiming to bring trendy clothing to market as quickly as possible. Traditional manual cutting methods are relatively inefficient and prone to human error, making it difficult to complete large-scale garment cutting tasks with high quality in a short period. The combination of machine vision technology and automated cutting equipment enables high-speed, high-precision cutting operations, significantly improving garment cutting production efficiency and meeting the demands of rapid turnover and frequent new product launches. In recent years, machine vision technology has made significant progress in both hardware (such as high-resolution, high-frame-rate industrial cameras, high-precision lenses, and stable lighting systems) and software algorithms (such as image processing, feature recognition, target localization, and size measurement algorithms). Industrial cameras can clearly and quickly capture image information of the fabric and cutting area; advanced image processing algorithms can accurately extract useful features even under complex backgrounds and fabric texture interference, such as identifying fabric edges, locating cutting patterns, and detecting defects; size measurement algorithms can achieve precision down to the millimeter or even smaller, providing accurate dimensional data for garment cutting. The maturity of these technologies has made the application of machine vision in the field of clothing cutting possible and has high reliability and practicality.

[0003] Chinese Patent Application Publication No. CN112541517A discloses a method and system for displaying clothing details. The method includes acquiring a target image and a display type; detecting the target image using a pre-trained type detection network to obtain a clothing image with details to be displayed; inputting the clothing image into a pre-trained attribute recognition network and recognizing it in conjunction with the display type to output corresponding clothing attribute values; inputting the clothing image into a pre-trained anchor point prediction network and generating a clothing anchor point image in conjunction with the display type; cropping the clothing anchor point image to output a corresponding detail image; matching the clothing attribute values ​​of the detail image; obtaining a matching display image from the detail image according to the display type; and displaying the display image while simultaneously displaying the matched clothing attribute values. By using a machine to recognize attributes and crop detail images, the method is highly efficient and, through image-text matching, comprehensively and meticulously displays the clothing.

[0004] However, the above method has the following problems: it cannot make the captured image clearer by adjusting the light intensity and angle of the lighting equipment used to capture the image, and it cannot improve the quality of garment processing by details such as patterns and fabric smoothness. Summary of the Invention

[0005] To address this, the present invention provides a machine vision-based garment cutting system and method to overcome the problems in the prior art where the captured images cannot be made clearer by adjusting the light intensity and illumination angle of the lighting equipment used to capture the images, and the quality of garment processing cannot be improved by details such as patterns and fabric smoothness.

[0006] To achieve the above objectives, in one aspect, the present invention provides a machine vision-based garment cutting system, comprising:

[0007] The image acquisition unit is used to take pictures of the fabric placed on the workbench before and after adjusting the lighting equipment and generate an initial fabric image and an adjusted fabric image, as well as to take pictures of a reference garment and a finished garment.

[0008] The lighting unit is connected to the image acquisition unit to detect the light intensity of the workbench surface and adjust the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity of the workbench surface.

[0009] The planning unit, which is connected to the image acquisition unit, is used to divide the reference garment into several fabric regions and record the pattern feature points of each fabric region.

[0010] A reserved unit is used to calculate the shrinkage and stretchability of the fabric by combining a laser displacement sensor and a fabric elasticity database, and to determine the reserved stitching area for each fabric region based on the shrinkage and stretchability.

[0011] The layout unit is connected to the planning unit and the reserved unit respectively, and is used to determine the position of each fabric area on the fabric according to the pattern feature points and the reserved sewing area to generate layout information.

[0012] The cutting unit is connected to the image acquisition unit and the layout unit respectively, and is used to determine the flatness of the fabric according to the adjusted fabric image, and to cut out each fabric area on the fabric according to the layout information and sew each fabric area to generate finished garments.

[0013] The quality inspection unit is connected to the image acquisition unit, the lighting unit, and the cutting unit, respectively. It is used to compare the weight of the finished garment with the weight of the reference garment to determine the qualified finished garment, and to compare the image of the qualified finished garment with the image of the reference garment to count the number and position of inconsistent pattern feature points. Based on the number and position of inconsistent pattern feature points, it issues an adjustment signal, which includes adjusting the illumination angle of the workbench lighting device.

[0014] The fabrics are all placed in the same position on the workbench, the fabric areas are all placed in the same position on the workbench, the finished garment and the reference garment are placed in the same position on the workbench, and the fabric includes several fabric areas.

[0015] Furthermore, the lighting unit adjusts the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity of the workbench surface, wherein,

[0016] If the shadow area on the initial fabric image is smaller than the preset shadow area, then the light intensity of the workbench surface is detected;

[0017] If the shadow area on the initial fabric image is greater than or equal to the preset shadow area, then increase the number of workbench lighting devices that are turned on.

[0018] The preset shadow area is positively correlated with the flat area of ​​the reference garment with its front side facing up.

[0019] Furthermore, after detecting the light intensity of the workbench surface, the lighting unit compares the light intensity of the workbench surface with a first preset light intensity, and determines whether to reduce the number of workbench lighting devices turned on based on the comparison result.

[0020] If the light intensity of the workbench surface is greater than or equal to the first preset light intensity, then it is determined to reduce the number of workbench lighting devices turned on.

[0021] The first preset light intensity is related to the material of the reference clothing.

[0022] Furthermore, the reserved unit combines a laser displacement sensor and a fabric elasticity database to calculate the shrinkage and elongation of the fabric, wherein,

[0023] Establish the fabric elasticity database;

[0024] Cut two pieces of the same size test fabric and record the initial length values;

[0025] The laser displacement sensor is set on two pieces of the test fabric respectively to detect the shrinkage length value of the test fabric after shrinkage and the elongation length value under the action of a preset tensile force.

[0026] The preset tensile force is positively correlated with the thickness of the test fabric.

[0027] Furthermore, the planning unit records the pattern feature points of each fabric region, wherein,

[0028] The pattern feature points include the endpoints and bends of the pattern boundary lines and the dividing lines of different colors in the pattern.

[0029] The pattern feature points of the same fabric area are merged to be consistent.

[0030] Furthermore, the layout unit determines the position of each fabric area on the fabric based on the pattern feature points and the reserved seam area to generate layout information, wherein,

[0031] The shape and area of ​​the fabric region are determined based on the reserved seam area;

[0032] The layout information is generated by combining the pattern feature points and the shape and area of ​​the fabric region on the fabric.

[0033] Furthermore, the cutting unit determines the flatness of the fabric based on the adjusted fabric image, including:

[0034] The grayscale gradient of the fabric is calculated by reading the adjusted fabric image in grayscale mode.

[0035] The number of pixels in the fabric whose grayscale value gradient per unit area is greater than the grayscale value gradient threshold is recorded as the number of high grayscale values.

[0036] If the number of high grayscale values ​​is greater than or equal to the first preset number, the fabric is determined to be uneven.

[0037] If the number of high grayscale values ​​is less than the first preset number, the fabric is determined to be flat.

[0038] The first preset quantity is related to the material of the fabric.

[0039] Further, the quality inspection unit compares the weight of the finished garment with the weight of the reference garment to determine whether the finished garment is qualified, including:

[0040] The difference between the weight of the finished garment and the weight of the reference garment is used to obtain the weight difference of the finished garment.

[0041] The absolute value of the finished product weight difference is compared with a preset weight difference.

[0042] The finished garments are judged to be qualified based on the comparison results.

[0043] If the absolute value of the finished product weight difference is less than the preset weight difference, the finished garment is deemed qualified.

[0044] If the absolute value of the finished product weight difference is greater than or equal to the preset weight difference, the finished garment is determined to be unqualified.

[0045] The preset weight difference is related to the material of the reference garment.

[0046] Further, the quality inspection unit compares the image of the qualified finished garment with the image of the reference garment, counts the number and location of inconsistent pattern feature points, and adjusts the illumination angle of the workbench lighting equipment according to the number and location of inconsistent pattern feature points, including...

[0047] The qualified finished garments are divided into several fabric areas.

[0048] Count the number of inconsistent pattern feature points within each fabric area.

[0049] The number of inconsistent pattern feature points within the fabric area is compared with a second preset number.

[0050] The illumination angle of the workbench lighting equipment is adjusted based on the comparison results, wherein...

[0051] If the number of inconsistent pattern feature points in the fabric area is greater than or equal to the second preset number, then the adjustment signal is issued;

[0052] The second preset quantity is positively correlated with the size of the pattern area on the fabric.

[0053] On the other hand, the present invention provides a machine vision-based garment cutting method, comprising:

[0054] Detect the light intensity on the workbench surface, and adjust the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity on the workbench surface.

[0055] The reference garment is divided into several fabric regions, and the pattern feature points of each fabric region are recorded.

[0056] The shrinkage and stretchability of the fabric are calculated by combining a laser displacement sensor and a fabric elasticity database, and the reserved seam area for each fabric region is determined based on the shrinkage and stretchability.

[0057] Based on the pattern feature points and the reserved stitching area, the positions of each fabric area on the fabric are determined to generate layout information;

[0058] The flatness of the fabric is determined by adjusting the fabric image, and each fabric area is cut out on the fabric according to the layout information and each fabric area is sewn together to generate a finished garment.

[0059] The qualified finished garments are determined based on the weight of the finished garments and the weight of the reference garments;

[0060] The finished garment image of the qualified garment is compared with the reference garment image to count the number and position of inconsistent pattern feature points, and the illumination angle of the workbench lighting device is adjusted according to the number and position of inconsistent pattern feature points.

[0061] Compared with existing technologies, the beneficial effects of this invention are as follows: By analyzing the shadow area on the captured image and combining it with the light intensity of the workbench, this invention adjusts the number of lighting devices turned on. Adjusting the number of lighting devices ensures uniform and appropriate lighting on the workbench, reduces blind spots caused by shadows, and allows for faster and more accurate identification of key features such as details, textures, and colors of objects. By analyzing the shadow area of ​​the image and combining it with the light intensity to adjust the number of lighting devices turned on, precise control of lighting is achieved. The number of devices turned on is only increased when more light is needed to eliminate shadows and meet work requirements, avoiding unnecessary lighting waste and achieving energy saving. This effectively improves the accuracy and reliability of the machine vision-based garment cutting system.

[0062] Furthermore, this invention determines the width of the reserved seam area by combining the shrinkage and stretch properties of the fabric. The shrinkage of the fabric will cause the garment to shrink in size after washing. If the seam area width is not reserved accordingly, the garment may become too tight and uncomfortable to wear. However, by reserving a seam area of ​​appropriate width according to the shrinkage properties, the garment can still maintain a relatively suitable size even after multiple washes, and will not cling to the body due to shrinkage. The looseness and comfort of wearing the garment are maintained, extending the comfortable wearing period of the garment. The reserved seam area width helps to keep the garment's pattern from deforming. For some garments with specific pattern requirements, such as suits and dresses, reserving a seam area according to the fabric characteristics can ensure that the overall outline and lines of the garment still meet the original design intention when the fabric shrinks or is subjected to stretching force. At the same time, it reduces problems such as excessive fabric wear and seam damage caused by size changes, making the garment more durable and further improving the accuracy and reliability of the machine vision-based garment cutting system.

[0063] Furthermore, this invention determines the flatness of the fabric based on adjusted fabric images. By accurately judging the flatness of the fabric, unsuitable fabrics can be screened out in advance, avoiding the use of uneven fabrics in garment production. This ensures that the cut garment pieces are accurate in size and regular in shape, and that the various parts fit and align better during sewing, thereby improving the overall manufacturing precision of the garment and reducing quality defects such as pattern distortion and uneven seams caused by fabric flatness issues. This improves the quality and aesthetics of the finished garment. By using the method of determining flatness by adjusting fabric images, unsuitable fabrics can be excluded or treated in the early stages, avoiding unnecessary rework, making the production process smoother and more efficient, improving overall production efficiency, and shortening the product production cycle. Fabrics with good flatness are easier to cut and plan precisely. After analyzing the fabric flatness through image analysis, the cutting layout can be reasonably arranged according to the flat areas, making fuller use of the fabric and reducing waste caused by ineffective use due to local unevenness. While improving fabric utilization, this further enhances the accuracy and reliability of the machine vision-based garment cutting system.

[0064] Furthermore, this invention compares images of qualified finished garments with reference garment images to statistically analyze the number and location of inconsistent pattern feature points. Based on the number and location of these inconsistent feature points, the illumination angle of the workbench lighting is adjusted. This precise statistical analysis of inconsistent pattern feature points allows for clear identification of specific problems in the pattern presentation of the finished garments. Different lighting angles may cause varying degrees of distortion, color changes, or shadow occlusion in the garment patterns. These visual errors can easily interfere with workers' judgment of the consistency of pattern feature points. By adjusting the illumination angle of the lighting equipment based on the statistical results, light is directed onto the garment at the optimal angle, minimizing visual errors caused by poor lighting. This allows workers to observe and compare pattern feature points more clearly and accurately, completing the inspection work faster and improving the efficiency of the inspection process. It avoids wasting time due to repeated confirmations or misjudgments, making the entire production process smoother and more efficient. While improving the stability of pattern features in each batch of products, this invention further enhances the accuracy and reliability of the machine vision-based garment cutting system. Attached Figure Description

[0065] Figure 1 This is a structural block diagram of the machine vision-based clothing cutting system of the present invention;

[0066] Figure 2 This is a determination diagram for adjusting the number of workbench lighting devices to be turned on, as shown in an embodiment of the present invention.

[0067] Figure 3 This is a determination diagram for fabric flatness in an embodiment of the present invention;

[0068] Figure 4 This is a flowchart of the machine vision-based clothing cutting method of the present invention. Detailed Implementation

[0069] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0070] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0071] It should be noted that in the description of this invention, the terms "upper", "lower", "left", "right", "inner", "outer", etc., which indicate directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings. This is only for the convenience of description and is not intended to indicate or imply that the device or element must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation of this invention.

[0072] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," and "linking" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0073] Please see Figure 1 The diagram shown is a structural block diagram of the machine vision-based garment cutting system of the present invention. An embodiment of the present invention provides a machine vision-based garment cutting system, comprising:

[0074] The image acquisition unit is used to take pictures of the fabric placed on the workbench before and after adjusting the lighting equipment and generate an initial fabric image and an adjusted fabric image, as well as to take pictures of a reference garment and a finished garment.

[0075] The lighting unit, which is connected to the image acquisition unit, is used to detect the light intensity on the workbench surface and adjust the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity on the workbench surface.

[0076] The planning unit, which is connected to the image acquisition unit, is used to divide the reference garment into several fabric regions and record the pattern feature points of each fabric region.

[0077] The reserved unit is used to combine the laser displacement sensor and the fabric elasticity database to calculate the shrinkage and stretch of the fabric, and to determine the reserved stitching area for each fabric area based on the shrinkage and stretch.

[0078] The layout unit, which is connected to the planning unit and the reserved unit respectively, is used to determine the position of each fabric area on the fabric based on the pattern feature points and the reserved seam area to generate layout information.

[0079] The cutting unit is connected to the image acquisition unit and the layout unit respectively. It is used to determine the flatness of the fabric based on the adjusted fabric image, and to cut out each fabric area on the fabric according to the layout information and sew each fabric area to generate finished garments.

[0080] The quality inspection unit is connected to the image acquisition unit, the lighting unit, and the cutting unit respectively. It is used to compare the weight of the finished garment with the weight of the reference garment to determine the qualified finished garment, and to compare the finished garment image of the qualified finished garment with the reference garment image to count the number and position of inconsistent pattern feature points. Based on the number and position of inconsistent pattern feature points, it issues adjustment signals, including adjusting the illumination angle of the workbench lighting equipment.

[0081] The placement of each fabric on the workbench is consistent, the placement of each fabric area on the workbench is consistent, and the placement of the finished garment and the reference garment on the workbench is consistent. The fabric includes several fabric areas.

[0082] It is understandable that when the image acquisition unit captures images of the reference garment and the finished garment, both the reference garment and the finished garment are laid flat on the worktable with their front faces up.

[0083] It is understandable to those skilled in the art that the lighting unit uses a lux meter to detect the light intensity on the work surface, and will not be elaborated further here.

[0084] Understandably, the lighting unit calculates the shadow area on the initial fabric image. Utilizing the characteristic that shadow areas are typically darker parts of an image, it converts the image into a binary image by setting an appropriate threshold. Pixel values ​​are only 0 and 255, with 0 generally representing non-shadow areas and 255 representing shadow areas. Then, it counts the number of pixels with a value of 255 and combines this with the correspondence between the image's actual physical size and pixels to calculate the shadow area.

[0085] Understandably, the layout unit determines the position of each fabric area on the fabric based on the pattern feature points and the reserved seam area to generate layout information. In order to maintain the consistency of the garment pattern, the position of the fabric area is determined on the fabric, taking into account the size and shape of the reserved seam area. At the same time, the position of different fabric areas is adjusted to make full use of the fabric and reduce waste. This will not be elaborated on here.

[0086] Please see Figure 2 As shown, this is a determination diagram for adjusting the number of workbench lighting devices turned on according to an embodiment of the present invention. The lighting unit adjusts the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity on the workbench surface.

[0087] If the shadow area on the initial fabric image is smaller than the preset shadow area, then the light intensity on the workbench surface is detected.

[0088] If the shadow area on the initial fabric image is greater than or equal to the preset shadow area, increase the number of workbench lighting devices turned on.

[0089] In practice, the preset shadow area is 0.005 square meters. If the shadow area on the initial fabric image is 0.003 square meters, which is less than the preset shadow area, the light intensity of the workbench is detected.

[0090] If the shadow area on the initial fabric image is 0.008, which is greater than the preset shadow area, then increase the number of workbench lighting devices that are turned on.

[0091] The preset shadow area is positively correlated with the flat area of ​​the reference garment with its front side facing up.

[0092] It is understandable that the larger the area of ​​the garment laid flat with its front side facing up, the greater the probability of it casting a shadow, the larger the area of ​​the shadow, and the larger the preset shadow area.

[0093] Specifically, this invention adjusts the number of lighting devices on the workbench by analyzing the shadow area in the captured image and combining it with the light intensity of the workbench. Adjusting the number of lighting devices ensures uniform and appropriate lighting on the workbench, reduces blind spots caused by shadows, and allows for faster and more accurate identification of key features such as details, textures, and colors of objects. By analyzing the shadow area in the image and combining it with the light intensity to adjust the number of lighting devices, precise control of lighting is achieved. The number of devices is only increased when more light is needed to eliminate shadows and meet work requirements, avoiding unnecessary lighting waste and achieving energy saving. This effectively improves the accuracy and reliability of the machine vision-based garment cutting system.

[0094] Specifically, after detecting the light intensity on the workbench surface, the lighting unit compares the light intensity on the workbench surface with a first preset light intensity, and determines whether to reduce the number of workbench lighting devices to be turned on based on the comparison result.

[0095] If the light intensity on the workbench is greater than or equal to the first preset light intensity, then it is determined to reduce the number of workbench lighting devices turned on.

[0096] In practice, after the light intensity of the workbench is tested, the first preset light intensity is 500 lux. If the light intensity of the workbench is 620 lux, which is greater than the first preset light intensity, it is determined that the number of workbench lighting devices turned on should be reduced.

[0097] The first preset light intensity is related to the material of the reference clothing.

[0098] Understandably, different clothing materials exhibit significant differences in their light reflection, absorption, and transmission characteristics. For instance, cotton clothing has a relatively rough surface, causing diffuse reflection when light shines on it, absorbing some light and producing a softer, more natural visual effect. Silk, on the other hand, has a smooth surface and high gloss, resulting in more regular light reflection and a tendency to produce highlights and strong reflections. Synthetic materials like polyester fibers may have different light transmittance than natural fibers, affecting the effect of light penetration. By presetting the light intensity to be related to the clothing material, appropriate light intensity can be provided based on these different optical characteristics.

[0099] Specifically, the reserved unit combines a laser displacement sensor and a fabric elasticity database to calculate the fabric's shrinkage and elongation.

[0100] Establish a fabric elasticity database;

[0101] Cut two pieces of the same size test fabric and record the initial length values;

[0102] Laser displacement sensors were installed on two test fabrics to detect the shrinkage length after shrinkage and the elongation length under a preset tensile force.

[0103] Understandably, representative fabric samples should be selected from the batch of fabric to be tested according to statistical principles. The number of samples should be sufficient to accurately reflect the characteristics of the entire batch; generally, at least 10 samples are recommended. The sample size should facilitate subsequent measurement operations; for example, they can be cut into squares with sides of approximately 30 cm or rectangles with a length of 50 cm and a width of 10 cm. Simultaneously, the cut edges of the samples should be neat and free of obvious defects to avoid interfering with the measurement results. The laser displacement sensor should be calibrated to ensure its measurement accuracy meets requirements. The laser displacement sensor should be installed appropriately according to the measurement scenario and the placement of the fabric to ensure stable and accurate measurement of displacement changes at various points on the fabric surface. Typically, the sensor is fixed on a stable bracket, and its position and angle are adjusted so that the laser beam is perpendicular to the surface. The test area of ​​the fabric should be directly illuminated, and the sensor should be kept at a suitable distance from the fabric, generally between a few centimeters and tens of centimeters, to obtain a clear and accurate measurement signal. Collect elasticity-related parameters of various common fabrics (such as pure cotton, polyester fiber, wool, silk, and other fabrics of different materials, weaves, and thicknesses). These parameters should include, but are not limited to, basic mechanical property data such as initial elastic modulus, yield strength, and ultimate elongation in different directions (warp and weft), as well as data such as shrinkage range and elongation range under different conditions (such as the influence of environmental factors such as different washing methods, temperature, and humidity) accumulated through previous experiments or actual production experience. Each record in the database should clearly indicate the detailed information of the fabric, such as material, composition ratio, weave, thickness, and manufacturer, to facilitate subsequent query and comparative analysis.

[0104] Understandably, based on the common washing or treatment methods used in the actual application of the fabric, a suitable simulated shrinkage treatment method can be selected. For example, a standard washing machine program can be used (selecting parameters such as water temperature, washing time, and spin speed suitable for the fabric material; for cotton fabrics, a water temperature of around 30℃, a gentle washing mode, and a washing time of around 30 minutes are commonly used). Alternatively, a soaking treatment can be performed (completely immersing the fabric sample in an appropriate amount of water, soaking for a certain period of time, and then removing it; the soaking water temperature is also set according to the fabric characteristics, and the time can be set to 1-2 hours, etc.). Afterward, it can be air-dried naturally or dried according to a standard drying program (such as drying in a 60℃ drying oven for a certain period of time, the specific time being adjusted according to factors such as fabric thickness). The shrunken fabric sample is then placed flat on a measuring platform, and the shrinkage length value is measured using a laser displacement sensor.

[0105] Understandably, when measuring fabric elongation, one end of the fabric sample is fixed to the fixed fixture of a tensile testing device such as a tensile testing machine, ensuring that the fixture clamps the fabric tightly without causing damage. The other end is connected to a movable tensile fixture, so that the laser beam of the laser displacement sensor is aligned with the test area of ​​the fabric sample (generally selected at the middle position in the length direction of the fabric, which can more evenly reflect the overall elongation of the fabric). The measurement parameters of the sensor are set so that it can accurately record the displacement change of the fabric during the stretching process in real time, which is the elongation length value.

[0106] The preset tensile force is positively correlated with the thickness of the test fabric.

[0107] It is understandable that thicker fabrics usually mean that they have more fiber layers or that the fibers are thicker and denser. This allows them to withstand greater tensile force when subjected to external stretching before they begin to show obvious deformation (such as elongation, yielding, etc.). On the other hand, thinner fabrics have relatively fewer fibers and a relatively thinner structure, so they can withstand relatively less tensile force. Therefore, the preset tensile force is positively correlated with the thickness of the test fabric.

[0108] Specifically, this invention determines the width of the reserved seam area on the fabric by combining the fabric's shrinkage and stretch properties. Fabric shrinkage causes garments to shrink after washing; without a corresponding seam area width, the garment may become too tight and uncomfortable to wear. However, by reserving a seam area of ​​appropriate width based on shrinkage, the garment maintains a relatively suitable size even after multiple washes, preventing it from clinging to the body due to shrinkage. This maintains a comfortable and relaxed fit, extending the garment's comfortable wearing period. A reasonably reserved seam area width also helps maintain the garment's shape. For garments with specific shape requirements, such as suits and dresses, reserving a seam area based on fabric characteristics ensures that the overall silhouette and lines of the garment still conform to the original design even when the fabric shrinks or stretches. This also reduces excessive fabric wear and seam damage caused by size changes, making the garment more durable and further improving the accuracy and reliability of the machine vision-based garment cutting system.

[0109] Specifically, the planning unit records the pattern feature points of each fabric area, among which...

[0110] The feature points of the pattern include the endpoints and bends of the pattern boundary lines and the dividing lines between different colors in the pattern;

[0111] Merge the pattern feature points of the same fabric area to make them consistent.

[0112] It is understandable that merging the pattern feature points of the same fabric area into a unified whole means merging the pattern feature points in the same position and marking the different pattern feature points in the unidentified fabric area.

[0113] Specifically, the layout unit determines the position of each fabric area on the fabric based on the pattern feature points and reserved seam areas to generate layout information.

[0114] Determine the shape and area of ​​the fabric area based on the reserved seam area;

[0115] The pattern features are combined with the shape and area of ​​the fabric region to create a layout on the fabric, generating layout information.

[0116] Understandably, the layout on the fabric involves combining the pattern feature points with the shape and area of ​​the fabric region. First, the area size and shape of the fabric region are recorded. Second, the distribution of the pattern feature points on the fabric region is determined. Then, the layout direction of the fabric is determined. The area size and shape information of the fabric, as well as the coordinates of the Luigi pattern feature points, are imported into the software using professional garment CAD software or graphic layout software. Then, the layout is tested using the software's automatic layout function (some software can automatically generate multiple layout schemes based on set parameters such as fabric utilization rate and cut piece interval) or by manually dragging and adjusting the position of the cut pieces.

[0117] Understandably, the exact position of each cut piece on the fabric must be recorded. If using a two-dimensional coordinate system (with a corner of the fabric as the origin, establishing horizontal and vertical axes), record the coordinates of each vertex of the cut piece. For example, the coordinates of the upper left vertex of the shirt front piece are (10 cm, 20 cm), and the coordinates of the lower right vertex are (40 cm, 80 cm), etc. This accurately indicates the placement of the cut piece in the fabric, facilitating accurate positioning during subsequent cutting operations. For irregularly shaped cut pieces, in addition to vertex coordinates, the coordinates of key curve control points or information describing their relative positional relationships, such as angles and distances, can also be recorded.

[0118] Understandably, it is important to record in detail the pattern feature points within the area covered by each piece of fabric. For example, if a piece contains several key feature points, their relative positions within the piece (such as the center, one side of the edge, etc.), and the relationship between these feature points and the shape of the piece, this information can be presented through a combination of text descriptions and labeled images (clearly marking the shape of the piece and the corresponding pattern feature points on the images). This will help cutting personnel understand the information and take care to protect these important pattern feature points during the cutting process.

[0119] Understandably, the fabric utilization rate after layout is calculated using the formula (used fabric area ÷ total fabric area × 100%). For example, if the used fabric area after layout is 8 square meters and the total fabric area is 10 square meters, then the fabric utilization rate is 80%. Simultaneously, the shape and size of any remaining blank areas of fabric, as well as their position within the overall fabric composition, should be clearly recorded. This provides a basis for decisions regarding whether to continue using these remaining fabrics for small part cutting or waste disposal.

[0120] Please see Figure 3 As shown, this is a determination diagram for fabric flatness in an embodiment of the present invention. The cutting unit determines the fabric flatness based on the adjusted fabric image, including:

[0121] Read and adjust the fabric image in grayscale mode to calculate the grayscale gradient of the fabric.

[0122] The number of pixels in a fabric whose grayscale value gradient per unit area is greater than a grayscale value gradient threshold is recorded as the number of high grayscale values.

[0123] If the number of high grayscale values ​​is greater than or equal to the first preset number, the fabric is determined to be uneven.

[0124] If the number of high grayscale values ​​is less than the first preset number, the fabric is determined to be flat.

[0125] It is understandable that if the fabric surface is flat and smooth, the gray value changes relatively slowly and evenly in the image, and the corresponding gray value gradient is relatively small. However, if the fabric has wrinkles or is uneven, the gray value of the corresponding area in the image will change suddenly, resulting in an increased gray value gradient. Therefore, the smoothness of the fabric can be inferred by analyzing the gray value gradient of different areas of the fabric image.

[0126] In practice, the number of high grayscale values ​​is 15,000. If the number of high grayscale values ​​is 31,852, which is greater than the first preset number, the fabric is determined to be uneven.

[0127] If the number of high grayscale values ​​is 9856, which is less than the first preset number, the fabric is determined to be flat.

[0128] The first preset quantity is related to the material of the fabric.

[0129] It can be understood that for smooth and flat fabric materials, such as silk, which has a very smooth surface, delicate and uniform texture, the pixel gray value changes relatively gently in its image, and the overall gray value gradient is small. Therefore, the corresponding threshold may be set relatively low (gradient threshold is between 50 - 100), and at such a threshold, the number of pixel points exceeding the threshold is usually also small, accounting for 10% - 20% of the total number of pixel points; for rough and textured fabric materials, like denim, coarse linen, etc., which have a rough surface, obvious and complex texture, the gray value changes in its image are relatively drastic, and naturally, there are more pixel points with a large gray value gradient. In this case, the threshold often needs to be set high (gradient threshold is between 150 - 250), and the proportion of pixel points exceeding the threshold is relatively high, reaching 30% - 60%.

[0130] Specifically, the present invention determines the flatness of the fabric based on the adjusted fabric image. By accurately judging the flatness of the fabric, it is possible to screen out fabrics that do not meet the requirements in advance, avoid using uneven fabrics for making clothing, ensure that the cut garment pieces have accurate dimensions and regular shapes, and each part can also fit and align better during sewing, thereby improving the overall production accuracy of the clothing, reducing quality defects such as distorted garment patterns and uneven seams caused by the flatness problem of the fabric itself, improving the quality and aesthetics of the finished clothing. Using the method of determining flatness based on the adjusted fabric image can exclude fabrics that do not meet the requirements or conduct targeted processing in the early stage, avoid unnecessary rework, make the production process smoother and more efficient, improve the overall production efficiency, shorten the production cycle of the product, and fabrics with good flatness are more suitable for precise cutting and layout planning. After knowing the flatness of the fabric through image analysis, the cutting layout can be reasonably arranged according to the flat area, enabling the fabric to be more fully utilized, reducing waste caused by the inability to effectively utilize due to local unevenness of the fabric, and further improving the accuracy and reliability of the clothing cutting system based on machine vision while increasing the fabric utilization rate.

[0131] Specifically, the quality inspection unit determines qualified finished clothing by comparing the weight of the finished clothing with the weight of the reference clothing, including

[0132] subtracting the weight of the reference clothing from the weight of the finished clothing to obtain the finished weight difference

[0133] comparing the absolute value of the finished weight difference with the preset weight difference

[0134] judging whether the finished clothing is qualified according to the comparison result, where

[0135] if the absolute value of the finished weight difference is less than the preset weight difference, it is determined that the finished clothing is qualified;

[0136] If the absolute value of the weight difference of the finished products is greater than or equal to the preset weight difference, the finished garments are deemed unqualified.

[0137] In practice, the preset weight difference is 100 grams. If the absolute value of the finished product weight difference is 86 grams, which is less than the preset weight difference, the finished garment is deemed to be qualified.

[0138] If the absolute value of the weight difference of the finished product is 246 grams, which is greater than the preset weight difference, the finished garment is deemed unqualified.

[0139] The preset weight difference is related to the material of the reference garment.

[0140] It is understandable that different garments have different fabric densities, so the weight will be different when the same amount of fabric is used, hence the different preset weight differences.

[0141] Specifically, the quality inspection unit compares images of qualified finished garments with reference garment images, statistically analyzing the number and location of inconsistent pattern feature points. Based on the number and location of these inconsistent feature points, the unit adjusts the illumination angle of the workbench lighting equipment, including...

[0142] The qualified finished garments are divided into several fabric areas.

[0143] Count the number of inconsistent pattern feature points in each fabric area.

[0144] The number of inconsistent pattern feature points within the fabric area is compared with a second preset number. Based on the comparison result, the illumination angle of the workbench lighting equipment is adjusted.

[0145] If the number of inconsistent pattern feature points in the fabric area is greater than or equal to the second preset number, an adjustment signal is issued.

[0146] In practice, the second preset quantity is 5. If the number of inconsistent pattern feature points in the fabric area is 7, which is greater than the second preset quantity, an adjustment signal is issued.

[0147] The second preset quantity is positively correlated with the size of the pattern area on the fabric.

[0148] It is understandable that the larger the area of ​​the pattern on the fabric, the more pattern feature points are generated, and the greater the number of inconsistent pattern feature points detected, so the larger the second preset number.

[0149] Specifically, this invention compares images of qualified finished garments with reference garment images to statistically analyze the number and location of inconsistent pattern feature points. Based on the number and location of these inconsistent feature points, the illumination angle of the workbench lighting is adjusted. This precise statistical analysis of inconsistencies in pattern feature points allows for clear identification of specific problems in the pattern presentation of the finished garments. Different lighting angles can cause varying degrees of distortion, color changes, or shadow occlusion in the garment patterns, and these visual errors can easily interfere with workers' judgment of the consistency of pattern feature points. By adjusting the illumination angle of the lighting equipment based on the statistical results, light is directed onto the garment at the optimal angle, minimizing visual errors caused by poor lighting. This allows workers to observe and compare pattern feature points more clearly and accurately, completing the inspection work faster and improving the efficiency of the inspection process. It avoids wasting time due to repeated confirmations or misjudgments, making the entire production process smoother and more efficient. While improving the stability of pattern features in each batch of products, it further enhances the accuracy and reliability of the machine vision-based garment cutting system.

[0150] Please see Figure 4 The diagram shows a flowchart of the machine vision-based garment cutting method of the present invention. The present invention provides a machine vision-based garment cutting method, comprising:

[0151] Step S1 detects the light intensity on the workbench surface and adjusts the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity on the workbench surface.

[0152] Step S2: Divide the reference garment into several fabric areas and record the pattern feature points of each fabric area.

[0153] Step S3: Combine the laser displacement sensor and the fabric elasticity database to calculate the shrinkage and stretch of the fabric, and determine the reserved seam area for each fabric region based on the shrinkage and stretch.

[0154] Step S4: Based on the pattern feature points and reserved seam areas, determine the position of each fabric area on the fabric to generate layout information;

[0155] Step S5: Determine the flatness of the fabric based on the adjusted fabric image, and cut out each fabric area on the fabric according to the layout information and sew each fabric area to generate the finished garment.

[0156] Step S6: Determine the qualified finished garments based on the weight of the finished garments and the weight of the reference garments.

[0157] Step S7: Compare the finished garment image of the qualified finished garment with the reference garment image to count the number and location of inconsistent pattern feature points, and adjust the illumination angle of the workbench lighting equipment according to the number and location of inconsistent pattern feature points.

[0158] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

[0159] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A machine vision-based garment cutting system, characterized in that, include: The image acquisition unit is used to take pictures of the fabric placed on the workbench before and after adjusting the lighting equipment and generate an initial fabric image and an adjusted fabric image, as well as to take pictures of a reference garment and a finished garment. The lighting unit is connected to the image acquisition unit to detect the light intensity of the workbench surface and adjust the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity of the workbench surface. The planning unit, which is connected to the image acquisition unit, is used to divide the reference garment into several fabric regions and record the pattern feature points of each fabric region. A reserved unit is used to calculate the shrinkage and stretchability of the fabric by combining a laser displacement sensor and a fabric elasticity database, and to determine the reserved stitching area for each fabric region based on the shrinkage and stretchability. The layout unit is connected to the planning unit and the reserved unit respectively, and is used to determine the position of each fabric area on the fabric according to the pattern feature points and the reserved sewing area to generate layout information. The cutting unit is connected to the image acquisition unit and the layout unit respectively, and is used to determine the flatness of the fabric according to the adjusted fabric image, and to cut out each fabric area on the fabric according to the layout information and sew each fabric area to generate finished garments. The quality inspection unit is connected to the image acquisition unit, the lighting unit, and the cutting unit, respectively. It is used to compare the weight of the finished garment with the weight of the reference garment to determine the qualified finished garment, and to compare the image of the qualified finished garment with the image of the reference garment to count the number and position of inconsistent pattern feature points. Based on the number and position of inconsistent pattern feature points, it issues an adjustment signal, which includes adjusting the illumination angle of the workbench lighting device. The lighting unit adjusts the number of workbench lighting devices to be turned on based on the shadow area on the initial fabric image and the light intensity of the workbench surface. If the shadow area on the initial fabric image is smaller than the preset shadow area, then the light intensity of the workbench surface is detected; If the shadow area on the initial fabric image is greater than or equal to the preset shadow area, then increase the number of workbench lighting devices that are turned on. The preset shadow area is positively correlated with the flat area of ​​the reference garment with its front side facing up; The fabrics are all placed in the same position on the workbench, the fabric areas are all placed in the same position on the workbench, the finished garment and the reference garment are placed in the same position on the workbench, and the fabric includes several fabric areas.

2. The machine vision-based garment cutting system according to claim 1, characterized in that, When the lighting unit has completed detecting the light intensity of the workbench surface, it compares the light intensity of the workbench surface with a first preset light intensity, and determines whether to reduce the number of workbench lighting devices that are turned on based on the comparison result. If the light intensity of the workbench surface is greater than or equal to the first preset light intensity, then it is determined to reduce the number of workbench lighting devices turned on. The first preset light intensity is related to the material of the reference clothing.

3. The machine vision-based garment cutting system according to claim 2, characterized in that, The reserved unit, in conjunction with a laser displacement sensor and a fabric elasticity database, calculates the shrinkage and elongation of the fabric, wherein... Establish the fabric elasticity database; Cut two pieces of the same size test fabric and record the initial length values; The laser displacement sensor is set on two pieces of the test fabric respectively to detect the shrinkage length value of the test fabric after shrinkage and the elongation length value under the action of a preset tensile force. The preset tensile force is positively correlated with the thickness of the test fabric.

4. The machine vision-based garment cutting system according to claim 3, characterized in that, The planning unit records the pattern feature points of each fabric region, wherein... The pattern feature points include the endpoints and bends of the pattern boundary lines and the dividing lines of different colors in the pattern. The pattern feature points of the same fabric area are merged to be consistent.

5. The machine vision-based garment cutting system according to claim 4, characterized in that, The layout unit determines the position of each fabric area on the fabric based on the pattern feature points and the reserved seam area, and generates layout information, wherein... The shape and area of ​​the fabric region are determined based on the reserved seam area; The layout information is generated by combining the pattern feature points and the shape and area of ​​the fabric region on the fabric.

6. The machine vision-based garment cutting system according to claim 5, characterized in that, The cutting unit determines the flatness of the fabric based on the adjusted fabric image, including: The grayscale gradient of the fabric is calculated by reading the adjusted fabric image in grayscale mode. The number of pixels in the fabric whose grayscale value gradient per unit area is greater than the grayscale value gradient threshold is recorded as the number of high grayscale values. If the number of high grayscale values ​​is greater than or equal to the first preset number, the fabric is determined to be uneven. If the number of high grayscale values ​​is less than the first preset number, the fabric is determined to be flat. The first preset quantity is related to the material of the fabric.

7. The machine vision-based garment cutting system according to claim 6, characterized in that, The quality inspection unit compares the weight of the finished garment with the weight of the reference garment to determine whether the finished garment is qualified, including: The difference between the weight of the finished garment and the weight of the reference garment is used to obtain the weight difference of the finished garment. The absolute value of the finished product weight difference is compared with a preset weight difference. The finished garments are judged to be qualified based on the comparison results. If the absolute value of the finished product weight difference is less than the preset weight difference, the finished garment is deemed qualified. If the absolute value of the finished product weight difference is greater than or equal to the preset weight difference, the finished garment is determined to be unqualified. The preset weight difference is related to the material of the reference garment.

8. The machine vision-based garment cutting system according to claim 7, characterized in that, The quality inspection unit compares the image of the qualified finished garment with the image of the reference garment, counts the number and location of inconsistent pattern feature points, and adjusts the illumination angle of the workbench lighting equipment based on the number and location of inconsistent pattern feature points. The qualified finished garments are divided into several fabric areas. Count the number of inconsistent pattern feature points within each fabric area. The number of inconsistent pattern feature points within the fabric area is compared with a second preset number. The illumination angle of the workbench lighting equipment is adjusted based on the comparison results, wherein... If the number of inconsistent pattern feature points in the fabric area is greater than or equal to the second preset number, then the adjustment signal is issued; The second preset quantity is positively correlated with the size of the pattern area on the fabric.

9. A machine vision-based garment cutting method using the machine vision-based garment cutting system according to any one of claims 1-8, characterized in that, include: Detect the light intensity on the workbench surface, and adjust the number of workbench lighting devices turned on based on the shadow area on the initial fabric image and the light intensity on the workbench surface. The reference garment is divided into several fabric regions, and the pattern feature points of each fabric region are recorded. The shrinkage and stretchability of the fabric are calculated by combining a laser displacement sensor and a fabric elasticity database, and the reserved seam area for each fabric region is determined based on the shrinkage and stretchability. Based on the pattern feature points and the reserved stitching area, the positions of each fabric area on the fabric are determined to generate layout information; The flatness of the fabric is determined by adjusting the fabric image, and each fabric area is cut out on the fabric according to the layout information and each fabric area is sewn together to generate a finished garment. The qualified finished garments are determined based on the weight of the finished garments and the weight of the reference garments; The finished garment image of the qualified garment is compared with the reference garment image to count the number and position of inconsistent pattern feature points, and the illumination angle of the workbench lighting device is adjusted according to the number and position of inconsistent pattern feature points.

Images