Yarn breakage image detection method for beam area of sizing machine

Abstract

The application discloses a yarn breakage image detection method for a warp beam area of a sizing machine and belongs to the technical field of yarn breakage detection of a warp beam area of a sizing machine.The method is characterized in that: firstly, the image of the warp beam of the sizing machine is collected; secondly, the straight edge detection, the circular arc edge detection and the warp beam area positioning algorithm of the warp beam of the sizing machine are designed according to the characteristics of the image of the warp beam of the sizing machine; and thirdly, the warp beam area detection and the yarn breakage statistical determination algorithm are designed, so that the yarn breakage image detection of the warp beam area of the sizing machine is realized.The method focuses on the warp beam area only, the industrial camera without a special light source is used to collect the image of the warp beam of the sizing machine, the hardware arrangement requirement is greatly simplified, the system complexity is reduced, and the reliability of the yarn breakage detection of the warp beam area of the sizing machine is ensured.

Description

Technical Field

[0001] This invention relates to a method for detecting yarn breakage images in the warp beam area of ​​a sizing machine, belonging to the technical field of yarn breakage detection in the warp beam area of ​​a sizing machine. Background Technology

[0002] The warp beam zone of a sizing machine is a crucial sizing stage in textile production, where yarn is sized before entering the weaving process. However, due to equipment wear, improper operation, or yarn quality issues, yarn breakage is prone to occur in the warp beam zone. These breakages not only affect weaving efficiency but can also lead to fabric quality problems and even equipment downtime. Therefore, timely and accurate detection of yarn breakage in the warp beam zone is of great significance for improving production efficiency, reducing losses, and ensuring product quality.

[0003] Many textile companies still rely on manual inspection to detect yarn breaks in the warp beam area. This method is not only inefficient but also susceptible to human error, leading to missed or false detections. Furthermore, manual inspection requires significant manpower, increasing labor costs, and worker fatigue during long inspection periods can further reduce accuracy. Therefore, there is an urgent need for an automated and intelligent inspection method to replace traditional manual inspection, improve efficiency and accuracy, and reduce operating costs for businesses.

[0004] Currently, there are image-based methods for detecting yarn breakage on sizing machines. For example, CN 117953234 A discloses a machine vision-based algorithm for detecting yarn breakage in the sizing groove area of ​​a sizing machine. This algorithm uses machine vision technology to acquire and process images of the sizing groove area, achieving automatic detection and alarm for yarn breakage faults, thus solving the problem of low efficiency in manual inspection to some extent. However, this method can only detect yarn breakage in the sizing groove area and cannot detect yarn breakage occurring in the warp beam area. CN117974610A discloses a method for detecting yarn breakage in the reed tooth area of ​​a sizing machine based on an image processing algorithm; CN113516628B discloses a machine vision-based method for detecting reed collisions during sizing; and CN119338814B discloses a method and device for detecting reed collisions during sizing based on image processing. These methods detect yarn breakage faults in the reed tooth area of ​​sizing machines in textile production using image processing algorithms or neural network models, achieving automated fault identification, solving the problems of high labor costs and resource waste in textile production, and improving the quality of sizing products and the level of intelligence in production equipment. However, the above methods can only detect yarn breaks in the reed area, not in the warp beam area. CN 214844889 U discloses a machine vision-based yarn breakage detection system for sizing machines. By setting up cameras and light sources at different locations on the sizing machine, a machine vision detection system is built to detect yarn breaks in real time and issue warnings, solving the problem that traditional manual inspections cannot detect yarn breaks in a timely manner. However, this detection system requires an industrial light source of model TPA40-90ADJ and an industrial camera of model HIKVISION in the warp beam area. Its computing module uses a model SG150-BGCM display and a model VOS 3671-R15 computer, which is relatively complex and costly. This system uses a high-cost camera and a specific light source to capture more ideal images of the warp beam area, thus enabling yarn breakage detection using conventional image processing methods. However, the production environment of textile enterprises is relatively harsh. On the one hand, setting up high-cost equipment often leads to higher maintenance costs. On the other hand, even with a specific light source, it is difficult to ensure stable imaging results, causing conventional image processing methods to fail. Therefore, the effectiveness of this system is affected by both cost and detection reliability, making it difficult to meet the detection needs of actual production. In summary, setting up a specific industrial camera and light source in the warp beam area, combined with conventional image processing methods to detect yarn breakage, results in a system that is highly complex, costly, difficult to maintain, and unstable in detection.

[0005] In summary, existing methods have limitations such as being unable to detect yarn breakage in the warp beam area, or having high system complexity, high cost, and difficulty in maintenance. Summary of the Invention

[0006] To address at least one of the aforementioned problems, this invention provides a method for detecting yarn breakage in the warp beam region of a sizing machine. This invention focuses solely on the warp beam region, employing an industrial camera that does not require a dedicated light source for warp beam image acquisition. Based on the characteristics of warp beam images, it designs algorithms for straight-line edge detection, arc edge detection, and warp beam region localization, as well as algorithms for warp beam region detection and yarn breakage statistical determination. By utilizing adaptive parameter settings and targeted algorithm design, it significantly simplifies hardware layout requirements, reducing system complexity while ensuring the reliability of yarn breakage detection in the warp beam region of the sizing machine.

[0007] The method for detecting yarn breakage images in the warp beam area of ​​a sizing machine according to the present invention includes:

[0008] A1. Image Acquisition; Acquire image I of the warp beam from the sizing machine;

[0009] A2. Warp beam region localization: From the warp beam image I of the sizing machine, extract the straight edges of the warp beam, extract the arc edges of the warp beam, and finally extract the binary image J of the independent warp beam region. k k∈[1,n], where n is the number of independent meridian regions;

[0010] A3. Meridian region detection; J in the binary image of each meridian region. k Calculate the axis l of the meridian region k Then, the Hough circle detection algorithm is applied to image I to obtain the binary image B containing the circle detection. R And record B R The center pixel coordinates p of each circle's edge i i∈[1,n r ], n r It is the number of circular edges; B R In, satisfying p i To axis l k Circles with a distance greater than d are removed to obtain the binary image B of the broken yarn. k d = [1, 50] pixels;

[0011] A4. Yarn breakage statistical determination; statistical analysis of binary images of yarn breaks. k The total number N of pixels whose gradient in the longitudinal image is greater than the threshold t2 at positions where the value is 1. k If N k If the value is greater than the set threshold t3, it is determined that a yarn breakage has been detected.

[0012] In one embodiment, in step A1, image acquisition is performed by acquiring warp beam image I of the sizing machine through a warp beam image acquisition device.

[0013] In one embodiment, in step A1, the image acquisition involves positioning the image acquisition device between the two left and right channel steel columns and the two upper and lower cross braces, and acquiring the warp beam image along the warp beam direction. Optionally, the image acquisition device is fixed to the warp beam frame cross brace of the sizing machine.

[0014] In one embodiment, in step A1, the image acquisition is performed using a CCD industrial camera, i.e., an image acquisition device.

[0015] In one embodiment, the calculation method for extracting the straight edge of the warp axis, the circular edge of the warp axis, and the independent warp axis region in step A2 is as follows:

[0016] Step 1: Extract straight edges of the warp beam; Use the Hough line detection algorithm on the warp beam image I from the sizing machine to extract straight edges that conform to the angle ranges of 125°~145° and 35°~55°, and record them as binary image B of the straight edges. l ;

[0017] Step 2: Extract the edge of the arc along the meridian; apply the Hough circle detection algorithm to image I to obtain the binary image B of the circle detection. R Locate the pixel positions in image I where the gradient is less than t1, and then add B. R Setting the values ​​at these pixel locations to 0 yields the binary image B of the arc edge. r ; where t1 is calculated using the binary image B of the straight line edge. l The image gradient information at the mid-edge position is adaptively set to avoid parameter sensitivity issues caused by unstable image imaging conditions. The calculation formula is as follows:

[0018]

[0019] Where D is the gradient map of image I, H and W are the pixel height and width of the image, respectively, and (x, u) represents the pixel coordinates of the image;

[0020] Step 3: Extract independent meridian regions; in the binary image at the meridian edge B=B r +B l Above, the regions enclosed by the edges in B at the four corner positions of the image matrix are used to construct four binary images J representing the longitudinal regions. k k∈[1,n], n=4 is the number of independent meridian regions;

[0021] In one embodiment, in step A3, the method for calculating the axis of the meridian region is to obtain a binary image J of the meridian region based on the results of Hough edge detection. k Given the slopes and intercepts of the two straight edges in the diagram, denoted as (m1, b1) and (m2, b2), calculate the mean value of both to determine the axis l. kThe slope and intercept (m) l ,b l ).

[0022] In one implementation, the formula for calculating the threshold t2 in step A4 is:

[0023] t2=γ2t1

[0024] Where γ2 is the gradient sensitivity parameter; optionally, γ2∈[0.1,0.75]; further optionally, γ2=0.5.

[0025] In one implementation, the threshold t3 is calculated using the following formula:

[0026]

[0027] Where γ3∈[1,30] is the regional sensitivity parameter; optionally, γ3=10.

[0028] This invention also claims a yarn breakage image detection system for the warp beam area of ​​a sizing machine, wherein the yarn breakage image detection system for the warp beam area of ​​a sizing machine performs the method described in this invention.

[0029] In one embodiment, the yarn breakage image detection system for the warp beam area of ​​the sizing machine includes an image acquisition device, a warp beam area detection device, a warp beam area detection device, and a yarn breakage statistical determination device.

[0030] Beneficial effects:

[0031] This invention focuses solely on the warp beam region, employing an industrial camera that does not require a dedicated light source to acquire warp beam images from the sizing machine. Algorithms for straight-line edge detection, arc edge detection, and warp beam region localization are designed specifically for the characteristics of warp beam images from the sizing machine. Algorithms for warp beam region detection and yarn breakage statistical determination are also included. By utilizing adaptive parameter settings and targeted algorithm design, hardware layout requirements are significantly simplified, reducing system complexity while ensuring the reliability of yarn breakage detection in the warp beam region of the sizing machine. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 This is a flowchart of the method of the present invention;

[0034] Figure 2 This is image I of the warp beam of the sizing machine;

[0035] Figure 3 It is a binary image B with straight line edges. l picture;

[0036] Figure 4 It is a binary image B of the arc edge. r picture;

[0037] Figure 5 It is the binary image B of the meridian edge;

[0038] Figure 6 J is a binary image of the meridian region. k picture;

[0039] Figure 7 It is the axis of the meridian l k Figure; the blue line is axis I. k ;

[0040] Figure 8 This is a diagram of the yarn breakage test results;

[0041] Figure 9 The image shows the detection results processed using the global Otsu thresholding algorithm.

[0042] Figure 10 This is a detection result image using a region adaptive threshold algorithm. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of the present invention clearer, the embodiments of the present invention will be described in further detail below with reference to the accompanying drawings.

[0044] Example 1

[0045] like Figure 1 The diagram shows a flowchart of the yarn breakage image detection method for the warp beam area of ​​a sizing machine according to the present invention. First, the warp beam image of the sizing machine is acquired, the warp beam area is located, then the warp beam area is detected, and finally, yarn breakage statistics are determined.

[0046] Specific methods for detecting yarn breakage images in the warp beam area of ​​a sizing machine include:

[0047] A1. Image Acquisition; An image acquisition device for the warp beam of a sizing machine (a CCD industrial camera is used in this embodiment) is used to acquire images of the warp beam of the sizing machine; the image acquisition device is fixed on the cross brace of the warp beam frame of the sizing machine, located in the middle of the two left and right channel steel columns and the two upper and lower cross braces, and acquires images of the warp beam along the warp beam direction; For example... Figure 2 Image I shows the warp beam of the sizing machine;

[0048] A2. Warp beam region localization: Sequentially extract the straight edges of the warp beam from the warp beam image I of the sizing machine, extract the arc edges of the warp beam, and finally extract the binary image J of the independent warp beam region.k k∈[1,n], where n is the number of independent meridian regions;

[0049] The calculation methods for extracting the straight edge of the warp axis, the circular edge of the warp axis, and the independent warp axis region in step A2 are as follows:

[0050] Step 1: Extract straight edges of the warp beam; Use the Hough line detection algorithm on the warp beam image I from the sizing machine to extract straight edges that conform to the angle ranges of 125°~145° and 35°~55°, and record them as binary image B of the straight edges. l ;like Figure 3 As shown, this is the binary image B of the straight line edge. l ;

[0051] Step 2: Extract the edge of the arc along the meridian; apply the Hough circle detection algorithm to image I to obtain the binary image B of the circle detection. R Locate the pixel positions in image I where the gradient is less than t1, and then add B. R Setting the values ​​at these pixel locations to 0 yields the binary image B of the arc edge. r (like Figure 4 (as shown); where t1 is calculated using the binary image B of the straight line edge. l The image gradient information at the mid-edge position is adaptively set to avoid parameter sensitivity issues caused by unstable image imaging conditions. The calculation formula is as follows:

[0052]

[0053] Where D is the gradient map of image I, H and W are the pixel height and width of the image, respectively, and (x, u) represents the pixel coordinates of the image;

[0054] Step 3: Extract independent meridian regions; in the binary image at the meridian edge B=B r +B l Above, the four corner positions of the image matrix are B (e.g., Figure 5 The regions enclosed by the edges of the image (as shown) constitute four binary images along the meridian region J. k (like Figure 6 (As shown), k∈[1,n], n=4 is the number of independent meridian regions;

[0055] A3. Meridian region detection; J in the binary image of each meridian region. k Calculate the axis l of the meridian region k (like Figure 7 (As shown); then apply the Hough circle detection algorithm to image I to obtain the circle detection binary image B. R And record B R The center pixel coordinates p of each circle's edge ii∈[1,n r ], n r It is the number of circular edges; B R In, satisfying p i To axis l k Circles with a distance greater than d are removed to obtain the binary image B of the broken yarn. k d = [1, 50] pixels;

[0056] The method for calculating the axis of the meridian region is as follows: based on the results of Hough edge detection, the binary image J of the meridian region is obtained. k Given the slopes and intercepts of the two straight edges in the diagram, denoted as (m1, b1) and (m2, b2), calculate the mean value of both to determine the axis l. k The slope and intercept (m) l ,b l );

[0057] A4. Yarn breakage statistical determination; statistical analysis of binary images of yarn breaks. k The total number N of pixels whose gradient in the longitudinal image is greater than the threshold t2 at positions where the value is 1. k If N k If the value is greater than the set threshold t3, it is determined that a yarn breakage has been detected (e.g., Figure 8 As shown in the blue mark);

[0058] The threshold t2 is set according to a certain proportion following the value of t1, and the calculation formula is as follows:

[0059] t2=γ2t1

[0060] Where γ2 = 0.5 is the gradient sensitivity parameter;

[0061] Since a broken yarn inevitably wraps around the warp beam, forming an arc-shaped protrusion, it appears as an arc-shaped edge in the image; therefore, the number of pixels occupied by the broken yarn must be larger than the binary image B of the arc-shaped edge. r The threshold t3 is a multiple of the pixels occupied by the middle edge, therefore it can be determined based on B. r The number of pixels with a median value of 1 is set to avoid misclassifying excessively small detected areas as yarn breaks. The calculation formula is as follows:

[0062]

[0063] Where γ3 = 10 is the region sensitivity parameter. H and W are the pixel height and width of image I, respectively, and (x, y) represents the pixel coordinates of the image.

[0064] Comparative example:

[0065] In actual production scenarios, without specially designated light sources, the information contained in the warp beam images of sizing machines captured by industrial cameras is quite complex, and areas of yarn breakage are difficult to identify. Figure 2 As shown. To demonstrate the beneficial effects of the present invention, the inventors attempted to use conventional image processing algorithms, but found it difficult to obtain... Figure 2 Effective detection results can be obtained in various ways; for example, the global Otsu thresholding algorithm and the region adaptive thresholding algorithm yield results such as... Figure 9 , 10 As shown, it is difficult to achieve effective detection and yarn breakage judgment.

[0066] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting yarn breakage images in the warp beam area of ​​a sizing machine, characterized in that, include: A1. Image Acquisition; Acquiring images of the warp beams of the sizing machine. ; A2. Warp beam area positioning; from the warp beam image of the sizing machine. In the process, straight edges along the meridian axis are extracted, circular edges along the meridian axis are extracted, and finally, binary images of independent meridian axis regions are extracted. , , It represents the number of independent meridian regions; A3. Meridian region detection; binary image of each meridian region. Calculate the axis of the meridian region Then examine the image. The Hough circle detection algorithm is used to obtain a binary image of the circle detection. and record The center pixel coordinates of each circle's edge , , It is the number of circular edges; In the middle, satisfy To the axis The distance is greater than The circle is removed to obtain the binary image of the broken yarn. , Pixel; A4. Determination of yarn breakage statistics; Statistical binary image of yarn breakage At positions where the value is 1, the gradient of the longitudinal image is greater than the threshold. The total number of pixels ,like Greater than the set threshold If the threshold is reached, it is determined that a broken yarn has been detected; The calculation formula is: in These are gradient sensitivity parameters; ; threshold The calculation formula is: in, It is a regional sensitivity parameter; Step A2 specifically involves: Step 1: Image of the warp beam of the sizing machine The Hough line detection algorithm is used to extract the edges of straight lines within the 125°~145° and 35°~55° angle ranges from the meridian image, and these edges are recorded as binary images of the line edges. ; Step 2: Process the image The Hough circle detection algorithm is used to obtain a binary image of the circle detection. Location image Gradient less than The pixel position, then Setting the values ​​at these pixel locations to 0 yields a binary image of the arc edge. ;in, The calculation formula is: It is an image gradient plot, These are the image's pixel height and width, respectively. Represents the pixel coordinates of the image; Step 3: Binary image at the edge of the meridian. Above, the four corner positions of the image matrix are... The regions enclosed by the middle edge constitute four binary images along the meridian. , , It represents the number of independent meridian regions.

2. The method according to claim 1, characterized in that, In step A3, the calculation The method is as follows: Based on the results of Hough edge detection, a binary image of the meridian region is obtained. The slopes and intercepts of the two straight lines in the figure are respectively Calculate the axis value for both. slope and intercept .

3. The method according to claim 1, characterized in that, The .

4. The method according to claim 1, characterized in that, In step A1, image acquisition involves acquiring images of the warp beam of the sizing machine using a warp beam image acquisition device. .

5. The method according to claim 1, characterized in that, In step A1, the image acquisition involves placing the image acquisition device between the two left and right channel steel columns and the two upper and lower cross braces, and acquiring the warp axis image along the warp axis direction.

6. The method according to claim 1, characterized in that, In step A1, the image acquisition is performed using a CCD industrial camera.

7. A yarn breakage image detection system for the warp beam area of ​​a sizing machine, characterized in that, The system includes an image acquisition device, a warp beam area detection device, a warp beam area detection device, and a yarn breakage statistics and determination device; the yarn breakage image detection system for the warp beam area of ​​the sizing machine executes the method described in any one of claims 1-6.

Images