Machine vision-based wet drawing online detection system and method

Abstract

The invention provides a wet drawing online detection system and method based on machine vision, the detection system comprises a traction module, a light source module, a detection module, a take-up module and a control module, the light source module comprises a backlight plate, the detection module comprises a lifting support and a camera, and a monofilament enters the space between the camera and the backlight plate after passing through the traction module; and then the cable enters the take-up module. According to the method, machine vision is added between the wet drawing process and the take-up process to detect the monofilament, the monofilament diameter is calculated by adopting a combined processing mode of camera linear calibration, image median filtering and denoising, Otsu threshold segmentation and canny operator edge detection, the monofilament detection precision and detection efficiency are improved, manpower and time can be saved, and the production efficiency is improved. The production cost is reduced; and meanwhile, an alarm can be given in time and processing can be carried out when the phenomena of wire expansion and wire shrinkage occur in production, so that economic loss is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of wire drawing machines, and in particular to a wet drawing online detection system and method based on machine vision. Background Art

[0002] In the production process of steel cord, all types of wire rods must go through multiple processing and drawing processes, and finally be drawn into various specifications of monofilaments through wet drawing machines. The existing inspection of wet drawn monofilament semi-finished products is to manually use a micrometer to measure the seam head of the I-shaped wheel with a fixed length to check whether the semi-finished steel wire has expanded or shrunk. This inspection method has many problems and defects.

[0003] The existing inspection of semi-finished monofilaments is basically carried out by manual sampling and manual measurement. This method is not only time-consuming and labor-intensive, but also has low measurement accuracy and is greatly affected by human factors. At the same time, human inspection errors may cause great economic losses. Moreover, if the I-shaped wheel with a fixed length expands or shrinks the wire during the wire drawing process, even if the expansion or shrinkage is found at the final seam, it is impossible to locate at what meter the unqualified phenomenon occurs, resulting in a great waste of semi-finished monofilaments.

[0004] Patent CN201220235468.5 provides a wire drawing machine with an online detection device, which uses a light source transmitter and a reflected light receiver to detect the unqualified wire diameter of the wire online and stop the wire drawing machine to ensure the quality of the wire. This method is to measure the width of the light shadow to check the actual wire diameter, but the width of the shadow depends on multiple factors, including the angle and intensity of the light, the position of the light source and the shape of the wire. Steel wires of the same width may produce shadows of different widths under different light conditions, and steel wires of different widths may also produce shadows of similar widths under the same light conditions. The formation of a shadow is a complex process, and its size and shape are affected by multiple factors, not limited to the width of the wire. Therefore, the width of the shadow is more suitable for expressing information such as the proportion or rough shape of the wire, rather than the specific diameter of the wire.

[0005] Patent CN201710947201.6 provides an online detection method and device for surface defects of wire ropes based on machine vision, which uses three industrial cameras to detect defects of finished wire ropes from three different angles at the same time. This method not only has a higher quality of finished products, but also the quality of the images acquired by the three industrial cameras will be seriously affected by the environment in which the finished wire rope is located when acquiring images, thereby seriously affecting the accuracy of detection; in addition, this defect detection is an inspection of the finished wire rope. Once the semi-finished monofilament is formed into a finished product, it means an increase in cost, waste of production capacity and resources.

[0006] Therefore, it is necessary to improve the online detection technology of steel wire. Summary of the invention

[0007] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a wet pull online detection system and method based on machine vision to improve detection accuracy and detection efficiency, save costs and reduce economic losses.

[0008] To achieve the above object, the present invention adopts the following technical solutions: According to a first aspect of the present invention, a machine vision-based wet-pull online detection system is provided, wherein the machine vision-based wet-pull online detection system comprises a traction module, a light source module, a detection module, a wire-receiving module and a control module.

[0009] The light source module and the detection module are respectively arranged between the traction module and the wire-taking module, and the light source module is located below the detection module; the control module is arranged on the wire-taking module, and the traction module, light source module, detection module, and wire-taking module are respectively electrically connected to the control module.

[0010] The light source module includes a backlight panel, which is horizontally arranged between the traction module and the wire-receiving module; the detection module includes a lifting bracket and a camera, the lifting bracket is arranged on the wire-receiving module, the camera is liftably arranged on the lifting bracket, and the field of view of the camera is toward the backlight panel.

[0011] The monofilament passes through the traction module and enters between the camera and the backlight plate, and then enters the take-up module.

[0012] The camera is used to capture single-wire images and upload them to the control module; the control module is used to process the received single-wire images, calculate the wire diameter of the single wire, and compare the calculated wire diameter with a preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is judged to be qualified; if the wire diameter exceeds the preset wire diameter range, the control module controls the machine tool to stop and alarm.

[0013] Preferably, the traction module comprises a wire passing wheel and a traction wheel, and the monofilament passes through the wire passing wheel and the traction wheel in sequence.

[0014] Preferably, the backlight panel is an LED backlight panel.

[0015] Preferably, the wire taking-up module comprises a wire taking-up frame, a meter wheel and a wire taking-up wheel. The meter wheel is arranged on the side of the wire taking-up frame facing the traction module, and the wire taking-up wheel is arranged in the wire taking-up frame. After the single wire is detected, it passes through the meter wheel and then is wound onto the wire taking-up wheel.

[0016] Preferably, the control module includes a controller and a host computer, and the host computer is electrically connected to the controller and the camera respectively.

[0017] The host computer is used to perform median filtering and denoising on the received single-filament image; perform Otsu threshold method segmentation on the image after median filtering and denoising to obtain a binary image of the single-filament image; use the canny operator edge detection algorithm to scan and extract edge coordinate points of the binary image of the single-filament image; use the least squares method to fit the scanned and extracted edge coordinate points into two edge lines, extract the midpoint coordinates of the two edge lines, calculate the distance between the midpoint coordinates of the two edge lines to obtain the wire diameter of the single wire, and then transmit the wire diameter of the single wire to the controller.

[0018] The controller is used to compare the calculated wire diameter with a preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is determined to be qualified; if the wire diameter exceeds the preset wire diameter range, the machine tool is controlled to stop and an alarm is issued.

[0019] Preferably, the control module further comprises a light source switch, the light source switch is electrically connected to the controller and the backlight panel respectively, and the light source switch is used to control the switching and intensity of the backlight panel.

[0020] Preferably, the control module further comprises an alarm, and the alarm is electrically connected to the controller. When it is detected that the wire diameter exceeds a preset wire diameter range, the controller controls the alarm to issue an alarm prompt.

[0021] According to a second aspect of the present invention, a detection method of a wet pull online detection system based on machine vision is provided, comprising the following steps: Step S1: Perform linear calibration on the camera model to obtain the calibration coefficient ; Step S2: using a camera to collect a single-filament image and uploading it to a control module; Step S3: the control module performs median filtering and denoising on the received single-filament image; Step S4: segmenting the image after median filtering and denoising using the Otsu threshold method to obtain a binary image of the single-filament image; Step S5: using the canny operator edge detection algorithm to scan and extract edge coordinate points of the binary image of the single-filament image; Step S6: Use the least squares method to fit the edge coordinate points extracted by scanning into two edge lines, extract the midpoint coordinates of the two edge lines, and calculate the distance between the midpoint coordinates of the two edge lines. , then the wire diameter of the single wire ; Step S7: Compare the calculated wire diameter with the preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is determined to be qualified; if the wire diameter exceeds the preset wire diameter range, the machine tool is controlled to stop and an alarm is issued.

[0022] Preferably, the step of linearly calibrating the camera model in step S1 is: Step S1-1: Provide a calibration monofilament with a wire diameter of d, place it in the camera field of view at 0°, 15°, 30°, 45°, 60° and 90°, and measure it six times. The pixel sizes obtained by the measurement are recorded as , , , , , ; Step S1-2: The wire diameter d and the pixel size of the six measurements , , , , , Substitute into the calibration coefficient formula: , the calibration coefficient k is calculated.

[0023] Preferably, the median filter window in step S3 adopts a linear window.

[0024] Preferably, the canny operator edge detection algorithm in step S5 includes the following steps: Step S5-1: smoothing the image using a Gaussian filter; Step S5-2: Calculate the magnitude and direction of the gradient using finite differences of the first-order partial derivatives; Step S5-3: performing non-maximum suppression on the gradient amplitude; Step S5-4: Use a dual threshold algorithm to detect and connect edges.

[0025] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention adds machine vision to detect the monofilament between the wet pulling and winding processes, and adopts a combined processing method of camera linear calibration, image median filtering denoising, Otsu threshold segmentation and canny operator edge detection to calculate the wire diameter of the monofilament, thereby improving the detection accuracy and detection efficiency of the monofilament, saving manpower and time, and reducing production costs; at the same time, the phenomenon of wire expansion and shrinkage in production can be alarmed and processed in time, thereby reducing economic losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It is a schematic diagram of the structure of a wet-pull online detection system based on machine vision according to some embodiments of the present invention.

[0027] Figure 2 The present invention is a flowchart of a detection method of a wet-pull online detection system based on machine vision in some embodiments of the present invention.

[0028] Figure 3 The figure is a flow chart of the canny operator edge detection algorithm in some embodiments of the present invention.

[0029] In the figure, 10-traction module, 11-line wheel, 12-traction wheel, 20-light source module, 21-backlight panel, 30-detection module, 31-lifting bracket, 32-camera, 40-receiving module, 41-receiving frame, 42-meter wheel, 43-receiving wheel, 50-control module, 51-controller, 52-host computer, 53-light source switch, 54-alarm, 60-monofilament. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention. It can be understood that some technical means of the various embodiments described herein can be replaced or combined with each other without conflict.

[0031] In the description of the present invention, if there are terms such as "first", "second", etc., they are only used to distinguish the objects described, and do not have any order or technical meaning. Therefore, the objects defined as "first", "second", etc. may explicitly or implicitly include one or more of the objects. In addition, "one" or "a" and similar words do not indicate a quantity limitation, but indicate the existence of at least one, and "multiple" means not less than two.

[0032] In the description of the present specification, reference to "one embodiment" or "some embodiments" etc. means that one or more embodiments of the present invention include a specific feature, structure or characteristic described in conjunction with the embodiment. Therefore, the phrases "in one embodiment", "in some embodiments", "in some other embodiments", "in some other embodiments", etc. appearing in different places in this specification do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways.

[0033] According to a first aspect of the present invention, a wet pull online detection system based on machine vision is provided. Figure 1 In some embodiments, a wet-pull online detection system based on machine vision includes a traction module 10 , a light source module 20 , a detection module 30 , a wire collection module 40 and a control module 50 .

[0034] The light source module 20 and the detection module 30 are respectively arranged between the traction module 10 and the wire taking-up module 40, and the light source module 20 is located below the detection module 30. The control module 50 is arranged on the wire taking-up module 40, and the traction module 10, the light source module 20, the detection module 30, and the wire taking-up module 40 are respectively electrically connected to the control module 50.

[0035] Specifically, the light source module 20 includes a backlight panel 21, which is horizontally arranged between the traction module 10 and the wire-receiving module 40. The detection module 30 includes a lifting bracket 31 and a camera 32, the lifting bracket 31 is arranged on the wire-receiving module 40, and the camera 32 is lifted and lowered on the lifting bracket 31, and the field of view of the camera 32 is toward the backlight panel 21. It should be noted that the lifting bracket 31 can adopt products such as linear guides, cylinders, and electric telescopic rods in the prior art, as long as they can drive the camera 32 to rise and fall.

[0036] During production, the monofilament 60 passes through the traction module 10 and enters between the camera 32 and the backlight plate 21 , and then enters the take-up module 40 .

[0037] Among them, the camera 32 is used to collect the image of the single wire 60 and upload it to the control module 50; the control module 50 is used to process the received single wire image, calculate the wire diameter of the single wire 60, and compare the calculated wire diameter with the preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is judged to be qualified; if the wire diameter exceeds the preset wire diameter range, the control module 50 controls the machine tool to stop and alarm.

[0038] The present invention adds machine vision to detect the monofilament between the wet-pulling and wire-collecting processes, thereby improving the detection accuracy and efficiency of the monofilament, saving manpower and time, and reducing production costs; at the same time, when the wire expansion and shrinkage phenomena occur during production, an alarm can be issued in time and processed, thereby reducing economic losses.

[0039] In some preferred embodiments, the traction module 10 includes a wire passing wheel 11 and a traction wheel 12. The drawn monofilament 60 passes through the wire passing wheel 11 and the traction wheel 12 in sequence, and then enters the detection area of ​​the camera 32 for detection.

[0040] In some preferred embodiments, the backlight board 21 is an LED backlight board. The backlight board 21 is located below the monofilament 60 to be measured and can provide backlight for image acquisition, thereby improving the quality of the image.

[0041] In some preferred embodiments, the wire taking-up module 40 includes a wire taking-up frame 41, a meter wheel 42 and a wire taking-up wheel 43. The meter wheel 42 is disposed on the side of the wire taking-up frame 41 facing the traction module 10, and the wire taking-up wheel 43 is disposed in the wire taking-up frame 41. After the monofilament 60 is detected, it passes through the meter wheel 42 and is wound on the wire taking-up wheel 43. The meter wheel 42 can be a meter wheel product in the prior art, which is used to measure the length of the monofilament 60.

[0042] In some preferred embodiments, the control module 50 includes a controller 51 and a host computer 52 , and the host computer 52 is electrically connected to the controller 51 and the camera 32 , respectively.

[0043] Among them, the host computer 52 is used to perform median filtering and denoising on the received single-filament image; perform Otsu threshold method segmentation on the image after median filtering and denoising to obtain a binary image of the single-filament image; use the canny operator edge detection algorithm to scan and extract the edge coordinate points of the binary image of the single-filament image; use the least squares method to fit the scanned and extracted edge coordinate points into two edge lines, and extract the midpoint coordinates of the two edge lines, calculate the distance between the midpoint coordinates of the two edge lines to obtain the wire diameter of the single wire, and then transmit the wire diameter of the single wire to the controller 51.

[0044] The controller 51 is used to compare the calculated wire diameter with the preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is determined to be qualified; if the wire diameter exceeds the preset wire diameter range, the machine tool is controlled to stop and an alarm is issued.

[0045] Furthermore, the controller 51 may be a controller such as a PLC (Programmable Logic Controller), the host computer 52 may be a Matlab platform, and the camera 32 may be a CMOS (Complementary Metal Oxide Semiconductor) industrial camera.

[0046] In some preferred embodiments, the control module 50 further includes a light source switch 53 , which is electrically connected to the controller 51 and the backlight panel 21 , respectively. The light source switch 53 is used to control the switching and intensity of the backlight panel 21 .

[0047] In other preferred embodiments, the control module 50 further includes an alarm 54, which is electrically connected to the controller 51. When the wire diameter is detected to be beyond the preset wire diameter range, the controller 51 controls the alarm 54 to issue an alarm prompt. In application, the alarm 54 can be a product such as an audible and visual alarm, which can issue an audible and visual alarm prompt so that the operator can handle the fault in time.

[0048] According to the second aspect of the present invention, a detection method of a wet pull online detection system based on machine vision is provided. Figure 2 In some embodiments, the detection method comprises the following steps: Step S1: Perform linear calibration on the camera model to obtain the calibration coefficient ; Step S2: using a camera to collect a single-filament image and uploading it to a control module; Step S3: the control module performs median filtering and denoising on the received single-filament image; Step S4: segmenting the image after median filtering and denoising using the Otsu threshold method to obtain a binary image of the single-filament image; Step S5: using the canny operator edge detection algorithm to scan and extract edge coordinate points of the binary image of the single-filament image; Step S6: Use the least squares method to fit the edge coordinate points extracted by scanning into two edge lines, extract the midpoint coordinates of the two edge lines, and calculate the distance between the midpoint coordinates of the two edge lines. , then the wire diameter of the single wire ; Step S7: Compare the calculated wire diameter with the preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is determined to be qualified; if the wire diameter exceeds the preset wire diameter range, the machine tool is controlled to stop and an alarm is issued.

[0049] The present invention adopts a combined processing method of camera linear calibration, image median filtering denoising, Otsu threshold segmentation and canny operator edge detection to calculate the wire diameter of the single wire, thereby improving the detection accuracy and detection efficiency of the single wire, saving manpower and time, and reducing production costs; at the same time, when the wire expansion and shrinkage phenomena occur in production, an alarm can be issued in time and processed, thereby reducing economic losses.

[0050] In some preferred embodiments, the step of linearly calibrating the camera model in step S1 is as follows: Step S1-1: Provide a calibration monofilament with a wire diameter of d, place it in the camera field of view at 0°, 15°, 30°, 45°, 60° and 90°, and measure it six times. The pixel sizes obtained by the measurement are recorded as , , , , , ; Step S1-2: The wire diameter d and the pixel size of the six measurements , , , , , Substitute into the calibration coefficient formula: , the calibration coefficient k is calculated.

[0051] Furthermore, in step S3, when filtering and denoising the image, the selection of the filter window is also very important, because the size of the filter window also affects the denoising effect to a great extent. A window that is too large may cause the edge information of the image to be lost, thereby affecting the quality of the image; but if the window is too small, the image denoising processing effect will not be obvious, and image information will also be lost. Therefore, when performing image filtering and denoising, a window of appropriate size should be selected according to the actual situation of the object being measured and the requirements of this system. Therefore, the present invention chooses to use a linear window, which can ensure a better filtering and denoising effect.

[0052] In some preferred embodiments, the canny operator edge detection algorithm in step S5 includes the following steps: Step S5-1: smoothing the image using a Gaussian filter; Step S5-2: Calculate the magnitude and direction of the gradient using finite differences of the first-order partial derivatives; Step S5-3: performing non-maximum suppression on the gradient amplitude; Step S5-4: Use a dual threshold algorithm to detect and connect edges.

[0053] Specifically, in one embodiment, the detection method of the wet pull online detection system based on machine vision includes the following steps: Step S1: Perform linear calibration on the camera model to obtain the calibration coefficient :Using a single wire with a wire diameter of 0.30mm, six measurements were made at 0°, 15°, 30°, 45°, 60° and 90° in the camera field of view. The measured pixel sizes were 1.488, 2.000, 1.500, 1.558, 1.748 and 1.068 respectively. Substituting the above data into the calibration coefficient formula: , the calibration coefficient k=0.284 is calculated.

[0054] Step S2: Use a camera to capture single-filament images and upload them to a host computer.

[0055] Step S3: performing median filtering to remove noise on the received single-filament image; wherein the median filtering window is a linear window.

[0056] Step S4: performing Otsu threshold segmentation on the image after median filtering and denoising to obtain a binary image of a single silk thread.

[0057] Step S5: Use the Canny operator edge detection algorithm to scan and extract the edge coordinate points of the binary image of the single-filament image; the Canny operator is more accurate in locating the edge and has a stronger anti-interference ability. The detected edge has better continuity and no false edges. The Canny operator edge detection algorithm includes the following steps: Step S5-1: smoothing the image using a Gaussian filter; Step S5-2: Calculate the magnitude and direction of the gradient using finite differences of the first-order partial derivatives; Step S5-3: performing non-maximum suppression on the gradient amplitude; Step S5-4: Use a dual threshold algorithm to detect and connect edges.

[0058] Step S6: Use the least squares method to fit the edge coordinate points extracted by scanning into two edge lines, extract the midpoint coordinates of the two edge lines, and calculate the distance between the midpoint coordinates of the two edge lines. , then the wire diameter of the single wire ; Step S7: The host computer transmits the calculated wire diameter to the controller, and the controller compares the wire diameter with the preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is judged to be qualified; if the wire diameter exceeds the preset wire diameter range, the machine tool is controlled to stop and the alarm is controlled to sound an alarm.

[0059] In production, the above detection system and detection method are used to measure the monofilament five times, and the measurement results in Table 1 are obtained: Table 1 Single-filament measurement results frequency 1st time 2nd time 3rd 4th 5th Average value (mm) Actual value (mm) Measurement value (mm) 0.309 0.298 0.308 0.299 0.307 0.304 0.300 It should be noted that, during the production process, for monofilaments with a specification of 0.30 mm, the wire diameter is generally controlled within 0.30±0.01. As can be seen from Table 1, after adopting the detection system and method of the present invention, the error between the measured value and the actual value will basically not affect the qualified judgment of the monofilament, and no alarm will be issued for monofilaments with a wire diameter within the above range, thereby ensuring production efficiency and avoiding system errors.

[0060] In summary, the present invention provides a wet-pulling online detection system and method based on machine vision, which adds machine vision to detect the monofilament between the wet-pulling and wire-winding processes, and adopts a combined processing method of camera linear calibration, image median filtering denoising, Otsu threshold segmentation and canny operator edge detection to calculate the wire diameter of the monofilament, thereby improving the detection accuracy and efficiency of the monofilament, saving manpower and time, and reducing production costs; at the same time, the phenomenon of wire expansion and shrinkage in production can be alarmed and processed in time, thereby reducing economic losses.

[0061] The present invention has been described by the above-mentioned relevant embodiments, however, the above-mentioned embodiments are only examples for implementing the present invention. It must be pointed out that the disclosed embodiments do not limit the scope of the present invention. On the contrary, changes and modifications made without departing from the spirit and scope of the present invention are all within the scope of patent protection of the present invention.

Claims

1. A wet pull online detection system based on machine vision, characterized by: The wet-pull online detection system based on machine vision includes a traction module, a light source module, a detection module, a wire collection module and a control module; The light source module and the detection module are respectively arranged between the traction module and the wire-receiving module, and the light source module is located below the detection module; the control module is arranged on the wire-receiving module, and the traction module, the light source module, the detection module, and the wire-receiving module are respectively electrically connected to the control module; The light source module includes a backlight panel, which is horizontally arranged between the traction module and the wire-receiving module; the detection module includes a lifting bracket and a camera, the lifting bracket is arranged on the wire-receiving module, the camera is liftably arranged on the lifting bracket, and the field of view of the camera is toward the backlight panel; The monofilament passes through the traction module and enters between the camera and the backlight plate, and then enters the take-up module; The camera is used to collect single-filament images and upload them to the control module; The control module is used to process the received single-wire image, calculate the wire diameter of the single wire, and compare the calculated wire diameter with a preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is judged to be qualified; if the wire diameter exceeds the preset wire diameter range, the control module controls the machine tool to stop and alarm.

2. The wet pull online detection system based on machine vision according to claim 1, characterized in that: The traction module comprises a wire passing wheel and a traction wheel, and the monofilament passes through the wire passing wheel and the traction wheel in sequence.

3. The wet pull online detection system based on machine vision according to claim 1, characterized in that: The wire-taking module comprises a wire-taking frame, a meter wheel and a wire-taking wheel. The meter wheel is arranged on a side of the wire-taking frame facing the traction module. The wire-taking wheel is arranged in the wire-taking frame. After the single wire is detected, it passes through the meter wheel and then is wound onto the wire-taking wheel.

4. The wet pull online detection system based on machine vision according to claim 1, characterized in that: The control module includes a controller and a host computer, and the host computer is electrically connected to the controller and the camera respectively; The host computer is used to perform median filtering and denoising on the received single-filament image; perform Otsu threshold method segmentation on the image after median filtering and denoising to obtain a binary image of the single-filament image; use the canny operator edge detection algorithm to scan and extract edge coordinate points of the binary image of the single-filament image; use the least squares method to fit the scanned and extracted edge coordinate points into two edge lines, extract the midpoint coordinates of the two edge lines, calculate the distance between the midpoint coordinates of the two edge lines to obtain the wire diameter of the single wire, and then transmit the wire diameter of the single wire to the controller; The controller is used to compare the calculated wire diameter with a preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is determined to be qualified; if the wire diameter exceeds the preset wire diameter range, the machine tool is controlled to stop and an alarm is issued.

5. The wet pull online detection system based on machine vision according to claim 4, characterized in that: The control module also includes a light source switch, which is electrically connected to the controller and the backlight panel respectively, and is used to control the switch and intensity of the backlight panel.

6. The wet pull online detection system based on machine vision according to claim 4, characterized in that: The control module also includes an alarm, which is electrically connected to the controller. When it is detected that the wire diameter exceeds a preset wire diameter range, the controller controls the alarm to issue an alarm prompt.

7. A detection method of a wet-pull online detection system based on machine vision according to any one of claims 1 to 6, characterized in that: The following steps are involved: Step S1: Perform linear calibration on the camera model to obtain the calibration coefficient ; Step S2: using a camera to collect a single-filament image and uploading it to a control module; Step S3: the control module performs median filtering and denoising on the received single-filament image; Step S4: segmenting the image after median filtering and denoising using the Otsu threshold method to obtain a binary image of the single-filament image; Step S5: using the canny operator edge detection algorithm to scan and extract edge coordinate points of the binary image of the single-filament image; Step S6: Use the least squares method to fit the edge coordinate points extracted by scanning into two edge lines, extract the midpoint coordinates of the two edge lines, and calculate the distance between the midpoint coordinates of the two edge lines. , then the wire diameter of the single wire ; Step S7: Compare the calculated wire diameter with the preset wire diameter range. If the wire diameter is within the preset wire diameter range, the single wire is determined to be qualified; if the wire diameter exceeds the preset wire diameter range, the machine tool is controlled to stop and an alarm is issued.

8. The detection method of the wet-pull online detection system based on machine vision according to claim 7, characterized in that: The steps of linear calibration of the camera model in step S1 are: Step S1-1: Provide a calibration monofilament with a wire diameter of d, place it in the camera field of view at 0°, 15°, 30°, 45°, 60° and 90°, and measure it six times. The pixel sizes obtained by the measurement are recorded as , , , , , ; Step S1-2: The wire diameter d and the pixel size of the six measurements , , , , , Substitute into the calibration coefficient formula: , the calibration coefficient k is calculated.

9. The detection method of the wet-pull online detection system based on machine vision according to claim 7, characterized in that: The median filter window in step S3 adopts a linear window.

10. The detection method of the wet-pull online detection system based on machine vision according to claim 7, characterized in that: The canny operator edge detection algorithm in step S5 includes the following steps: Step S5-1: smoothing the image using a Gaussian filter; Step S5-2: Calculate the magnitude and direction of the gradient using finite differences of the first-order partial derivatives; Step S5-3: performing non-maximum suppression on the gradient amplitude; Step S5-4: Use a dual threshold algorithm to detect and connect edges.

Images