Image acquisition device and fabric defect real-time detection system for circular knitting machines
Through multi-camera arrangement around the rotation axis and cloud processing, the stability and cost problems of the image acquisition system in the circular knitting machine are solved, low-cost and high-quality fabric defect detection is achieved, and detection efficiency and accuracy are improved.
Patent Information
- Application Number
- CN202010575323.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2020-06-22
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2040-06-22
AI Technical Summary
The image acquisition system in the existing circular knitting machine is difficult to stabilize image acquisition due to relative motion, and it is difficult to provide high-quality images, which affects the effect of subsequent defect detection and is costly.
An image acquisition device arranged around the rotation axis is adopted for multi-camera, the camera rotates synchronously with the knitting machine, and a fish-eye wide-angle lens and lighting mechanism are used, combined with cloud image processing modules and wireless communications to realize real-time image acquisition and defect recognition.
It reduces the difficulty and cost of image acquisition, improves image quality and stability, reduces hardware and resource configuration requirements, and realizes low-cost and high-quality real-time detection of fabric defects.
Smart Images

Figure CN111812108B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of textile equipment, and in particular to an image acquisition device for a circular knitting machine and a real-time fabric defect detection system and a detection method. Background Art
[0002] Circular knitting machine, also known as large circular knitting machine, circular weft knitting machine or circular weft knitting machine, has the advantages of multiple loop forming systems (number of routes), high speed, high output, fast pattern change, good fabric quality, fewer processes, and strong product adaptability. It is widely used in the textile industry.
[0003] When the circular knitting machine is working, the yarn placed on the creel enters the machine head, which weaves it through the needle head and weaves it into a cylindrical fabric, which is then collected by the receiving roller. In the process of producing fabrics, due to some failures of the creel, machine head or other parts, defects often occur on the fabric, such as attached Figure 4 The holes and attached Figure 5 The horizontal stripes / flying ammonia or bad needles / spoiled needles shown. The formation of defects will affect the quality of the fabric, and if the faults causing the defects cannot be eliminated in time, they will continue to form defects in subsequent production, causing serious losses. Therefore, it is necessary to conduct real-time inspection of the produced fabrics during the production process of the circular knitting machine to find problems in time and reduce losses.
[0004] In the prior art, many systems or devices have been developed for fabric defect detection in circular knitting machines, as well as related algorithms for defect recognition in images. However, existing image acquisition systems for fabric defect detection in circular knitting machines usually use a fixed camera setting. When the circular knitting machine is in production, the fabric rotates along with the circular knitting machine, and the fixed camera captures images of the fabric in production for defect recognition. Due to the relative motion between the image acquisition device and the fabric, expensive high-speed cameras are required to ensure the good quality of the captured fabric images. Even so, due to the inherent challenges of capturing moving objects, similar image acquisition systems still find it difficult to provide stable, high-quality images, and also pose a great challenge to subsequent defect detection algorithms. Summary of the invention
[0005] The object of the present invention is to provide a low-cost, stable, high-quality image acquisition device and a corresponding defect detection system to solve at least one of the above-mentioned problems.
[0006] In order to achieve the above object, the present invention adopts the following technical solutions:
[0007] A first aspect of the present invention provides an image acquisition device for a circular knitting machine, comprising:
[0008] A plurality of cameras, the plurality of cameras being arranged circumferentially and spaced apart around a rotating shaft of the circular knitting machine, for collecting real-time images of the fabric being produced inside or outside the circular knitting machine;
[0009] A control transmission module, which is electrically connected to the plurality of cameras, and is used to control the shooting action of the cameras, and store / transmit image data collected by the cameras;
[0010] And, a driving / transmission mechanism, which is used to drive or bring the camera so that it rotates synchronously with the circular knitting machine when the circular knitting machine is running.
[0011] Furthermore, the lens of the camera faces the area below the needle of the circular knitting machine, and the upper edge of the shooting area covers the needle.
[0012] Furthermore, the camera lens adopts a fisheye wide-angle lens.
[0013] Optionally, the cameras are configured to be three evenly spaced around the rotating axis of the circular knitting machine, and the camera lens adopts a fisheye wide-angle lens with a field of view angle greater than 120°.
[0014] Optionally, the cameras are arranged as two cameras evenly spaced around the rotating axis of the circular knitting machine, and the camera lens adopts a fisheye ultra-wide-angle lens with a field of view angle greater than 180°.
[0015] Furthermore, the image acquisition device also includes a lighting mechanism, which includes a plurality of light sources adapted to the camera and used to provide auxiliary light when the camera is shooting.
[0016] Furthermore, the driving / transmission mechanism includes a conductive slip ring installed on the support shaft of the circular knitting machine, and a transmission rod connecting the rotating part of the circular knitting machine and the conductive slip ring, and the plurality of cameras are connected to the conductive slip ring via a camera mounting bracket.
[0017] Furthermore, the control transmission module and the camera are synchronously connected to the conductive slip ring and receive power.
[0018] A second aspect of the present invention provides a real-time fabric defect detection system for a circular knitting machine, comprising the image acquisition device as described in the first aspect above, and further comprising:
[0019] The image processing module is communicatively connected to the control transmission module, and is used to receive and process image data collected by the camera based on a preset defect recognition algorithm, and to issue an alarm message when a defect is detected.
[0020] Furthermore, the image processing module is configured in the cloud, including a defect recognition algorithm and corresponding cloud computing power and cloud storage.
[0021] Furthermore, the image processing module and the control transmission module are remotely connected via wireless communication, and the wireless communication method includes but is not limited to WIFI or 3G / 4G / 5G.
[0022] Furthermore, the real-time fabric defect detection system also includes a marker provided on the needle of the circular knitting machine, and the marker is arranged at intervals on the needle of the circular knitting machine. The image processing module is also used to identify the position of the marker in the image captured by the camera, and then locate the defect position or the faulty needle.
[0023] Furthermore, the real-time fabric defect detection system also includes an alarm module, which is used to receive the alarm information sent by the image processing module and send out an audio / visual alarm signal.
[0024] Furthermore, the alarm module is integrated on the alarm module at the top end of the circular knitting machine support shaft.
[0025] The third aspect of the present invention provides a detection method based on the real-time fabric defect detection system according to the second aspect, comprising the following steps:
[0026] S1, the camera in the image acquisition device rotates synchronously with the circular knitting machine when the circular knitting machine is running and acquires real-time images of the fabric in production inside or outside the circular knitting machine;
[0027] S2, the image captured by the camera is transmitted to the image processing module via the control transmission module, the image processing module processes the image data based on a preset defect recognition algorithm, and issues an alarm message when a defect is detected;
[0028] The camera collects real-time images of fabrics being produced inside or outside the circular knitting machine by interval shooting, and the interval time t≤L / s; wherein L is the length of the fabric produced by the circular knitting machine per unit time, and s is the field of view height of the camera in the vertical direction.
[0029] Furthermore, in step S2, the preset defect recognition algorithm is an intelligent defect recognition algorithm based on deep learning.
[0030] The beneficial technical effects of the present invention are as follows:
[0031] The image acquisition device for a circular knitting machine of the present invention adopts a multi-camera-coaxial rotation technical solution, that is, the camera and the control transmission module rotate coaxially with the circular knitting machine, thereby achieving a relative static state between the camera and the fabric to be detected, greatly reducing the difficulty of image acquisition, and improving the image acquisition quality and stability. At the same time, due to the use of the coaxial rotation technology, the index requirements for the image acquisition hardware are greatly reduced, thereby greatly reducing the hardware cost of the image acquisition system, and having good practical value.
[0032] Based on the above-mentioned image acquisition device, the real-time fabric defect detection system for circular knitting machines of the present invention can greatly reduce the installation requirements of the entire system on the site and save on-site space by configuring the image processing module in the cloud. At the same time, the rich computing power and resources in the cloud can be reasonably utilized and dynamically allocated, further reducing the operating cost of the entire system while ensuring the detection effect.
[0033] In the detection method based on the above-mentioned real-time fabric defect detection system of the present invention, the camera adopts an interval shooting mode, and a reasonable shooting interval time is set based on the production parameters of the circular knitting machine, thereby minimizing the number of images taken while ensuring complete detection of the produced fabric, thereby reducing the calculation amount of the image processing module, and further optimizing the configuration requirements and operating costs of related resources. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 The figure is a schematic diagram of the composition of an embodiment of an image acquisition device for a circular knitting machine according to the present invention.
[0035] Figure 2 Schematic diagram of camera arrangement in an embodiment of an image acquisition device for a circular knitting machine of the present invention.
[0036] Figure 3 The figure is a schematic diagram of the working process of an embodiment of the real-time fabric defect detection system for a circular knitting machine of the present invention.
[0037] Figure 4 Schematic diagram of a hole defect on a fabric to be inspected, generated for a circular knitting machine.
[0038] Figure 5 Schematic diagram of bad stitches / spoiled stitches and horizontal stripes / flying ammonia defects on the fabric to be inspected, generated for a circular knitting machine. DETAILED DESCRIPTION
[0039] In order to further understand the present invention, preferred embodiments of the present invention are described below in conjunction with examples. However, it should be understood that these descriptions are only for further illustrating the features and advantages of the present invention, rather than limiting the claims of the present invention.
[0040] like Figure 1 As shown, an illustrative embodiment of the present invention provides an image acquisition device for a circular knitting machine.
[0041] The circular knitting machine comprises a housing 101, a head portion 102 located at the upper portion of the housing, a winding device 103 located inside the housing and mounted on a rotating platform, and a support shaft 104 located at the center of the housing. In addition, it also comprises a yarn rack for placing yarns, etc., which are not shown in the figure for the sake of simplicity. Since the structure of the circular knitting machine is well known to those skilled in the art, only a schematic description is provided here without detailed description.
[0042] When the circular knitting machine is in operation, the head part 102 and the winding device 103 rotate synchronously, the yarn on the creel enters the head part, is woven into a tubular fabric by the circularly arranged textile pillows of the head part 102, and is further wound by the winding device.
[0043] In order to monitor the fabric being produced in real time and promptly discover defects on the fabric, the image acquisition device in this embodiment is equipped with a plurality of cameras 106, and the plurality of cameras are arranged at intervals around the rotating shaft of the circular knitting machine. When the circular knitting machine is running, the plurality of cameras are driven or driven by the driving / transmission mechanism to rotate synchronously with the circular knitting machine to collect real-time images of the fabric being produced inside or outside the circular knitting machine.
[0044] By adopting the above-mentioned multi-camera-coaxial rotation technical solution, a relative static state between the camera and the fabric to be inspected can be achieved, thereby greatly reducing the difficulty of image acquisition, improving the image acquisition quality and stability, and correspondingly reducing the hardware cost of the image acquisition system.
[0045] As a preferred embodiment, the lens of the camera 106 is a fisheye wide-angle lens.
[0046] More preferably, Figure 2 As shown, the cameras 106 are arranged in three evenly spaced arrangements around the rotating shaft of the circular knitting machine, and the camera lenses use fisheye wide-angle lenses with a field of view greater than 120°. Alternatively, the cameras 106 are arranged in two evenly spaced arrangements around the rotating shaft of the circular knitting machine, and the camera lenses use fisheye ultra-wide-angle lenses with a field of view greater than 180°.
[0047] Generally, when capturing images of objects, fisheye wide-angle lenses can obtain a larger field of view, but they can also cause image distortion and other effects, causing trouble for subsequent image data processing. Therefore, in existing vision-based image detection devices or systems, fisheye lenses with large or ultra-wide field of view are consciously avoided.
[0048] However, in the technical solution of this embodiment, since the object to be photographed is a tubular fabric 111 produced by a circular knitting machine, the use of a fisheye wide-angle lens can overcome the adverse effects caused by image distortion when photographing a flat object, and can better adapt to the photographing of tubular fabrics in a circular curved surface state. The image obtained by the photographing can make the defects appear more clearly, further improving the accuracy of subsequent defect recognition and detection.
[0049] As a preferred embodiment, the lens of the camera 106 faces the area below the needle of the circular knitting machine, and the upper edge of the shooting area covers the needle. With this setting, it can be ensured that the camera field of view can cover the fabric to be inspected and detect defects as early as possible, and it can be more convenient to locate the defect or faulty needle when the defect is detected. This part will be further explained in the subsequent embodiments.
[0050] As a further preferred embodiment, the image acquisition device in this embodiment may also be equipped with a lighting mechanism, which includes a plurality of light sources adapted to the camera and used to provide auxiliary light when the camera is shooting.
[0051] In one embodiment, the above-mentioned lighting mechanism can be a plane light source or an annular light source for illuminating the fabric to be inspected, and can be arranged to be a plurality of light sources matching the camera, which is not specifically limited here.
[0052] The image acquisition device in this embodiment also includes a control transmission module 107, which is electrically connected to the plurality of cameras 106, and is used to control the shooting action of the camera, store and transmit the image data acquired by the camera, and perform some pre-processing on the image data acquired by the camera.
[0053] Specifically, the control transmission module may be a host computer equipped with a processor, a memory, a data transmission module, a housing and related interfaces, and its form is not specifically limited.
[0054] In one embodiment, Figure 1 and Figure 2 As shown, the driving / transmission mechanism for driving or driving the camera to rotate synchronously with the circular knitting machine includes a conductive slip ring 109 mounted on the circular knitting machine support shaft 104, and a transmission rod 110 connecting the rotating part of the circular knitting machine with the conductive slip ring 109. At the same time, the camera 106 is connected to the conductive slip ring 109 through a camera mounting bracket 112. Those skilled in the art should understand that the specific shape of the mounting bracket can be designed according to actual conditions, and will not be described in detail here.
[0055] In a further preferred embodiment, the control transmission module 107 and the camera 106 are synchronously connected to a conductive slip ring 109 through the same or mutually connected bracket, and are powered by the conductive slip ring 109 .
[0056] By adopting the above design, the conductive slip ring 109 and the transmission rod 110 installed on the circular knitting machine bracket shaft 104 can not only achieve the synchronous rotation of the camera and the rotating part of the circular knitting machine, but also cleverly solve the power supply problem of the rotating camera and the control transmission module, thereby ensuring the stable working state of the camera and the control transmission module.
[0057] Another embodiment of the present invention provides a real-time fabric defect detection system for a circular knitting machine, which includes the image acquisition device as described in the above embodiment; and also includes a corresponding image processing module 108 (see the attached Figure 1 ), the image processing module 108 is communicatively connected to the control transmission module 107, and is used to receive and process the image data collected by the camera based on a preset defect recognition algorithm, and to issue an alarm message when a defect is detected.
[0058] In one embodiment, the above-mentioned image processing module is configured in the cloud, including a defect recognition algorithm and corresponding cloud computing power and cloud storage.
[0059] Correspondingly, the image processing module and the control transmission module are remotely connected via wireless communication, and the wireless communication method includes but is not limited to WIFI or 3G / 4G / 5G.
[0060] By configuring the image processing module in the cloud, the installation requirements of the entire system on the site can be greatly reduced, saving on-site space. At the same time, the rich computing power and resources in the cloud can be reasonably utilized and dynamically allocated, further reducing the operating costs of the entire system while ensuring the detection effect.
[0061] In a further preferred embodiment, the fabric defect real-time detection system further comprises a marker disposed on the needle of the circular knitting machine. The marker is disposed on the needle of the circular knitting machine at intervals, so that the image processing module can locate the defect position or the faulty needle by identifying the position of the marker in the image captured by the camera.
[0062] In a specific embodiment, the marker can be a color mark or shape mark that is sensitive to detection, which can be set on the needle of the circular knitting machine at certain angles and assigned a certain serial number to facilitate recognition by the image processing module, and then locate the detected defect position based on the mark, so that the staff can quickly find the defect or faulty needle.
[0063] In one embodiment, the above-mentioned fabric defect real-time detection system also includes an alarm module, which is used to receive the alarm information sent by the image processing module and send out an audio / visual alarm signal.
[0064] In a specific implementation, the alarm module mentioned above can be integrated into the alarm light module 105 configured as a standard part at the top end of the circular knitting machine support shaft 104.
[0065] Another exemplary embodiment of the present invention further provides a detection method of the above fabric defect real-time detection system, see the attached Figure 3 , which comprises the following steps:
[0066] When the circular knitting machine is turned on, the camera initialization configuration operation is performed;
[0067] The camera then rotates synchronously with the circular knitting machine and captures real-time images of the fabric being produced inside or outside the circular knitting machine;
[0068] The image captured by the camera is transmitted to the image processing module through the control transmission module. The image processing module processes the image data based on the preset defect recognition algorithm and issues an alarm message when a defect is detected;
[0069] The worker makes a manual judgment to determine whether to shut down the machine; if shutdown is not necessary, the above detection steps are continued; if shutdown is required, the above detection steps are performed again after correction and restart.
[0070] In the above process, the camera collects real-time images of the fabric in production inside or outside the circular knitting machine by interval shooting, thereby minimizing the number of images taken and reducing the calculation amount of the image processing module.
[0071] In one embodiment, the interval time of camera shooting is t≤L / s. Wherein, L is the length of fabric produced by the circular knitting machine per unit time, and s is the height of the camera's field of view in the vertical direction. As a preferred solution, t=L / s or slightly less than L / s, which can be set according to actual conditions to achieve a balance between continuous and complete shooting detection and reducing the number of shots.
[0072] In one embodiment, the preset defect recognition algorithm may be an intelligent defect recognition algorithm based on deep learning. There are many public examples of such algorithms in the prior art, and no detailed description or specific limitation is given here.
[0073] It should be noted that, due to the use of the image acquisition device of the multi-camera-coaxial rotation scheme of the present invention, the complexity and development cost of the defect recognition algorithm used can be greatly reduced, thereby greatly improving and optimizing the detection efficiency, accuracy, and detection cost.
[0074] The above embodiments are only used to help understand the method and core idea of the present invention. It should be noted that, for those skilled in the art, several improvements and modifications can be made to the present invention without departing from the principles of the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.
Claims
1. An image acquisition device for a circular knitting machine, It is characterized in that include: A plurality of cameras, the plurality of cameras are arranged at intervals around the rotating shaft of the circular knitting machine, the lenses of the cameras are fisheye wide-angle lenses, and the field of view of the plurality of cameras covers the entire circular area, the lenses of the cameras are directed toward the area below the needles of the circular knitting machine, and the upper edge of the shooting area covers the needles, and are used to collect real-time images of fabrics being produced inside the circular knitting machine; A control transmission module, which is electrically connected to the plurality of cameras, and is used to control the shooting action of the cameras, and store / transmit image data collected by the cameras; And, a driving or transmission mechanism, which is used to drive or drive the camera so that it rotates synchronously with the circular knitting machine when the circular knitting machine is running; the driving or transmission mechanism includes a conductive slip ring installed on the support shaft of the circular knitting machine, and a transmission rod connecting the rotating part of the circular knitting machine and the conductive slip ring, and the plurality of cameras are connected to the conductive slip ring through a camera mounting bracket.
2. The image acquisition device for a circular knitting machine according to claim 1, It is characterized in that The cameras are arranged in three evenly spaced circumferential directions around the rotating shaft of the circular knitting machine, and the camera lenses are fisheye wide-angle lenses with a field of view angle greater than 120°.
3. The image acquisition device for a circular knitting machine according to claim 1, It is characterized in that The cameras are arranged in two evenly spaced circumferentially around the rotating shaft of the circular knitting machine, and the camera lenses adopt fisheye ultra-wide-angle lenses with a field of view angle greater than 180°.
4. The image acquisition device for a circular knitting machine according to claim 1, It is characterized in that It also includes a lighting mechanism, which includes a plurality of light sources adapted to the camera and is used to provide auxiliary light when the camera is shooting.
5. The image acquisition device for a circular knitting machine according to claim 1, It is characterized in that The control transmission module and the camera are synchronously connected to the conductive slip ring and receive power.
6. A real-time fabric defect detection system for circular knitting machines, It is characterized in that The device comprises the image acquisition device according to any one of claims 1 to 5; and further comprises: The image processing module is communicatively connected to the control transmission module, and is used to receive and process image data collected by the camera based on a preset defect recognition algorithm, and to issue an alarm message when a defect is detected.
7. The real-time fabric defect detection system for a circular knitting machine according to claim 6, It is characterized in that The image processing module is configured in the cloud, including a defect recognition algorithm and corresponding cloud computing power and cloud storage.
8. The fabric defect real-time detection system for a circular knitting machine according to claim 6, It is characterized in that The image processing module and the control transmission module are remotely connected via wireless communication, and the wireless communication method includes WIFI or 3G or 4G or 5G.
9. The real-time fabric defect detection system for a circular knitting machine according to claim 7 or 8, It is characterized in that It also includes a marker arranged on the needle of the circular knitting machine, and the marker is arranged on the needle of the circular knitting machine at intervals. The image processing module is also used to identify the position of the marker in the image captured by the camera, and then locate the defect position or the faulty needle.
10. The fabric defect real-time detection system for a circular knitting machine according to claim 9, It is characterized in that It also includes an alarm module, which is used to receive the alarm information sent by the image processing module and send out an audible or optical alarm signal.
11. The fabric defect real-time detection system for a circular knitting machine according to claim 10, It is characterized in that The alarm module is integrated on the alarm module at the top end of the circular knitting machine support shaft.
12. A detection method based on the real-time fabric defect detection system according to any one of claims 6 to 11, It is characterized in that The steps include: S1, the camera in the image acquisition device rotates synchronously with the circular knitting machine when the circular knitting machine is running and acquires real-time images of fabrics being produced inside the circular knitting machine; S2, the image captured by the camera is transmitted to the image processing module via the control transmission module, the image processing module processes the image data based on a preset defect recognition algorithm, and issues an alarm message when a defect is detected; The camera collects real-time images of fabrics produced inside the circular knitting machine by interval shooting, and the interval time t≤L / s; wherein L is the length of the fabric produced by the circular knitting machine per unit time, and s is the field of view height of the camera in the vertical direction.
13. The detection method according to claim 12, It is characterized in that In step S2, the preset defect recognition algorithm is an intelligent defect recognition algorithm based on deep learning.
Citation Information
Patent Citations
Online flaw detection device and online flaw detection method of circular knitting machine
CN105044121A
Missed stitch detecting system of circular knitting machine and detecting method
CN106637655A
Device and method for acquiring real-time cloth cover image of circular weft knitting machine
CN106937046A
Be used for knitting quick -witted defective cloth monitoring devices of circular weft
CN206570496U
Detection device and system
CN209486004U
