A machine vision-based loom fault lamp state acquisition system and method
By using a machine vision-based fault light status acquisition system for circular looms, combined with a transparent protective cover and an adjustable bracket, along with image preprocessing and color recognition algorithms, the system solves the problems of accuracy and cost in identifying fault light status on circular looms, achieving low-cost and highly robust fault data acquisition.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- DONGHUA UNIV
- Filing Date
- 2026-03-09
- Publication Date
- 2026-06-12
AI Technical Summary
Existing technologies are insufficient to accurately identify the status of fault lights on circular looms in harsh industrial environments. Traditional methods are costly, pose significant safety risks, and have poor adaptability.
A machine vision-based fault light status acquisition system for circular looms is adopted, including a transparent protective cover, an adjustable bracket, and an edge computing unit. Combined with image preprocessing and color recognition algorithms, it achieves non-invasive and low-cost fault light status acquisition.
It achieves highly robust and low-cost fault light status acquisition, is applicable to various models and installation angles of circular looms, provides high-quality fault data recording, and supports production management and predictive maintenance.
Smart Images
Figure CN122199890A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of industrial Internet of Things (IoT) technology, and in particular to a machine vision-based system and method for acquiring the status of fault lights on a circular loom. Background Technology
[0002] Large circular looms are key equipment in the textile industry. Currently, many of the circular looms still in service are outdated models with control panels consisting only of simple indicator lights, lacking digital communication interfaces (such as RS485 or Ethernet). To achieve digital status monitoring of these machines (such as collecting fault stop signals), the traditional solution is to replace the entire intelligent control panel, which is costly and time-consuming. Another solution is to connect external sensors or PLCs to the equipment circuitry to collect the on / off current of the indicator lights; however, this method requires disconnecting and rewiring, posing safety hazards, and has poor compatibility with various equipment models, making installation and maintenance complex.
[0003] In recent years, machine vision technology has been applied to industrial inspection. Industrial environments are harsh (with abundant dust, oil, and pollutants), and lighting conditions are complex (variable brightness and glare). The installation angles of panels on different machines vary greatly. When setting up cameras to capture images of the panels to identify the status of lights, directly captured images suffer from severe distortion, uneven lighting, and occlusion. General-purpose vision algorithms have extremely poor robustness in such scenarios and cannot guarantee recognition accuracy. Therefore, there is an urgent need for a reliable solution specifically designed for such scenarios that can overcome these difficulties. Summary of the Invention
[0004] The purpose of this application is to provide a machine vision-based system and method for acquiring the status of fault lights on a circular weaving machine, which can improve the accuracy of acquiring the status of fault lights on a circular weaving machine.
[0005] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a machine vision-based circular knitting machine fault light status acquisition system, including: a protective cover, an image acquisition module, a bracket, and an edge computing unit; Both the protective cover and the image acquisition module are located at the adjustment end of the bracket; and the image acquisition module is located inside the protective cover. The adjustment end of the bracket is located above the target circular knitting machine panel; The image acquisition module is connected to the edge computing unit; The image acquisition module is used to acquire images of the target circular weaving machine panel at a preset frequency to obtain a sequence of images of the target circular weaving machine panel. The edge computing unit is used to determine the status of different fault lights on the target circular weaving machine panel based on the target circular weaving machine panel image sequence and using machine vision technology, and generate a status record. The edge computing unit is connected to a remote server; the edge computing unit is used to transmit status records to the remote server.
[0006] Optionally, the protective cover is made of a transparent material; The protective cover has an opening.
[0007] Optionally, the image acquisition module uses an industrial camera.
[0008] Optionally, the adjustment end of the bracket supports multi-degree-of-freedom angle adjustment and locking.
[0009] Optionally, the bracket adopts a multi-section hinged arm or a universal joint structure with a snap-fit.
[0010] Optionally, the edge computing unit employs an embedded computing device.
[0011] Optionally, the machine vision-based circular loom fault light status acquisition system further includes: a supplementary lighting module; The supplementary lighting module is disposed inside the protective cover; the supplementary lighting module is connected to the edge computing unit; The edge computing unit is used to control the switching on and off and the brightness of the edge computing unit.
[0012] Secondly, this application provides a machine vision-based method for acquiring the status of fault lights on a circular weaving machine, including: A machine vision-based circular weaving machine fault light status acquisition system is installed on the target circular weaving machine panel. The angle of the adjustment end of the bracket is adjusted to obtain the image of the target circular weaving machine panel as a calibration image. Based on the calibration image, determine the mapping relationship between fault light area and fault type; The target circular weaving machine panel image is acquired at a preset frequency to obtain a sequence of target circular weaving machine panel images. Based on the target circular loom panel image sequence, machine vision technology and the fault light area-fault type mapping relationship are used to determine the status of different fault lights on the target circular loom panel and generate a status record. The status record is transmitted to a remote server.
[0013] Optionally, the step of determining the status of different fault lights on the target circular loom panel based on the target circular loom panel image sequence, using machine vision technology and the fault light area-fault type mapping relationship, and generating a status record, specifically includes: The target circular loom panel image at the current moment is preprocessed to obtain a front view of the target circular loom panel; the preprocessing includes distortion correction and perspective transformation. Convert the front view of the standard circular knitting machine panel from the RGB color space to the HSV color space; Designate any fault light area as the calibration area; Based on the fault light area-fault type mapping relationship, the fault light corresponding to the calibration area is determined as the target light; different fault light areas correspond to different fault lights; different fault lights correspond to different fault types. Query the color characteristics of the target light; Determine the percentage of pixels in the hue and saturation channels within the calibration area that match the color characteristics of the target light. Update the target light's status record using the pixel count ratio; Update the calibration area and return to the step "Determine the fault light corresponding to the calibration area as the target light based on the fault light area-fault type mapping relationship" until all fault light areas are traversed, and the status record update of the fault light corresponding to all fault light areas is completed.
[0014] Optionally, the status record of the target light is updated using the pixel count ratio, specifically including: The last element in the real-time state sequence of the target light is the previous state of the target light; Determine whether the pixel count ratio reaches the pixel count ratio threshold to obtain a first determination result; If the first judgment result is negative, then the real-time status of the target light is determined to be off; Add the real-time status code of the target light as the last element to the real-time status sequence of the target light. Determine whether the target light was previously lit to obtain the second determination result; If the second judgment result is negative, return to the step "update the calibration area"; If the second judgment result is yes, a status record is generated and the process returns to the step "update the calibration area"; the status record includes: the timestamp of the target circular loom panel image at the current moment, the target circular loom panel ID, the fault type corresponding to the fault light, and the duration of the fault; If the first judgment result is yes, then the real-time state of the target light is determined to be on; Add the real-time status code of the target light as the last element to the real-time status sequence of the target light. Determine whether the target light was previously off to obtain the third judgment result; If the third judgment result is negative, return to the step "update the calibration area"; If the third judgment result is yes, then the timestamp of the target circular loom panel image at the current moment is determined as the fault start time, and the process returns to the step "update the calibration area".
[0015] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a machine vision-based system and method for acquiring fault light status of a circular knitting machine, including a protective cover fixed to the machine base, a camera installed inside the cover and fixed by an adjustable bracket, and an edge computing unit (such as a Raspberry Pi) connected to the camera. The adjustable-angle transparent protective cover creates a stable and directly facing visual environment for image acquisition. Combined with a software algorithm that allows for one-time calibration and long-term recognition, this system achieves non-invasive, low-cost, and highly robust status data acquisition for various older circular knitting machines.
[0016] The system captures panel images periodically using a camera; preprocesses the images using techniques such as perspective transformation; identifies the colors within pre-marked fault light areas to determine the on / off status of the lights; if a light is detected on, it records the fault type, timestamp, and duration, and uploads this information to the server. This method offers the following advantages: Non-invasive and low-cost: It does not affect the original structure and electrical safety of the equipment, and the cost is far lower than replacing the smart panel.
[0017] High versatility and easy deployment: Through physical bracket adjustment and software calibration, it can be quickly adapted to different models and installation angles of circular knitting machines, making it highly scalable.
[0018] High robustness: The transparent protective cover solves the problem of environmental interference; the adjustable bracket solves the problem of angle distortion; and the built-in supplementary light solves the problem of lighting, making the core recognition algorithm simple and reliable.
[0019] High data value: It not only collects status but also accurately records the time and duration of failures, providing a high-quality data foundation for production management, equipment efficiency analysis (OEE), and predictive maintenance. Attached Figure Description
[0020] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0021] Figure 1 This is a schematic diagram of a machine vision-based circular weaving machine fault light status acquisition system according to an embodiment of this application. Figure 2 This is a schematic diagram of a protective cover in one embodiment of this application; Figure 3 This is a schematic diagram illustrating the working principle of the edge computing unit in one embodiment of this application. Detailed Implementation
[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0023] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.
[0024] In one exemplary embodiment, such as Figure 1 As shown, a machine vision-based fault light status acquisition system for a circular loom is provided, including: a protective cover, an image acquisition module, a support, and an edge computing unit.
[0025] like Figure 2The protective cover is made of transparent material and has openings. The transparent protective cover, made of transparent materials such as acrylic, is fixed to the top of the circular knitting machine control panel via clips, magnets, or bolts, forming a sealed or semi-sealed space to effectively isolate external lint, dust, and oil. Both the protective cover and the image acquisition module are located at the adjustable end of the bracket; the image acquisition module is housed inside the protective cover. The adjustable end of the bracket supports multi-degree-of-freedom angle adjustment and locking. The bracket uses a multi-section hinged arm or a universal joint structure with clips, or has multiple clip positioning points. The bracket, located inside the transparent protective cover, is used to fix and adjust the shooting angle of the image acquisition module, ensuring its optical axis is substantially perpendicular to the surface of the control panel. The adjustable bracket connects the protective cover and the camera, allowing technicians to quickly and precisely adjust the camera's pitch and rotation angles on-site, ensuring its shooting plane is parallel to the panel plane, fundamentally eliminating image perspective distortion caused by different installation angles. The adjustable end of the bracket is located above the target circular knitting machine panel. The image acquisition module is connected to the edge computing unit. The image acquisition module uses an industrial camera. It acquires images of the target circular knitting machine panel at a preset frequency, resulting in an image sequence. The edge computing unit, based on this image sequence and using machine vision technology, determines the status of different fault lights on the target circular knitting machine panel and generates a status record. The edge computing unit connects to a remote server and transmits the status records to the remote server. The edge computing unit, using a Raspberry Pi, Jetson Nano, or other embedded computing device, has the camera built-in or connected. It controls image capture, runs visual recognition algorithms, records data, and communicates with the upper layer. The edge computing unit controls image acquisition, processes image data, identifies light statuses, and generates status data; a data communication module, located on the edge computing unit, uploads the status data to the remote server.
[0026] The machine vision-based circular loom fault light status acquisition system also includes a supplementary lighting module. This module is located inside the protective cover and is connected to the edge computing unit. The edge computing unit controls its on / off state and brightness, providing a stable and uniform lighting environment for image acquisition and overcoming the problem of varying ambient light.
[0027] like Figure 3 The edge computing unit is configured to perform the following methods: Initialization calibration steps: Receive user input, select each fault light area in the image and establish a mapping relationship with the fault type.
[0028] Image acquisition steps: The image acquisition module is triggered periodically to acquire panel images.
[0029] Image preprocessing steps: Perform perspective transformation correction on the acquired image and convert it to the HSV color space.
[0030] Status recognition steps: Within the calibrated fault light areas, calculate the average or percentage of pixel values for a specific color channel. If the value exceeds a preset threshold, the light is determined to be on.
[0031] Data recording steps: Monitor the changes in the status of the lights, record the start time of the lights turning on, calculate the duration, and generate status records by associating them with the fault type.
[0032] In another embodiment, a machine vision-based method for acquiring the status of fault lights on a circular loom is provided, including: Step 1: Install a machine vision-based circular weaving machine fault light status acquisition system on the target circular weaving machine panel, adjust the angle of the adjustment end of the bracket, and acquire the image of the target circular weaving machine panel as a calibration image.
[0033] Secure the protective cover to the panel and adjust the camera to the optimal angle using the bracket. In the system software interface, perform a one-time calibration of the current panel—the operator outlines the area of each indicator light in the real-time view and assigns it a fault type. For different models of circular loom panels, manually or automatically outline the area of each fault light in the acquired image and input the corresponding fault type information for each area to establish an area-type mapping table.
[0034] Step 2: Based on the calibration image, determine the mapping relationship between fault light area and fault type.
[0035] Step 3: Acquire images of the target circular loom panel at a preset frequency to obtain a sequence of target circular loom panel images. Control the camera to continuously acquire panel images at fixed time intervals (e.g., 3 seconds).
[0036] Step 4: Based on the target circular loom panel image sequence, use machine vision technology and the fault light area-fault type mapping relationship to determine the status of different fault lights on the target circular loom panel and generate a status record.
[0037] The acquired images are subjected to distortion correction and perspective transformation to obtain a front view of the panel. Each calibrated region in the image is extracted, and the color characteristics of the pixels within the region are analyzed. The on / off state of each faulty light is determined based on a preset threshold. The state sequence of each light is tracked; when a change from "off" to "on" is detected, the time is recorded as the fault start time. When a change from "on" to "off" is detected, the fault duration is calculated. A state record containing a timestamp, device ID, fault type, duration, and start time is generated. The specific method for analyzing color features is as follows: the image is converted from the RGB color space to the HSV color space. For each calibrated region, the percentage of pixels in the hue (H) and saturation (S) channels that match the target light's color characteristics is calculated. If the percentage exceeds a threshold, the light is determined to be on. A time-window-based state filtering algorithm is used, requiring multiple consecutive detections of a light being on before confirming a state change to prevent false positives.
[0038] Step 5: Transmit the status record to a remote server (database or cloud platform) via 4G, Wi-Fi or Ethernet.
[0039] Step 4 specifically includes: The target circular loom panel image at the current moment is preprocessed to obtain a front view of the target circular loom panel. The preprocessing includes distortion correction and perspective transformation.
[0040] The front view of the circular knitting machine panel is converted from the RGB color space to the HSV color space. The acquired image is preprocessed, including applying perspective transformation to correct the image to a front view and converting it to the HSV color space for color recognition.
[0041] Define any fault light area as the calibration area.
[0042] Based on the fault light area-fault type mapping relationship, the fault light corresponding to the calibrated area is determined as the target light. Different fault light areas correspond to different fault lights. Different fault lights correspond to different fault types.
[0043] Query the color characteristics of the target light.
[0044] Determine the percentage of pixels in the hue and saturation channels within the calibration area that match the color characteristics of the target light.
[0045] Update the target light's status record using the percentage of pixels.
[0046] Update the calibration area and return to the step "Determine the fault light corresponding to the calibration area as the target light based on the fault light area-fault type mapping relationship" until all fault light areas are traversed, and the status record update of the fault light corresponding to all fault light areas is completed.
[0047] The process of updating the target light's status record using pixel count ratio specifically includes: The last element in the real-time state sequence of the target light is the previous state of the target light.
[0048] Determine whether the percentage of pixels reaches the pixel percentage threshold to obtain the first judgment result.
[0049] If the first judgment result is negative, then the real-time status of the target light is determined to be off.
[0050] The real-time status code of the target light is added as the last element to the real-time status sequence of the target light.
[0051] Determine whether the target light was on in the previous state to obtain the second judgment result.
[0052] If the second judgment result is negative, return to the step "Update calibration area".
[0053] If the second judgment result is yes, a status record is generated, and the process returns to the "Update Calibration Area" step. The status record includes: the timestamp of the target circular loom panel image at the current moment, the target circular loom panel ID, the fault type corresponding to the fault light, and the duration of the fault.
[0054] If the first judgment result is yes, then the real-time status of the target light is determined to be on.
[0055] The real-time status code of the target light is added as the last element to the real-time status sequence of the target light.
[0056] Determine whether the target light was in an off state in the previous step to obtain the third judgment result.
[0057] If the third judgment result is negative, return to the step "Update calibration area".
[0058] If the third judgment result is yes, then the timestamp of the target circular loom panel image at the current moment is determined as the fault start time, and the process returns to the step "Update calibration area".
[0059] The hardware deployment process for the machine vision-based method for acquiring the status of fault lights on a circular weaving machine is as follows: Select a suitable transparent protective cover according to the size of the circular knitting machine panel, and fix it to the machine base with bolts.
[0060] Mount the camera on the adjustable bracket inside the cover, and connect the power and network cables.
[0061] Turn on the camera video stream and observe through the software interface. Manually adjust each joint of the bracket until the panel image is displayed correctly and completely on the screen.
[0062] Tighten all adjustment knobs or clips on the bracket to fix the camera angle.
[0063] Power on and connect the edge computing unit (Raspberry Pi) to the network.
[0064] Software configuration and operation: The camera takes pictures of the control panel of the large circular machine at set time intervals (e.g., once every 3 seconds) and transmits the pictures to a local computer or server.
[0065] The software first preprocesses the image, using image enhancement algorithms to improve parameters such as brightness and contrast, and employing methods such as Gaussian filtering to remove noise from the image.
[0066] Using a pre-trained color recognition model, traffic light colors are detected in the pre-processed image in the HSV color space. For each color of traffic light, a corresponding color threshold range is set. When the color value of a certain area in the image falls within this range, that area is determined to be an on traffic light area.
[0067] Once a traffic light is detected to be on, the software records the time the current image was generated and continuously monitors the status of the traffic light area. When the traffic light goes out, it calculates the duration the traffic light was on.
[0068] Based on a pre-established database linking fault types to illuminated traffic light areas, the software automatically matches the fault type corresponding to the illuminated area and records information such as fault type, illumination start time, and illumination duration in an Excel spreadsheet for subsequent analysis and querying.
[0069] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0070] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.
Claims
1. A machine vision-based fault light status acquisition system for a circular loom, characterized in that, include: Protective shield, image acquisition module, support frame, and edge computing unit; Both the protective cover and the image acquisition module are located at the adjustment end of the bracket; Furthermore, the image acquisition module is housed within the protective cover; The adjustment end of the bracket is located above the target circular knitting machine panel; The image acquisition module is connected to the edge computing unit; The image acquisition module is used to acquire images of the target circular weaving machine panel at a preset frequency to obtain a sequence of images of the target circular weaving machine panel. The edge computing unit is used to determine the status of different fault lights on the target circular weaving machine panel based on the target circular weaving machine panel image sequence and using machine vision technology, and generate a status record. The edge computing unit is connected to a remote server; the edge computing unit is used to transmit status records to the remote server.
2. The machine vision-based circular loom fault light status acquisition system according to claim 1, characterized in that, The protective shield is made of a transparent material; The protective cover has an opening.
3. The machine vision-based circular loom fault light status acquisition system according to claim 1, characterized in that, The image acquisition module uses an industrial camera.
4. The machine vision-based circular loom fault light status acquisition system according to claim 1, characterized in that, The adjustable end of the bracket supports multi-degree-of-freedom angle adjustment and locking.
5. The machine vision-based circular loom fault light status acquisition system according to claim 4, characterized in that, The bracket adopts a multi-section hinged arm or a universal joint structure with a snap-fit.
6. The machine vision-based circular loom fault light status acquisition system according to claim 1, characterized in that, The edge computing unit uses an embedded computing device.
7. The machine vision-based circular loom fault light status acquisition system according to claim 1, characterized in that, The machine vision-based circular loom fault light status acquisition system also includes: a supplementary lighting module; The supplementary lighting module is disposed inside the protective cover; the supplementary lighting module is connected to the edge computing unit; The edge computing unit is used to control the switching on and off and the brightness of the edge computing unit.
8. A method for acquiring the status of fault lights on a circular loom based on machine vision, characterized in that, include: Install the machine vision-based circular weaving machine fault light status acquisition system as described in any one of claims 1-7 on the target circular weaving machine panel, adjust the angle of the adjustment end of the bracket, and acquire the image of the target circular weaving machine panel as a calibration image; Based on the calibration image, determine the mapping relationship between fault light area and fault type; The target circular weaving machine panel image is acquired at a preset frequency to obtain a sequence of target circular weaving machine panel images. Based on the target circular loom panel image sequence, machine vision technology and the fault light area-fault type mapping relationship are used to determine the status of different fault lights on the target circular loom panel and generate a status record. The status record is transmitted to a remote server.
9. The method for acquiring the status of fault lights on a circular loom based on machine vision according to claim 8, characterized in that, The step of determining the status of different fault lights on the target circular weaving machine panel based on the target circular weaving machine panel image sequence, using machine vision technology and the fault light area-fault type mapping relationship, and generating a status record, specifically includes: The target circular loom panel image at the current moment is preprocessed to obtain a front view of the target circular loom panel; the preprocessing includes distortion correction and perspective transformation. Convert the front view of the standard circular knitting machine panel from the RGB color space to the HSV color space; Designate any fault light area as the calibration area; Based on the fault light area-fault type mapping relationship, the fault light corresponding to the calibration area is determined as the target light; different fault light areas correspond to different fault lights; different fault lights correspond to different fault types. Query the color characteristics of the target light; Determine the percentage of pixels in the hue and saturation channels within the calibration area that match the color characteristics of the target light. Update the target light's status record using the pixel count ratio; Update the calibration area and return to the step "Determine the fault light corresponding to the calibration area as the target light according to the fault light area-fault type mapping relationship" until all fault light areas are traversed, and the status record update of the fault light corresponding to all fault light areas is completed.
10. The method for acquiring the status of fault lights on a circular loom based on machine vision according to claim 9, characterized in that, Update the target light's status record using pixel count ratio, specifically including: The last element in the real-time state sequence of the target light is the previous state of the target light; Determine whether the pixel count ratio reaches the pixel count ratio threshold to obtain a first determination result; If the first judgment result is negative, then the real-time status of the target light is determined to be off; Add the real-time status code of the target light as the last element to the real-time status sequence of the target light. Determine whether the target light was previously lit to obtain the second determination result; If the second judgment result is negative, return to the step "update the calibration area"; If the second judgment result is yes, a status record is generated and the process returns to the step "update the calibration area"; the status record includes: the timestamp of the target circular loom panel image at the current moment, the target circular loom panel ID, the fault type corresponding to the fault light, and the fault duration; If the first judgment result is yes, then the real-time state of the target light is determined to be on; Add the real-time status code of the target light as the last element to the real-time status sequence of the target light. Determine whether the target light was previously off to obtain the third judgment result; If the third judgment result is negative, then return to the step "update the calibration area"; If the third judgment result is yes, then the timestamp of the target circular loom panel image at the current moment is determined as the fault start time, and the process returns to the step "update the calibration area".