Woven label production control method and system

Abstract

The invention provides a woven label production control method and system, and relates to the technical field of woven label production control, and the key points of the technical scheme are as follows: obtaining an expected shutdown position of a label weaving machine; in the process of executing the shutdown braking operation, image information of the motion state of a key weaving part is collected in real time, and motion state parameters of a main shaft of the label weaving machine are reflected; calculating a final shutdown position of the label weaving machine based on the image information and the main shaft motion state parameters; and comparing the calculated final stop position with an expected stop position, and if the final stop position deviates from the expected stop position, adjusting a braking effect applied to the label weaving machine before the label weaving machine finishes braking so as to enable the actual final stop position of the label weaving machine to approach to the expected stop position. The woven label production control method and system provided by the invention have the advantage of reducing the deviation between the actual shutdown position and the expected shutdown position.

Description

Technical Field

[0001] The present application relates to the technical field of woven label production control, and in particular, to a woven label production control method and system. Background Art

[0002] In modern woven label production environments, multiple high-speed woven label machines are typically deployed to significantly improve production efficiency and product quality. These advanced woven label machines are often equipped with a real-time yarn breakage defect recognition and control system based on multimodal data fusion. By integrating various sensors, such as industrial cameras and microphone arrays, this system can capture and analyze image and sound information from the weaving process in real time. This allows for rapid and accurate identification of abnormal events, such as warp or weft yarn breaks, and triggers the woven label machine's automatic shutdown mechanism.

[0003] In the woven label production process, stop control after a yarn break is crucial. An ideal control goal is to ensure that, upon detecting a yarn break, the woven label machine stops precisely at a predetermined position that is optimal for the operator to handle the break. This predetermined position is not fixed but closely depends on the specific type of yarn break. For example, in the event of a warp break, the desired stop position should ensure that the broken warp yarn end is clearly visible, allowing the operator to quickly and accurately connect the yarn. This prevents the broken warp yarn from becoming entangled in the fabric or with other yarns during the subsequent weaving process. This not only significantly simplifies repair efforts but also effectively reduces waste. Conversely, in the event of a weft break, the ideal stop position might be at a specific mechanical phase, either after the weft beater completes one reed beat-up operation or before the next weft insertion begins. This stop position facilitates the operator's removal of the broken weft yarn at the weft fell and creates favorable conditions for re-insertion. If the weaving label machine fails to stop at the appropriate position, for example, it stops before the beating-up action is completed, it will not only cause inconvenience in yarn breakage processing, but may also cause damage to mechanical parts or leave obvious defects on the fabric.

[0004] Existing multimodal data fusion systems have made significant progress in the identification of yarn break events (e.g., determining whether a yarn break exists) and preliminary classification (e.g., distinguishing between warp breakage, weft breakage, single yarn breakage, or multiple yarn breakage). These systems are able to generate preliminary stop control instructions containing the desired stop position or stop window based on the identified yarn break type. Typically, this desired stop position is defined relative to the fabric position or the operating angle of the main shaft of the weaving machine when the sensor captures the yarn breakage event. For example, the system may calculate that after the warp breakage point is detected, the main shaft of the weaving machine needs to rotate a specific angle unit before starting to brake; or for weft breakage, the system may require the weaving machine to complete the entire braking process within a preset phase window within the current weft beating cycle.

[0005] However, a woven label machine is a highly complex mechatronic system, comprising a high-speed motor, a precision transmission mechanism, and numerous key components that reciprocate or rotate at high speed. When the control system issues a stop command, the machine undergoes a deceleration and braking process from high-speed operation to complete rest. This process is influenced by a variety of factors, including but not limited to the inherent delay in motor response, the inertia of mechanical components, the micro-clearances and elastic deformation in the transmission chain, and the dynamic changes in loads (such as yarn tension and fabric structure) during the weaving process. These factors combine to influence the deceleration characteristics of the woven label machine during braking, as well as its actual stopping position, which are difficult to accurately calculate and control in real time, including the current operating speed, the material properties of the yarn used, the fabric structure, and even the ambient temperature.

[0006] Therefore, although existing multimodal systems can accurately identify yarn break types and calculate theoretically expected stop command parameters, due to the complex dynamic characteristics of the woven label machine and the uncertainty in the control execution process, there may still be an incalculable and significant deviation between the actual stop position of the woven label machine after executing the stop action and the expected stop position calculated by the system based on the characteristic points of the yarn break event. This deviation directly affects the efficiency of yarn break processing and the final quality of the fabric.

[0007] For example, if the system identifies a warp yarn break and calculates the ideal stopping point A (which ensures that the broken yarn end is fully exposed), but due to the excessive inertia of the weaving machine or the lag in the braking torque response, the weaving machine actually stops at a position after point A, which may cause the end of the broken warp yarn to be covered by the subsequently woven weft yarn. The operator has to spend extra time and energy to pick up the weft yarn to find the broken end, which not only prolongs the fault recovery time but also increases the length of waste products. On the contrary, if the system determines that the weft yarn is broken and instructs the weaving machine to stop at point B after the beating-up action is completed to facilitate the removal of the broken weft, but if the weaving machine stops before point B due to premature braking, the beating-up action may not be completely completed and the reed is still pressed on the weaving mouth, which is also not conducive to the cleaning operation.

[0008] This deviation between the actual and desired stop positions, caused by the dynamic characteristics of the loom and uncertainties in the control execution process, highlights the shortcomings of existing systems at the control execution level. Current multimodal synchronous acquisition and fusion technologies primarily focus on improving the accuracy and real-time performance of yarn break event recognition. However, after issuing a stop command, the system generally lacks an effective mechanism to sense or verify in real time the final actual stop position of the loom's main shaft or key actuator components (such as the reed and heald frame) relative to the point of the yarn break. Furthermore, the system lacks the ability to dynamically adjust control based on the deviation between this actual stop position and the desired position, or to provide precise positioning assistance to the operator. In short, while existing systems can identify the time of yarn break and its relative position on the fabric and issue a stop command, they are unable to accurately sense the loom's state in real time during the braking process, calculate the final stop position, and dynamically adjust based on the calculated deviation. This makes it difficult for carefully designed stop control strategies for different yarn break types to achieve the desired results, seriously hindering the further improvement of automation levels and maximizing production efficiency.

[0009] In view of the above problems, the existing technology is in urgent need of improvement. Summary of the Invention

[0010] The purpose of the present application is to provide a woven label production control method and system, which has the advantage of reducing the deviation between the actual stop position and the expected stop position.

[0011] In the first aspect, the present application provides a woven label production control method, the technical solution is as follows: Obtaining an expected stopping position of the woven label machine preset for responding to a woven label yarn breakage event and corresponding to the type of the woven label yarn breakage event; In the process of the woven label machine performing a stop braking operation in response to the woven label yarn breakage event, real-time collection of image information at least including image information reflecting the yarn breakage state in the weaving area of ​​the woven label machine or the motion state of key weaving components, and parameters reflecting the motion state of the main shaft of the woven label machine; Based on the real-time collected image information and the spindle motion state parameters, and in combination with the braking action currently applied to the woven label machine, calculating the final stop position of the woven label machine under the braking action; The calculated final stop position is compared with the expected stop position. If the comparison result indicates that the calculated final stop position deviates from the expected stop position, the braking action applied to the woven label machine is adjusted before the braking of the woven label machine is completed so that the actual final stop position of the woven label machine approaches the expected stop position.

[0012] Furthermore, in the present application, the step of calculating the final stop position of the woven label machine under the braking action based on the real-time collected image information and the spindle motion state parameters, and in combination with the braking action currently applied to the woven label machine, includes: During the process of the woven label machine performing a stop braking operation, determining the actual running deceleration of the woven label machine under the braking action currently applied to the woven label machine according to the change of the main shaft motion state parameter collected in real time over time; Based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameter at the current moment, and the yarn breakage state or the current physical position of the key weaving component obtained by the real-time image information collected at the current moment, the remaining motion amount of the woven label machine from the current moment to the complete stop is estimated; The final stopping position of the woven label machine under the braking action is calculated by combining the current physical position with the estimated remaining motion amount.

[0013] Furthermore, in the present application, during the process of the woven label machine performing the stop braking operation, the step of determining the actual operating deceleration of the woven label machine under the braking action currently applied to the woven label machine according to the change of the spindle motion state parameter collected in real time over time includes: During the process of the woven label machine performing a stop braking operation, monitoring the change in the magnitude or mode of the braking action currently applied to the woven label machine; Based on the monitored change in the magnitude or mode of the braking action, dividing the parking braking operation process in time into at least one braking sub-interval corresponding to the change; For each of the at least one braking sub-period, based on the change of the spindle motion state parameters collected in real time during the braking sub-period over time, the actual operating deceleration of the woven label machine during the braking sub-period and under the braking action currently applied to the woven label machine corresponding to the braking sub-period is determined.

[0014] Furthermore, in the present application, the step of calculating the remaining motion amount of the woven label machine from the current moment to the complete stop based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameter at the current moment, and the yarn breakage state or the current physical position of the key weaving component obtained by the real-time image information collected at the current moment includes: determining, for the yarn breakage state or the current physical position of the key weaving component acquired through the real-time collected image information at the current moment, a total image information acquisition delay time associated with an acquisition process of the current physical position; Calculating a physical position compensation value for correcting the current physical position based on the total delay time for acquiring the image information and the kinematic parameters of the woven label machine during the total delay time for acquiring the image information; Applying the physical position compensation value to the current physical position to obtain a delay-compensated physical position; Acquiring a confidence parameter obtained by performing target recognition on the yarn breakage state or the key weaving component through the real-time collected image information; determining a final physical position for estimating the remaining motion amount based on whether the delay-compensated physical position is lower than a preset validity threshold and, when the confidence parameter is lower than the preset validity threshold, according to a preset confidence processing rule; Based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameter at the current moment, and the final physical position, the remaining motion amount of the woven label machine from the current moment to complete stop is calculated.

[0015] Furthermore, in the present application, the step of determining the final physical position for estimating the remaining motion amount based on whether the confidence parameter is lower than a preset validity threshold and, when the confidence parameter is lower than the preset validity threshold, according to a preset confidence processing rule, includes: When the confidence parameter is lower than the preset validity threshold: Obtaining a current value of the confidence parameter and a changing trend of the confidence parameter within a preset time period; selecting, based on the obtained current value of the confidence parameter and the change trend, at least one alternative physical location determination strategy from a plurality of pre-configured alternative physical location determination strategies in accordance with a pre-configured selection condition corresponding to the current value of the confidence parameter and / or the change trend, wherein each of the alternative physical location determination strategies is associated with at least one of the selection conditions, and the setting of the selection conditions takes into account expected processing effects of different alternative strategies under corresponding confidence parameter values ​​or change trends; Applying the selected at least one candidate physical position determination strategy, combined with the delay-compensated physical position, to determine the final physical position for estimating the remaining motion; If the confidence parameter is not lower than the preset validity threshold, the delay-compensated physical position is determined as the final physical position for calculating the remaining motion.

[0016] Furthermore, in the present application, in the process of the woven label machine performing a stop braking operation in response to the woven label yarn breakage event, the step of collecting in real time at least image information reflecting the yarn breakage state or the motion state of key weaving components in the weaving area of ​​the woven label machine, and reflecting the motion state parameters of the main shaft of the woven label machine includes: Predetermining a time delay reference value between the image information acquisition system of the woven label machine and the spindle motion state parameter acquisition system; In the process of the woven label machine performing a stop braking operation in response to the woven label yarn breakage event, real-time acquisition of image information at least including image information reflecting the yarn breakage state in the weaving area of ​​the woven label machine or the movement state of key weaving components, and adding first time information to each data unit of the acquired image information; In the process of the woven label machine performing a stop braking operation in response to the woven label yarn breakage event, parameters reflecting the motion state of the main shaft of the woven label machine are collected in real time, and second time information is added to the data unit of each acquired main shaft motion state parameter; Based on the time delay reference value and in combination with the first time moment information and the second time moment information, the data unit of the image information or the data unit of the spindle motion state parameter is time-adjusted to obtain image information reflecting the interrupted yarn state or the motion state of the key weaving component in the weaving area of ​​the woven label machine under the same time reference, as well as parameters reflecting the motion state of the spindle of the woven label machine.

[0017] Furthermore, in the present application, the step of adjusting the braking action applied to the woven label machine before the woven label machine is braked includes: Before the woven label machine is braked, when the deviation between the final stop position and the desired stop position satisfies a preset adjustment start condition, starting the adjustment of the braking action applied to the woven label machine; After the adjustment is initiated, determining adjustment parameters for adjusting the braking effect applied by the woven label machine and / or a timing for executing the next adjustment based on the deviation amount, a variation trend of the deviation amount determined by monitoring a variation of the deviation amount within a preset monitoring time window, and a remaining braking time of the woven label machine from a current moment to completion of braking; The braking action applied to the woven label machine is adjusted according to the determined adjustment parameters and / or the execution timing. Furthermore, when performing the adjustment, the execution frequency of the adjustment is limited according to a preset frequency control strategy to balance the correction effect of the deviation and control the system load of the woven label machine.

[0018] Furthermore, in the present application, the woven label yarn breakage event type is a yarn breakage defect type obtained by identifying multimodal data during the weaving process of the woven label machine, including at least one of warp yarn breakage, weft yarn breakage, single yarn breakage, and multiple yarn breakage.

[0019] Furthermore, in the present application, the key weaving components include one or a combination of knitting needles, reeds, and jacquard needles of a jacquard device.

[0020] In a second aspect, the present application also proposes a woven label production control system, which includes: An acquisition module is used to acquire an expected stop position of the woven label machine that is preset to deal with a woven label yarn breakage event and corresponds to the type of the woven label yarn breakage event; an acquisition module for acquiring, in real time, image information reflecting at least a yarn break state in a weaving area of ​​the woven label machine or a motion state of a key weaving component, and parameters reflecting a motion state of a main shaft of the woven label machine during a process in which the woven label machine performs a stop braking operation in response to the woven label yarn break event; a calculation module, configured to calculate a final stopping position of the woven label machine under the braking action based on the real-time collected image information and the spindle motion state parameters, and in combination with the braking action currently applied to the woven label machine; an adjustment module, configured to compare the final stop position calculated by the final stop position calculation module with the expected stop position obtained by the expected position acquisition module; if the comparison result indicates that the calculated final stop position deviates from the expected stop position, then, before the braking of the woven label machine is completed, adjust the braking action applied to the woven label machine so that the actual final stop position of the woven label machine approaches the expected stop position.

[0021] From the above, it can be seen that the present application provides a woven label production control method and system, which obtains the expected stop position, collects image information and spindle motion parameters in real time during the braking process, calculates the final stop position, and dynamically adjusts the braking action according to the deviation between the calculated position and the expected position. It has the advantages of significantly reducing the deviation between the actual stop position and the expected stop position by collecting multimodal data in real time during the braking process of the woven label machine and dynamically adjusting the braking action, so that the woven label machine can stop more accurately at a position that is conducive to yarn breakage processing, thereby improving the efficiency of yarn breakage repair, reducing downtime and waste generation, and improving product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 A flow chart of a woven label production control method provided in this application.

[0023] Figure 2 This is a schematic diagram of the structure of a woven label production control system provided in this application.

[0024] In the figure: 210, acquisition module; 220, collection module; 230, calculation module; 240, adjustment module. DETAILED DESCRIPTION

[0025] The technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. The components of the present application generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application for which protection is claimed, but merely represents selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of this application.

[0026] It should be noted that similar reference numerals and letters represent similar items in the following drawings. Therefore, once an item is defined in one drawing, it does not need to be further defined or explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and should not be understood as indicating or implying relative importance.

[0027] In traditional woven label production control systems, when a yarn break occurs and a stop brake is executed, the inherent complex dynamic characteristics of the woven label machine, such as mechanical inertia, backlash and elastic deformation in the drive chain, response delay and nonlinearity of the braking system, and dynamic changes in the weaving process, including load changes caused by speed adjustments, yarn material changes, and fabric structure changes, all work together to make the deceleration characteristics of the woven label machine from high-speed operation to complete standstill difficult to accurately calculate and control. As a result, there will be an incalculable and significant deviation between the actual stop position of the woven label machine and the expected stop position calculated by the system based on the type of yarn break event.

[0028] For example, on a high-speed weaving label machine, the system identifies a warp break. Based on a preset strategy, the machine is expected to stop at a position where the broken warp yarn end is exposed, allowing the user to quickly continue weaving. After the system issues a stop command, the machine begins braking. However, due to machine inertia or a delayed braking response, the machine may not decelerate enough, ultimately stopping at a position beyond the desired position, causing the broken warp yarn end to be covered by the subsequently woven weft yarn. Alternatively, when a weft break is detected, the machine is expected to stop at a specific mechanical phase after the beating-up operation is complete, allowing the user to clear the broken weft. However, if the machine stops prematurely due to a change in braking characteristics, the beating-up operation may not be complete, and the reed may still be pressed against the weaving fell, hindering the clearing operation. In this scenario, while the system can identify the type of yarn break and issue a stop command, it lacks real-time awareness of the machine's state during braking, making it impossible to calculate the final stopping position online or dynamically adjust based on deviations from the calculated position.

[0029] In this regard, refer to Figure 1 , this application proposes a woven label production control method, including: S110, obtaining an expected stopping position of the woven label machine preset for responding to a woven label yarn breakage event and corresponding to the type of the woven label yarn breakage event; S120, in the process of the woven label machine performing a stop braking operation in response to a woven label yarn breakage event, collecting in real time at least image information reflecting a yarn breakage state in a weaving area of ​​the woven label machine or a motion state of a key weaving component, and parameters reflecting a motion state of a main shaft of the woven label machine; S130, based on the real-time collected image information and the main shaft motion state parameters, and in combination with the braking action currently applied to the woven label machine, calculating the final stop position of the woven label machine under the braking action; S140, comparing the calculated final stop position with the expected stop position. If the comparison result indicates that the calculated final stop position deviates from the expected stop position, adjusting the braking action applied to the woven label machine before the braking of the woven label machine is completed so that the actual final stop position of the woven label machine approaches the expected stop position.

[0030] The acquisition of the expected stop position refers to a target stop position pre-set according to the type of yarn break event, which can be obtained by looking up a table or calculating according to the type of yarn break.

[0031] Among them, real-time collection of image information refers to continuously obtaining image data of the weaving area or key components during the braking process, which can be achieved by using visual sensors such as industrial cameras.

[0032] Among them, real-time collection of spindle motion state parameters refers to continuously obtaining the spindle motion data during the braking process, which can be achieved by using encoders, sensors, etc. to obtain parameters such as speed, angle, and acceleration.

[0033] Among them, calculating the final stopping position based on the real-time acquired image information and spindle motion state parameters and combined with the currently applied braking action means using the real-time acquired multimodal data and the currently applied braking action to calculate the position where the machine will eventually stop, which can be specifically achieved by using a data fusion algorithm, etc.

[0034] Comparing the calculated final stop position with the expected stop position refers to comparing the calculated stop position with a preset target position, which can be specifically achieved by using numerical comparison.

[0035] Among them, adjusting the braking effect applied to the woven label machine means changing the braking force applied to the machine before the braking is completed according to the calculated deviation. Specifically, it can be achieved by controlling the motor or brake, such as changing the braking torque or braking mode.

[0036] The core innovation of this application lies in that during the braking process of the woven label machine due to yarn breakage, the final stopping position of the machine is calculated online by real-time fusion of multimodal data such as image information and spindle motion parameters, and based on the deviation between the calculated position and the expected position, the braking action is dynamically adjusted before the braking is completed, thereby achieving precise control of the stopping position of the woven label machine.

[0037] The working process and principle of this application are as follows: first, the desired stop position of the woven label machine is obtained, which is preset according to the type of yarn break event. This desired stop position is predetermined based on different yarn break types, and is intended to provide an ideal stopping point for subsequent yarn break processing.

[0038] Furthermore, during the woven label machine's braking operation in response to a yarn break event, multimodal data is collected in real time. This data includes at least image information reflecting the yarn break status in the woven label machine's weaving area or the motion status of key weaving components, as well as parameters reflecting the woven label machine's main shaft motion status. The image information provides a visual representation of the weaving area and the physical location of key components, while the main shaft motion parameters reflect the overall machine's motion speed, angle, acceleration, and other dynamic information. Real-time data collection ensures continuous and comprehensive information about the machine's current state during deceleration.

[0039] Based on the real-time image information and spindle motion parameters, and combined with the braking action currently applied to the woven label machine, the final stopping position of the woven label machine under this braking action is calculated. This calculation process integrates the fine position reference provided by the visual information, the overall motion state and deceleration characteristics provided by the spindle parameters, and the impact of the current braking action on the deceleration process, thereby dynamically calculating the final stopping position of the machine under the current conditions.

[0040] Specifically, in some implementations, the current physical position of a specific reference point on the woven label machine (e.g., a yarn breakage point or a key weaving component) can be first acquired through real-time image information, denoted as P_1. Simultaneously, the instantaneous velocity of that reference point is determined from the spindle motion state parameters, denoted as v_1. Furthermore, by analyzing the time-dependent variation in spindle speed under the currently applied braking action, the actual operating deceleration of the woven label machine at that moment is calculated, denoted as a_1. The value of a_1 is typically negative.

[0041] After obtaining these three key real-time parameters (P_1, v_1, and a_1), the system then uses basic kinematic principles to calculate the remaining motion of the reference point from its current moment to a complete stop (i.e., a final velocity v_2 equal to 0), denoted as S_1. This remaining motion S_1 can be calculated using the formula: S_1 = (v_2^2 - v_1^2) / (2 * a_1). Since the woven label machine will eventually stop, with a final velocity v_2 of 0, the above formula can be simplified to: S_1 = -v_1^2 / (2 * a_1). After calculating the remaining motion S_1, the system then vectorially adds it to the current physical position of the reference point, P_1, previously determined through image information, to obtain the predicted final stopping position of the woven label machine under braking. This is denoted as P_2, calculated using the formula: P_2 = P_1 + S_1. Through this series of calculation steps that combine real-time image positioning, spindle dynamic parameter monitoring and kinematic calculation, the present application can dynamically and relatively accurately predict the final stop point of the woven label machine under the current braking conditions.

[0042] The calculated final stop position is then compared with the preset desired stop position. If the comparison indicates that the calculated final stop position deviates from the desired stop position, the braking action applied to the woven label machine is adjusted before braking is complete. This braking adjustment aims to alter the machine's deceleration trajectory to correct the calculated final stop position closer to the desired position. This real-time comparison and dynamic adjustment creates a closed-loop control process, overcoming the uncertainties introduced by the machine's dynamic characteristics and changing operating conditions, and guiding the actual final stop position closer to the desired position.

[0043] As a preferred embodiment, the solution of this application is specifically implemented as follows: The control system for a woven label machine pre-stores a data table of desired stop positions corresponding to different types of yarn break events (e.g., warp breakage, weft breakage). When the yarn break detection system identifies a specific type of yarn breakage event, the control system queries this data table and obtains the corresponding desired stop position.

[0044] After the woven label machine initiates a stop brake command, an image acquisition system (e.g., an industrial camera mounted above the weaving area) begins capturing images of the weaving area in real time at a preset frequency. The image processing unit analyzes these images to identify the location of yarn breaks or the current physical position of key weaving components (e.g., reeds and needles). Simultaneously, a spindle encoder or sensor collects real-time motion parameters of the woven label machine's main shaft, such as spindle angle, angular velocity, and angular acceleration.

[0045] The system receives real-time image information (including position information) and spindle motion state parameters, as well as information about the braking effect currently being applied by the braking system (e.g., braking torque, braking current, or braking mode). Based on this real-time data, it applies a preset motion model or algorithm to calculate where the woven label machine will ultimately stop if it continues to decelerate according to the current braking effect. For example, it can calculate the remaining distance to travel based on the current speed, acceleration (determined by spindle parameters), current position (determined by image information), and braking effect, and then calculate the final stopping position based on the current position.

[0046] The calculated final stop position is compared with the previously obtained desired stop position to determine the deviation between the two.

[0047] If the deviation exceeds a preset threshold, the adjustment control unit dynamically adjusts the braking force applied to the woven label machine based on the magnitude and direction of the deviation before the machine comes to a complete stop. For example, if the calculated position exceeds the desired position, the braking torque is increased or the braking strategy is modified to accelerate deceleration; if the calculated position falls short of the desired position, the braking torque is reduced or the braking method is adjusted to slow deceleration. The adjusted braking force is then applied to the woven label machine's brake actuator, thereby altering the machine's deceleration process and bringing the actual final stop position closer to the desired position. This calculation, comparison, and adjustment process continues until the machine stops or the deviation meets the required level.

[0048] Through the above solution, the present application solves the problem of deviation between the actual stop position and the desired stop position when the woven label machine stops due to a yarn break, caused by the machine's inherently complex dynamic characteristics and the dynamic changes in weaving conditions. This allows the woven label machine to more accurately stop at the desired preset position after a yarn break, improving the operator's efficiency in handling yarn breaks, reducing waste and machine damage caused by improper stop positions, and enhancing the automation level and product quality of woven label production.

[0049] Specifically, some of the aforementioned solutions of this application propose calculating the final stopping position of the woven label machine under braking based on real-time multimodal data (image information and spindle motion parameters) and the current braking action applied to the machine. This is then compared with a preset desired stopping position, and the braking action is adjusted based on the deviation to bring the actual final stopping position closer to the desired stopping position. However, during the braking operation, the motion state and deceleration characteristics of the woven label machine are dynamically affected by a variety of complex factors, such as mechanical inertia, backlash and elasticity in the drive chain, the nonlinear response of the braking system, and load fluctuations caused by changes in operating conditions such as operating speed, yarn properties, and fabric structure. These factors make it difficult to accurately calculate the motion process of the woven label machine from the current moment to a complete stop. As a result, the final stopping position calculated simply based on real-time data and the current braking action may deviate significantly from the actual situation. This inaccuracy in the calculation directly affects the effectiveness of subsequent deviation determination and braking adjustments, making it difficult for the woven label machine to accurately stop at the desired position carefully designed for a specific yarn break type, thereby reducing yarn break handling efficiency and increasing scrap rates. Therefore, a more accurate method that can better reflect the actual dynamic characteristics of the woven label machine is needed to calculate the final stop position in order to overcome the uncertainty caused by the above complex factors.

[0050] In this regard, the present application further proposes that based on the real-time collected image information and spindle motion state parameters, and in combination with the braking action currently applied to the woven label machine, the steps of calculating the final stop position of the woven label machine under the braking action include: During the process of the woven label machine performing a stop braking operation, the actual running deceleration of the woven label machine under the braking action currently applied to the woven label machine is determined according to the change of the main shaft motion state parameters collected in real time over time; Based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameters at the current moment, and the yarn breakage state or the current physical position of the key weaving components obtained through real-time image information at the current moment, the remaining motion amount of the woven label machine from the current moment to the complete stop is estimated; The final stopping position of the woven label machine under the braking action is calculated by combining the current physical position with the estimated remaining motion.

[0051] During the process of the woven label machine performing a stop braking operation, the actual operating deceleration of the woven label machine under the braking action currently applied to the woven label machine is determined based on the time-varying changes in the spindle motion state parameters acquired in real time. The spindle motion state parameters may include the angular velocity, angular acceleration, or angular position of the spindle. The actual operating deceleration can be determined by calculating the time-varying rate of change of the spindle angular velocity acquired in real time, or by performing a quadratic differential on the spindle angular position over time.

[0052] Furthermore, based on the determined actual operating deceleration, the instantaneous motion state of the weaving machine determined by the spindle motion state parameters at the current moment, and the yarn breakage state or the current physical position of the key weaving components obtained through real-time image information collected at the current moment, the remaining motion amount of the weaving machine from the current moment to complete stop is calculated.

[0053] The current instantaneous motion state can be provided by spindle motion parameters, such as the current spindle angular velocity and angular position. Real-time image data can be used to identify the current physical location of yarn breaks or key weaving components (e.g., needles, reeds, or jacquard needles) within the weaving area. This physical location can be expressed as a pixel position in the image coordinate system or converted to a spatial position in the loom coordinate system through calibration. To estimate the remaining motion, the spindle's instantaneous motion state (e.g., angular velocity) can be combined with the actual deceleration to calculate the total spindle rotation angle from the current moment to the point of stop using kinematic formulas. Furthermore, the current physical position provided by the image data correlates with key feature points in the weaving process. The estimation of the remaining motion can convert the total spindle rotation angle into the physical displacement of the key feature point or yarn break point on the fabric or machine body. For example, if the correspondence between the spindle rotation angle and the reed position is known, the remaining spindle rotation angle can be converted into the remaining reed travel. By combining the overall motion calculation (based on spindle parameters and deceleration) with the local physical reference point position (based on image information), the estimated remaining motion has physical meaning and is directly linked to the target position where precise stopping is required. For example, if the image recognizes that the yarn break is at a certain position at the current moment, and the current spindle speed and deceleration calculate that the spindle will rotate X degrees, and this X degree rotation corresponds to Y millimeters of fabric forward movement, the calculated remaining motion can be expressed as Y millimeters.

[0054] Thus, by combining the current physical position with the estimated remaining motion, the final stopping position of the woven label machine under the braking action is calculated. The current physical position is obtained based on real-time image information and represents the actual spatial position of the yarn break point or key component at the current moment. The estimated remaining motion is the total displacement from the current moment to the complete stop calculated based on the deceleration determined in real time, the instantaneous motion state and the current physical position. Adding the current physical position to the estimated remaining motion can obtain the physical position of the yarn break point or key component when it finally stops, that is, the final stopping position of the woven label machine. For example, if the current image recognizes that the yarn break point is located at point P on the fabric, and it is estimated that the fabric will move forward a distance D from the current moment to the stop, then the final stopping position is the position after point P moves forward a distance D.

[0055] This calculation method fully utilizes multimodal data collected in real time, dynamically reflecting the actual motion characteristics of the woven label machine during the braking process. The calculation results provide a highly accurate calculation of where the woven label machine will ultimately stop under the current braking action. This high-precision calculation result is then compared with the preset expected stopping position. If there is a deviation, the braking action is adjusted before the woven label machine completes braking. By providing a more accurate calculation of the final stopping position, this method enables subsequent deviation judgment and braking adjustments to be performed more effectively, thereby improving the accuracy of the woven label machine's stopping control. This helps to accurately control the actual final stopping position near the expected position preset for different yarn break types, thereby improving yarn break handling efficiency and reducing scrap rates.

[0056] Specifically, some of the aforementioned solutions of this application propose determining the actual operating deceleration of a woven label machine under the currently applied braking action based on the temporal changes in the real-time acquired spindle motion state parameters, in order to estimate the remaining motion and calculate the final stopping position. However, during the actual braking process of a woven label machine, the applied braking action (e.g., the magnitude of the braking torque or the braking method) may not be constant but may be dynamically adjusted based on control requirements or system status. If a single actual operating deceleration is simply determined based on the spindle motion parameters throughout the entire braking process, or if the impact of changes in the braking action on the deceleration characteristics is not fully considered, the calculated deceleration may not accurately reflect the actual deceleration capability of the woven label machine at different braking stages. Such inaccurate deceleration values ​​directly affect the accuracy of subsequent kinematic calculations of the woven label machine's remaining motion from the current moment to a complete stop, ultimately resulting in a deviation between the calculated final stopping position and the actual situation. This, in turn, affects the effectiveness of braking adjustments based on the calculated results, making it difficult to achieve the goal of accurately stopping the woven label machine at the desired position.

[0057] In this regard, the present application further proposes that during the process of the woven label machine performing the stop braking operation, the actual running deceleration of the woven label machine under the braking action currently applied to the woven label machine is determined according to the changes of the main shaft motion state parameters collected in real time over time, including: during the process of the woven label machine performing the stop braking operation, the changes in the amplitude or mode of the braking action currently applied to the woven label machine are monitored; based on the changes in the amplitude or mode of the monitored braking action, the stop braking operation process is divided into at least one braking sub-period corresponding to the changes in time; for each braking sub-period in at least one braking sub-period, according to the changes of the main shaft motion state parameters collected in real time over time in the braking sub-period, the actual running deceleration of the woven label machine under the braking action currently applied to the woven label machine in the braking sub-period and corresponding to the braking sub-period is determined.

[0058] During the woven label machine's braking operation, changes in the magnitude or mode of the braking action currently applied to the machine are monitored. Specifically, this monitoring may include real-time detection of the braking control signal output by the control system, such as monitoring the torque command value applied to the brake actuator, the pressure signal, or the switching state of different braking modes (e.g., electric braking, mechanical braking). When the monitored braking action magnitude changes significantly, such as exceeding a preset threshold, or when the braking mode switches, the change is recorded. This recording is required to improve the accuracy of the final stop position prediction.

[0059] Specifically, the stopping and braking process of a woven label machine, especially under a control strategy that pursues a fast and smooth stop, may not apply a single and constant braking force, but rather adopt a phased, adjustable braking scheme. For example, a larger braking force may be used in the early stage to quickly reduce the speed, and the braking force may be reduced or the braking method may be changed in the later stage to achieve a smooth and precise stop. If these changes in braking action are not recorded, and the average deceleration is simply calculated by simply treating the entire braking process as a stage with unchanged deceleration characteristics, then this deceleration value will not accurately reflect the true dynamic behavior of the woven label machine in different braking stages, especially near the node where the braking action switches. This will directly lead to large errors in the remaining motion and final stop position subsequently calculated based on this deceleration.

[0060] By recording changes in the magnitude or mode of the braking action (including the time of the change and the braking parameters after the change), the technical solution of this application can divide the entire complex stop braking process into several braking sub-periods with relatively stable or specific braking characteristics. Each braking sub-period corresponds to a known or measurable braking action state. Then, for each divided braking sub-period, the system can determine the actual operating deceleration of the woven label machine in that sub-period more accurately based on the spindle motion state parameters (such as the rate of change of the spindle speed) collected in real time during that specific sub-period, combined with the specific braking action recorded during that period.

[0061] That is, based on the monitored changes in the magnitude or pattern of the braking action, the parking brake operation is temporally divided into at least one braking sub-period corresponding to the change. For example, if the braking action changes at time points t1 and t2, the entire braking process (from start time t0 to end time tend) can be divided into sub-periods [t0, t1), [t1, t2), and [t2, tend]. Each sub-period corresponds to a relatively stable braking action state.

[0062] For each of the at least one braking sub-period, the actual operating deceleration of the woven label machine during the braking sub-period and under the braking action currently applied to the woven label machine corresponding to the sub-period is determined based on the time-dependent changes in the spindle motion state parameters collected in real time during the braking sub-period. The spindle motion state parameters may include the angular velocity, angular acceleration, or angular position of the spindle. The actual operating deceleration can be determined using a variety of methods, such as calculating the rate of change of the spindle angular velocity within the sub-period, performing a quadratic differential on the spindle angular position data collected within the sub-period, or performing a linear regression analysis on the speed-time data within the sub-period to obtain a slope. By independently determining the deceleration within each sub-period, the deceleration characteristics of the woven label machine under that specific braking action can be more accurately reflected.

[0063] Thus, by segmenting the braking process as the braking action changes and determining the actual operating deceleration for each sub-period, a more accurate description of the deceleration characteristics can be obtained than a single deceleration value calculated based on the entire braking process. This more accurate actual operating deceleration value is used in subsequent calculation steps. For example, when estimating the remaining motion of the woven label machine from the current moment to complete stop, the calculation can be based on the current braking sub-period and its corresponding actual operating deceleration. This improves the accuracy of the remaining motion estimation, thereby making the calculation result of the final stop position closer to the actual situation. The improved accuracy of the final stop position calculation provides a more reliable basis for subsequent braking adjustments based on this calculation result, helping to achieve a more accurate parking of the woven label machine at the desired position.

[0064] Specifically, in some of the above-mentioned schemes of the present application, it is proposed to calculate the remaining motion of the woven label machine from the current moment to the complete stop based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameters at the current moment, and the yarn break state or the current physical position of the key woven component obtained by real-time image information at the current moment, to calculate the final stop position. However, in this process, there is an inherent image acquisition and processing delay in obtaining the yarn break state or the current physical position of the key woven component through real-time image information, resulting in the acquired physical position not being a completely real-time current position. In addition, the confidence of the image recognition process may vary due to factors such as ambient lighting, yarn properties, fabric structure or component occlusion. A low confidence may result in the identified physical position being inaccurate or unreliable. These delays and uncertainties cause deviations in the remaining motion calculated based on the physical position, which in turn affects the accuracy of the final stop position calculation, resulting in an incalculable deviation between the actual stop position and the expected stop position, affecting the yarn break processing efficiency and fabric quality.

[0065] In this regard, the present application further proposes a step of estimating the remaining motion of the woven label machine from the current moment to the complete stop based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the main shaft motion state parameters at the current moment, and the yarn breaking state or the current physical position of the key woven component obtained by the real-time image information at the current moment, including: determining the total delay time for image information acquisition associated with the acquisition process of the current physical position for the yarn breaking state or the current physical position of the key woven component obtained by the real-time image information at the current moment; calculating the image information acquisition delay time based on the total delay time for image information acquisition and the kinematic parameters of the woven label machine during the total delay time for image information acquisition, a physical position compensation value of the current physical position; applying the physical position compensation value to the current physical position to obtain a delay-compensated physical position; obtaining a confidence parameter obtained by performing target recognition on a broken yarn state or a key weaving component through real-time collected image information; determining a final physical position for estimating the remaining motion amount based on whether the confidence parameter of the delay-compensated physical position is lower than a preset validity threshold, and according to a preset confidence processing rule when the confidence parameter is lower than the preset validity threshold; estimating the remaining motion amount of the weaving label machine from the current moment to a complete stop based on the determined actual operating deceleration, the instantaneous motion state of the weaving label machine determined by the spindle motion state parameter at the current moment, and the final physical position.

[0066] For the yarn breakage status or current physical position of a key weaving component, as determined by real-time image information, the total image information acquisition delay associated with the acquisition process of that current physical position is determined. This delay can be determined in advance through system calibration or calculated in real time during runtime by monitoring the timestamps of each step in image acquisition, transmission, and processing. For example, the timestamp of the image sensor exposure start and the timestamp of the position output by the image processing algorithm can be recorded, and the difference between the two can be the total delay time. This delay time typically includes the camera exposure time, image data transmission time, and image processing unit calculation time.

[0067] Among them, one or more image analysis methods can be used to judge the broken yarn status of the woven label through real-time collected image information. The core of the method is to identify the characteristic changes in the weaving area image caused by yarn breakage.

[0068] For example, one implementation is based on the analysis of yarn structural integrity. First, the real-time acquired image is preprocessed, such as by using a filter to enhance the yarn edges and performing a binarization process to separate the yarn from the background. Subsequently, image processing algorithms, such as connected domain analysis or skeleton extraction, can be used to track the path and continuity of each yarn in the field of view. When the algorithm detects that one or more yarns are suddenly interrupted along their expected extension path, that is, the length of the connected domain of the yarn in the image is significantly shorter than the normal length, or a free end point appears at its end that is different from the fabric edge, the system determines that the yarn has broken. This judgment can be further combined with an analysis of the pixel features of the neighborhood near the breakpoint. For example, the yarn end at the breakpoint usually appears fluffy, scattered, or curled, which is significantly different from the straight, continuous shape of normal yarn. By extracting these local texture or morphological features, the occurrence of a yarn break can be further confirmed.

[0069] In order to ensure the accuracy of the control of the stop position of the woven label machine after a yarn breakage event, this application introduces a confidence parameter to quantify the reliability of the yarn breakage status or the physical position of key woven components obtained through image information.

[0070] The confidence parameter is determined based on a comprehensive assessment of the multi-dimensional features of image information in a dynamic and complex weaving environment. Specifically, it is based on intrinsic evaluation metrics derived from the image processing algorithm performing the target recognition task (e.g., image classification or object detection). This metric is typically a probability value or a standardized score output by the algorithm for each recognition result, directly reflecting the algorithm's confidence level in the current judgment. This score is adjusted accordingly in response to disturbances such as yarn flutter or sudden changes in lighting.

[0071] Specifically, the quality of the visual features of the target to be identified in the image (i.e., yarn break points or key weaving components) can be analyzed. For example, by calculating the edge strength, contrast, texture clarity, and whether there are partial occlusions, motion blur, etc. in the target area, it is possible to evaluate whether the current image can provide sufficiently clear and significant visual evidence for reliable identification. Under high-speed operation and complex lighting conditions, the degree of degradation of these features is an important reference for confidence assessment. Furthermore, confidence assessment relies on considering the consistency of recognition results in the temporal sequence. By tracking the target's frequency of occurrence, position stability, and the rationality of its motion trajectory in multiple consecutive frames, it can effectively distinguish between real and ongoing events and artifacts caused by short-term interference in the environment (such as scattered fibers and occasional light spots), thereby giving higher confidence to recognition results with good temporal consistency.

[0072] In addition, the verification of the degree of conformity between the recognition results and the inherent operating parameters of the woven label machine, the kinematic model of key components or known fabric defect patterns can also be integrated. For example, if the component position recognized by the image deviates greatly from the theoretical position calculated by the spindle encoder, or the recognized yarn break feature does not match the common breakage pattern under the current yarn type and tension setting, its confidence level will be lowered accordingly.

[0073] Based on the total delay time for acquiring the image information and the kinematic parameters of the woven label machine during this total delay time, a physical position compensation value is calculated to correct the current physical position. The kinematic parameters of the woven label machine during this delay time, such as instantaneous velocity and acceleration, can be obtained from the spindle motion state parameters. Assuming that the woven label machine moves at an approximately constant speed during the delay time, the compensation value can be simply calculated as the velocity multiplied by the delay time. If acceleration is taken into account, the compensation value can be calculated using the formula for uniformly accelerated linear motion. For example, if the delay time is 50 milliseconds and the instantaneous velocity of the woven label machine is 1 meter per second, the compensation value is approximately 50 millimeters. This compensation value represents the displacement of key components or yarn breakage points of the woven label machine relative to their position at the time of image acquisition during the period from image acquisition to availability.

[0074] The physical position compensation value is applied to the current physical position to obtain the delay-compensated physical position. This is typically accomplished through vector addition, where the calculated compensation value (representing the displacement) is added to the original physical position obtained by image recognition. This provides a more accurate estimate of the yarn breakage state or the actual physical position of a critical weaving component at the current moment.

[0075] The confidence parameter obtained by real-time image data acquisition for target recognition of yarn breakage or key weaving components is obtained. This confidence parameter is the output of the image recognition algorithm and reflects the reliability of the recognition result.

[0076] The delay-compensated physical position is then determined as the final physical position for estimating the remaining motion based on whether the confidence parameter is below a preset validity threshold, and based on preset confidence processing rules when the confidence parameter is below the preset validity threshold. The preset validity threshold can be set based on actual application requirements and system performance, such as 0.7 or 0.8. When the confidence parameter is not below the threshold, the delay-compensated physical position is considered reliable and can be directly determined as the final physical position. When the confidence parameter is below the threshold, it indicates that the image recognition result may have a large error, and the delay-compensated physical position cannot be directly used. The preset confidence processing rules can include various strategies, such as: completely ignoring the image data and instead using the position calculated by the motion model based on the spindle motion state parameters; taking a weighted average of the delay-compensated physical position and the position calculated by the motion model, with the weight related to the confidence level; through this processing, the physical position used for subsequent calculations is ensured to have a higher reliability, reducing the negative impact of low-quality image data.

[0077] Based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine at the current moment, as determined by the spindle motion state parameters, and the final physical position, the remaining motion of the woven label machine from the current moment until it comes to a complete stop is estimated. The actual operating deceleration can be determined based on the time-dependent changes in the spindle motion state parameters. The instantaneous motion state, including parameters such as the instantaneous velocity, is also derived from the spindle motion state parameters. The final physical position is the position after delay compensation and confidence processing. These parameters are input into the motion model, which calculates the total distance traveled by the woven label machine from the current moment until its velocity drops to zero, based on the dynamic characteristics of the woven label machine and the current braking action. For example, the kinematic formula S = V^2 / (2*a) can be used, where S is the remaining motion, V is the instantaneous velocity, and a is the actual operating deceleration. This estimated result is key information for calculating the woven label machine's final stop position. By using the corrected and verified final physical position as a reference point or input parameter for the calculation, the calculated remaining motion is more accurate, thereby improving the accuracy of the final stop position calculation.

[0078] Specifically, in some of the above-mentioned schemes of the present application, it is proposed to obtain the current physical position of the broken yarn state or key weaving components based on the real-time collected image information, and perform delay compensation and confidence processing to determine the final physical position for estimating the remaining motion amount in order to more accurately estimate the remaining motion amount of the weaving machine from the current moment to complete stop. However, in this process, when the confidence parameters of the broken yarn state or key weaving components obtained through image information are low, simply using the position information or adopting a single processing strategy may lead to inaccurate final physical position for estimating the remaining motion amount, thereby affecting the accuracy of the final stop position calculated subsequently, and failing to effectively deal with the confidence fluctuation problem that may occur in image recognition under complex working conditions.

[0079] In this regard, the present application further proposes determining the final physical position for estimating the remaining motion based on whether the confidence parameter of the delay-compensated physical position is lower than a preset validity threshold, and determining the final physical position for estimating the remaining motion based on a preset confidence processing rule when the confidence parameter is lower than the preset validity threshold, including the following steps: when the confidence parameter is lower than the preset validity threshold: obtaining the current value of the confidence parameter and the changing trend of the confidence parameter within a preset time length; selecting at least one alternative physical position determination strategy from a plurality of pre-configured alternative physical position determination strategies based on the obtained current value and changing trend of the confidence parameter and based on pre-configured selection conditions corresponding to the current value and / or changing trend of the confidence parameter, wherein each alternative physical position determination strategy is associated with at least one selection condition, and the setting of the selection condition takes into account the expected processing effects of different alternative strategies under the corresponding confidence parameter value or changing trend; applying the selected at least one alternative physical position determination strategy, in combination with the delay-compensated physical position, to determine the final physical position for estimating the remaining motion; and if the confidence parameter is not lower than the preset validity threshold, determining the delay-compensated physical position as the final physical position for estimating the remaining motion.

[0080] The solution establishes a preset validity threshold for assessing the confidence level of yarn breakage or key woven component positions acquired through image information. When the confidence level is at least this threshold, the image recognition result demonstrates relatively high reliability. The delay-compensated physical position is then used as the final physical position for subsequent calculation of the remaining motion. The delay-compensated physical position is the current physical position acquired from the original image, corrected by a compensation value calculated based on the total delay in the image acquisition process and the kinematic parameters of the woven label machine during that delay. This step is considered sufficiently accurate when the confidence level is high.

[0081] Furthermore, when the confidence parameter falls below a preset validity threshold, it indicates that the reliability of the image recognition results may be reduced, with the risk of increased uncertainty or error. At this point, the system no longer simply relies on the delay compensation physical location, but instead initiates a dynamic processing process. This process first obtains the current value of the confidence parameter and its temporal trend over a preset time period. The current value of the confidence parameter reflects the current level of recognition quality, while its trend (e.g., a continuous decline, fluctuations, a brief decline followed by a rebound, or remaining at a low level) provides dynamic information about the low confidence state. The preset time period can be a fixed duration, such as the last 100 milliseconds, or it can be associated with the operating cycle of the woven label machine. The trend can be determined by calculating the slope or variance of the confidence parameter over the preset time period or through pattern matching.

[0082] Next, based on the current value of the acquired confidence parameter and / or its changing trend, at least one of a plurality of pre-configured alternative physical location determination strategies is selected according to pre-configured selection conditions. The plurality of alternative physical location determination strategies are pre-designed and configured processing methods for different low-confidence situations. For example, alternative strategies may include: a calculation strategy based on historical reliable position data, which calculates the current position based on the past motion trajectory of the woven label machine or known mechanical periodicity; an inference strategy based on spindle motion parameters, which uses information such as the spindle's speed and acceleration combined with the kinematic model of the woven label machine to estimate the position of key components; a filtering strategy, such as a Kalman filter or a moving average filter, which is used to smooth the position sequence and reduce the impact of noise. This strategy can fuse image position data and spindle motion data; or a rule-based correction strategy, which applies a preset correction factor or offset based on the confidence value or trend. The selection condition associates a specific current value range and / or changing trend of the confidence parameter with one or more alternative strategies. For example, one selection condition could be set as "When the confidence parameter is below a threshold and is rapidly decreasing, select the extrapolation strategy based on the spindle motion parameters," while another condition could be set as "When the confidence parameter is below a threshold but fluctuates significantly, select the filtering strategy." These selection conditions fully consider the applicability and expected performance of different alternative strategies under specific confidence conditions, aiming to select the method that provides the most accurate and robust position estimation.

[0083] Subsequently, at least one selected alternative physical position determination strategy is applied, combined with the delay-compensated physical position, to determine the final physical position used to estimate the remaining motion. The term "combination" here means that the alternative strategy does not completely replace the delay-compensated physical position, but rather processes, modifies, or fuses it. For example, if a filtering strategy is selected, the delay-compensated physical position will be one of the inputs to the filter; if a calculation strategy based on historical data is selected, the final position can be a weighted average of the calculated position and the delay-compensated physical position, with the weights dynamically adjusted based on the confidence level; if a correction strategy is selected, the final position can be the delay-compensated physical position plus a correction determined by the alternative strategy. Through this combination, even in cases where image recognition confidence is low, the system can comprehensively utilize image information (after delay compensation) and other information sources or processing methods to obtain a final physical position that is more reliable than directly using the original low-confidence image position.

[0084] When the confidence level of the yarn break state or the position of the key weaving component obtained by the image information is low, this solution dynamically selects and applies an appropriate alternative strategy to process or correct the delay compensation physical position, thereby obtaining a more accurate and reliable final physical position for estimating the remaining movement. This final physical position is then used to estimate the remaining movement of the woven label machine from the current moment to the complete stop. The accurate estimation of the remaining movement is crucial for calculating the final stop position of the woven label machine under the braking action. The accurate calculation of the final stop position enables the system to compare it with the expected stop position and, before the braking is completed, when the deviation meets the adjustment start condition, determine the adjustment parameters and / or the timing of the next adjustment based on the deviation, the deviation change trend and the remaining braking time, and then adjust the braking action applied to the woven label machine. This method of estimating the remaining motion and calculating the final stop position based on more accurate real-time position information, and the braking adjustment based on this, significantly improves the ability of the weaving label machine's actual final stop position to approach the expected stop position, effectively solves the problem of inaccurate stop position caused by fluctuations in image recognition confidence, and improves yarn breakage processing efficiency and product quality.

[0085] Specifically, some of the aforementioned solutions of the present application propose that, during a woven label machine's braking operation in response to a yarn break event, real-time acquisition of at least image information reflecting the yarn break state or the motion state of key weaving components in the woven label machine's weaving area, as well as parameters reflecting the machine's spindle motion state, be used to calculate the machine's final stop position under the braking action based on multimodal data fusion. However, during the real-time acquisition of image information and spindle motion state parameters, different acquisition systems (e.g., the image acquisition system and the spindle motion state parameter acquisition system) may have different inherent processing or transmission delays. Consequently, the image information and spindle motion state parameters acquired simultaneously may actually reflect the machine's state at different points in time. This temporal asynchrony between the multimodal data makes it impossible to accurately reflect the machine's true, comprehensive state at a specific moment in time when the woven label machine's final stop position is subsequently calculated based on this data fusion. This introduces calculation errors, affects the accuracy of the final stop position calculation, and further reduces the effectiveness of braking adjustments based on the calculation results to achieve a precise stop.

[0086] In this regard, the present application further proposes to pre-determine the time delay reference value between the image information acquisition system and the spindle motion state parameter acquisition system of the woven label machine; In the process of the woven label machine performing a stop braking operation in response to a yarn break event of the woven label machine, real-time acquisition includes at least image information reflecting the yarn break state in the weaving area of ​​the woven label machine or the movement state of key weaving components, and adding first time moment information to each data unit of the acquired image information; In the process of the woven label machine performing a stop braking operation in response to a woven label yarn breakage event, parameters reflecting the motion state of the main shaft of the woven label machine are collected in real time, and second time information is added to the data unit of each acquired main shaft motion state parameter; Based on the time delay reference value and in combination with the first time moment information and the second time moment information, the data unit of the image information or the data unit of the main shaft motion state parameter is time-adjusted to obtain image information reflecting the interrupted yarn state of the weaving area of ​​the woven label machine or the motion state of the key weaving components under the same time reference, as well as parameters reflecting the motion state of the main shaft of the woven label machine.

[0087] This technical solution aims to address the potential time asynchrony between multimodal data (image information and spindle motion parameters) collected in real time during the woven label machine's braking process, thereby improving the accuracy of the final stop position calculation based on this data. By synchronizing data from different sources, the data used in subsequent calculations can more accurately reflect the woven label machine's comprehensive status at the same moment.

[0088] Specifically, the solution first establishes a reference for the inherent time difference between the woven label machine's image information acquisition system and the spindle motion state parameter acquisition system by pre-determining the time delay baseline value. This baseline value reflects the average or typical delay difference at the system level. Determining the time delay baseline value can be achieved in a variety of ways. For example, when the woven label machine is in a stable operating state, both acquisition systems can be triggered to collect data simultaneously and record their respective timestamps. By collecting multiple times and comparing the timestamp differences in the corresponding data of the same physical event (such as a key component passing through a specific location) in the two systems, the average delay can be calculated or a delay model can be established as a baseline value. Alternatively, an external high-precision clock source can be used to synchronize the clocks of the two acquisition systems. The delay can then be determined by sending synchronization signals and measuring the response time.

[0089] Next, during the real-time process of the woven label machine performing a stop-and-brake operation, image information and spindle motion state parameters are collected separately. Key to this is the addition of first time information to each acquired image information data unit, and second time information to each acquired spindle motion state parameter data unit. Adding precise real-time timestamps is fundamental to achieving dynamic time synchronization, recording the specific time each piece of data was actually acquired. The first and second time information can be provided by high-precision clocks within their respective acquisition systems, or distributed by a unified system clock. For example, the image acquisition system binds the current system timestamp to the image data when capturing each frame. Similarly, the spindle motion state parameter acquisition system (such as an encoder or sensor interface) also binds the current system timestamp to the parameter data when reading parameters such as spindle angle and speed.

[0090] Finally, based on the pre-determined time delay reference value and in combination with the first time moment information and the second time moment information acquired in real time, the image information data unit or the spindle motion state parameter data unit is time-adjusted. This processing process utilizes the inherent delay characteristics of the system (reference value) and the actual time difference (timestamp) during the real-time acquisition process. Through algorithms, the data is interpolated, extrapolated, or offset, and the data from different acquisition systems are aligned to the same time reference point. For example, the time point of the image data can be adjusted to coincide with the time point of the spindle parameter data, or the time point of the spindle parameter data can be adjusted to coincide with the time point of the image data, or both can be adjusted to a common reference time point.

[0091] Various techniques can be used to perform time alignment. One approach involves selecting one data stream (e.g., spindle parameters) as a time reference. Based on the time-delayed reference and the real-time timestamp, an estimate of the value of the other data stream (e.g., image information) at the same point in time is then calculated. This can be accomplished by interpolating the image data (if the required data point lies between two acquisition points) or extrapolating it (if the required data point lies outside the acquisition range, although extrapolation is typically avoided in real-time systems). Another approach involves aligning both data streams to an intermediate point in time or to an externally synchronized clock. For example, if the image data is acquired at time T1 and the spindle parameters at time T2, and it is known that the image acquisition system is typically delayed by ΔT relative to the spindle parameter acquisition system, the image data timestamp can be adjusted to T1 + ΔT, or the spindle parameter data timestamp can be adjusted to T2 - ΔT, aligning them closer to the actual physical event. Through this time adjustment process, what is finally obtained is the image information and spindle motion state parameters under the same time reference. These synchronized data can more accurately reflect the visual state and mechanical motion state of the weaving label machine at a certain moment.

[0092] The present application further proposes that before the braking of the woven label machine is completed, when the deviation between the final stop position and the expected stop position meets a preset adjustment start condition, the adjustment of the braking effect applied to the woven label machine is started; after the adjustment is started, based on the deviation amount, the deviation amount change trend determined by monitoring the deviation amount change within a preset monitoring time window, and the remaining braking time of the woven label machine from the current moment to the completion of braking, the adjustment parameters for adjusting the braking effect applied by the woven label machine and / or the execution timing of the next adjustment are determined; according to the determined adjustment parameters and / or execution timing, the braking effect applied to the woven label machine is adjusted, and when the adjustment is performed, the execution frequency of the adjustment is limited according to a preset frequency control strategy to balance the correction effect of the deviation amount and control the system load of the woven label machine.

[0093] Specifically, this solution proposes initiating adjustment of the braking action applied to the woven label machine before braking is complete, when the deviation between the calculated final stop position and the desired stop position meets a preset adjustment activation condition. This means that the adjustment process is not continuous but rather conditionally triggered. The adjustment mechanism is only activated when the deviation between the calculated final stop position and the desired position reaches or exceeds a preset threshold. For example, the adjustment can be triggered only when the deviation exceeds a minimum unit length of the fabric or a minimum step angle of the spindle rotation. This deviation threshold-based activation condition avoids unnecessary adjustments when the deviation is small or within an acceptable range, thereby reducing the system's computational and execution burden and improving control efficiency and stability.

[0094] Furthermore, after the adjustment is initiated, the adjustment parameters for adjusting the braking force applied by the woven label machine and / or the timing for executing the next adjustment are determined based on the deviation, the deviation trend determined by monitoring the deviation within a preset monitoring time window, and the remaining braking time from the current moment until the woven label machine completes braking. This system not only relies on the current deviation but also comprehensively considers the deviation trend and the remaining braking time. The deviation itself provides information about the current degree and direction of deviation from the desired position. The deviation trend (e.g., whether the deviation is increasing or decreasing, and the rate of change) reflects whether the current braking effect is insufficient, excessive, or moderate. For example, if the deviation continues to increase during braking, it may indicate insufficient braking force; if the deviation decreases rapidly and may overshoot, it may indicate excessive braking force. The remaining braking time provides a time window for adjustment and the remaining travel distance of the woven label machine. By integrating this information, more precise adjustment parameters can be determined, such as the amount of braking torque to increase or decrease, or the optimal time for the next adjustment, such as when the spindle rotates to a specific angle. This multi-factor decision-making approach allows adjustment parameters and their timing to more precisely match the current braking state and the remaining process. This step receives the deviation information provided by the step that calculates the final stop position and uses this information to guide subsequent braking adjustments, forming a closed feedback loop. In this way, the system can dynamically respond to the actual motion state of the woven label machine during braking, overcome the influence of uncertainties such as mechanical inertia and transmission delay, and improve the control accuracy of the stop position.

[0095] Next, the braking action applied to the woven label machine is adjusted based on the determined adjustment parameters and / or execution timing. This involves actually adjusting the braking action based on the optimized parameters and timing calculated in the previous step. For example, if it is determined that the braking force needs to be increased, a corresponding control signal is sent to the brake to change the braking torque applied to the woven label machine's main shaft. If the next adjustment timing is determined, the system will wait until that time to calculate and apply the next adjustment.

[0096] Furthermore, during adjustments, a preset frequency control strategy limits the frequency of adjustments to balance the effect of deviation correction and control the system load on the woven label machine. During dynamic adjustments, excessively frequent adjustments can lead to system instability, oscillation, and excessive load on mechanical components and the control system. Preset frequency control strategies, such as setting a minimum adjustment interval (e.g., no less than 50 milliseconds) or a maximum number of adjustments (e.g., a maximum of three adjustments during the entire braking process), effectively limit the frequency of adjustments. This ensures effective deviation correction while maintaining smooth system operation, avoiding unnecessary losses and failures, and ensuring the safety and reliability of the entire control process. This frequency control strategy works in conjunction with the adjustment parameters and timing determined in the previous step to ensure the effectiveness and stability of adjustments.

[0097] Specifically, the woven label yarn breakage event type is a yarn breakage defect type obtained by identifying multimodal data during the weaving process of the woven label machine, and includes at least one of warp yarn breakage, weft yarn breakage, single yarn breakage, and multiple yarn breakage.

[0098] Specifically, the key weaving components include one or a combination of knitting needles, reeds, and jacquard needles of a jacquard device.

[0099] Secondly, refer to Figure 2 , this application proposes a woven label production control system, the system includes: An acquisition module 210 is used to acquire an expected stop position of the woven label machine corresponding to the type of the woven label yarn break event, which is preset to deal with the woven label yarn break event; The acquisition module 220 is used to collect in real time at least image information reflecting the yarn break state in the weaving area of ​​the woven label machine or the motion state of key weaving components, as well as parameters reflecting the motion state of the main shaft of the woven label machine during the process of the woven label machine performing a stop braking operation in response to the woven label machine yarn break event; The calculation module 230 is used to calculate the final stopping position of the woven label machine under the braking action based on the real-time collected image information and the spindle motion state parameters and in combination with the braking action currently applied to the woven label machine; The adjustment module 240 is used to compare the final stop position calculated by the final stop position calculation module with the expected stop position obtained by the expected position acquisition module. If the comparison result indicates that the calculated final stop position deviates from the expected stop position, the braking action applied to the woven label machine is adjusted before the braking of the woven label machine is completed, so that the actual final stop position of the woven label machine approaches the expected stop position.

[0100] By obtaining the expected stop position, collecting image information and spindle motion parameters in real time during the braking process, calculating the final stop position, and dynamically adjusting the braking action according to the deviation between the calculated position and the expected position, the method has the advantages of significantly reducing the deviation between the actual stop position and the expected stop position by collecting multimodal data in real time during the braking process of the woven label machine and dynamically adjusting the braking action, so that the woven label machine can be stopped more accurately at a position that is conducive to yarn breakage processing, thereby improving the efficiency of yarn breakage repair, reducing downtime and waste generation, and improving product quality.

[0101] In addition, in some preferred embodiments, a woven label production control system proposed in this application can perform any one of the steps in the above method.

[0102] The above description is merely an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, various modifications and variations of the present application are possible. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A woven label production control method, characterized in that: include: Obtaining an expected stopping position of the woven label machine preset for responding to a woven label yarn breakage event and corresponding to the type of the woven label yarn breakage event; In the process of the woven label machine performing a stop braking operation in response to the woven label yarn breakage event, real-time collection of image information at least including image information reflecting the yarn breakage state in the weaving area of ​​the woven label machine or the motion state of key weaving components, and parameters reflecting the motion state of the main shaft of the woven label machine; Based on the real-time collected image information and the spindle motion state parameters, and in combination with the braking action currently applied to the woven label machine, calculating the final stop position of the woven label machine under the braking action; The calculated final stop position is compared with the expected stop position. If the comparison result indicates that the calculated final stop position deviates from the expected stop position, the braking action applied to the woven label machine is adjusted before the braking of the woven label machine is completed so that the actual final stop position of the woven label machine approaches the expected stop position.

2. A woven label production control method according to claim 1, characterized in that: The step of calculating the final stop position of the woven label machine under the braking action based on the real-time collected image information and the spindle motion state parameters and in combination with the braking action currently applied to the woven label machine includes: During the process of the woven label machine performing a stop braking operation, determining the actual running deceleration of the woven label machine under the braking action currently applied to the woven label machine according to the change of the main shaft motion state parameter collected in real time over time; Based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameter at the current moment, and the yarn breakage state or the current physical position of the key weaving component obtained by the real-time image information collected at the current moment, the remaining motion amount of the woven label machine from the current moment to the complete stop is estimated; The final stopping position of the woven label machine under the braking action is calculated by combining the current physical position with the estimated remaining motion amount.

3. A woven label production control method according to claim 2, characterized in that: The step of determining the actual operating deceleration of the woven label machine under the braking action currently applied to the woven label machine according to the real-time acquired change of the main shaft motion state parameter over time during the process of the woven label machine performing the stop braking operation comprises: During the process of the woven label machine performing a stop braking operation, monitoring the change in the magnitude or mode of the braking action currently applied to the woven label machine; Based on the monitored change in the magnitude or mode of the braking action, dividing the parking braking operation process in time into at least one braking sub-interval corresponding to the change; For each of the at least one braking sub-period, based on the change of the spindle motion state parameters collected in real time during the braking sub-period over time, the actual operating deceleration of the woven label machine during the braking sub-period and under the braking action currently applied to the woven label machine corresponding to the braking sub-period is determined.

4. A woven label production control method according to claim 2, characterized in that: The step of calculating the remaining motion amount of the woven label machine from the current moment to the complete stop based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameter at the current moment, and the yarn breakage state or the current physical position of the key weaving component obtained by the real-time image information collected at the current moment includes: determining, for the yarn breakage state or the current physical position of the key weaving component acquired through the real-time collected image information at the current moment, a total image information acquisition delay time associated with an acquisition process of the current physical position; Calculating a physical position compensation value for correcting the current physical position based on the total delay time for acquiring the image information and the kinematic parameters of the woven label machine during the total delay time for acquiring the image information; Applying the physical position compensation value to the current physical position to obtain a delay-compensated physical position; Acquiring a confidence parameter obtained by performing target recognition on the yarn breakage state or the key weaving component through the real-time collected image information; determining a final physical position for estimating the remaining motion amount based on whether the delay-compensated physical position is lower than a preset validity threshold and, when the confidence parameter is lower than the preset validity threshold, according to a preset confidence processing rule; Based on the determined actual operating deceleration, the instantaneous motion state of the woven label machine determined by the spindle motion state parameter at the current moment, and the final physical position, the remaining motion amount of the woven label machine from the current moment to complete stop is calculated.

5. A woven label production control method according to claim 4, characterized in that: The step of determining the delay-compensated physical position based on whether the confidence parameter is lower than a preset validity threshold, and determining the final physical position for estimating the remaining motion amount according to a preset confidence processing rule when the confidence parameter is lower than the preset validity threshold includes: When the confidence parameter is lower than the preset validity threshold: Obtaining a current value of the confidence parameter and a changing trend of the confidence parameter within a preset time period; selecting, based on the obtained current value of the confidence parameter and the change trend, at least one alternative physical location determination strategy from a plurality of pre-configured alternative physical location determination strategies in accordance with a pre-configured selection condition corresponding to the current value of the confidence parameter and / or the change trend, wherein each of the alternative physical location determination strategies is associated with at least one of the selection conditions, and the setting of the selection conditions takes into account expected processing effects of different alternative strategies under corresponding confidence parameter values ​​or change trends; Applying the selected at least one candidate physical position determination strategy, combined with the delay-compensated physical position, to determine the final physical position for estimating the remaining motion; If the confidence parameter is not lower than the preset validity threshold, the delay-compensated physical position is determined as the final physical position for calculating the remaining motion.

6. A woven label production control method according to claim 1, characterized in that: The step of collecting, in real time, image information at least including image information reflecting the yarn break state or the motion state of key weaving components in the weaving area of ​​the woven label machine and parameters reflecting the motion state of the main shaft of the woven label machine during the woven label machine performing a stop braking operation in response to the woven label yarn break event comprises: Predetermining a time delay reference value between the image information acquisition system of the woven label machine and the spindle motion state parameter acquisition system; In the process of the woven label machine performing a stop braking operation in response to the woven label yarn breakage event, real-time acquisition of image information at least including image information reflecting the yarn breakage state in the weaving area of ​​the woven label machine or the movement state of key weaving components, and adding first time information to each data unit of the acquired image information; In the process of the woven label machine performing a stop braking operation in response to the woven label yarn breakage event, parameters reflecting the motion state of the main shaft of the woven label machine are collected in real time, and second time information is added to the data unit of each acquired main shaft motion state parameter; Based on the time delay reference value and in combination with the first time moment information and the second time moment information, the data unit of the image information or the data unit of the spindle motion state parameter is time-adjusted to obtain image information reflecting the interrupted yarn state or the motion state of the key weaving component in the weaving area of ​​the woven label machine under the same time reference, as well as parameters reflecting the motion state of the spindle of the woven label machine.

7. The woven label production control method according to claim 1, characterized in that: The step of adjusting the braking effect applied to the woven label machine before the woven label machine is braked comprises: Before the woven label machine is braked, when the deviation between the final stop position and the desired stop position satisfies a preset adjustment start condition, starting the adjustment of the braking action applied to the woven label machine; After the adjustment is initiated, determining adjustment parameters for adjusting the braking effect applied by the woven label machine and / or a timing for executing the next adjustment based on the deviation amount, a variation trend of the deviation amount determined by monitoring a variation of the deviation amount within a preset monitoring time window, and a remaining braking time of the woven label machine from a current moment to completion of braking; The braking action applied to the woven label machine is adjusted according to the determined adjustment parameters and / or the execution timing. Furthermore, when performing the adjustment, the execution frequency of the adjustment is limited according to a preset frequency control strategy to balance the correction effect of the deviation and control the system load of the woven label machine.

8. The woven label production control method according to claim 1, characterized in that: The woven label yarn breakage event type is a yarn breakage defect type obtained by identifying multimodal data during the weaving process of the woven label machine, and includes at least one of warp yarn breakage, weft yarn breakage, single yarn breakage, and multiple yarn breakage.

9. The woven label production control method according to claim 1, characterized in that: The key weaving components include one or a combination of knitting needles, reeds, and jacquard needles of a jacquard device.

10. A woven label production control system, characterized in that: The system includes: An acquisition module is used to acquire an expected stop position of the woven label machine that is preset to deal with a woven label yarn breakage event and corresponds to the type of the woven label yarn breakage event; an acquisition module for acquiring, in real time, image information reflecting at least a yarn break state in a weaving area of ​​the woven label machine or a motion state of a key weaving component, and parameters reflecting a motion state of a main shaft of the woven label machine during a process in which the woven label machine performs a stop braking operation in response to the woven label yarn break event; a calculation module, configured to calculate a final stopping position of the woven label machine under the braking action based on the real-time collected image information and the spindle motion state parameters and in combination with the braking action currently applied to the woven label machine; an adjustment module, configured to compare the final stop position calculated by the final stop position calculation module with the expected stop position obtained by the expected position acquisition module; if the comparison result indicates that the calculated final stop position deviates from the expected stop position, then, before the braking of the woven label machine is completed, adjust the braking action applied to the woven label machine so that the actual final stop position of the woven label machine approaches the expected stop position.

Images