Dynamic detection and feature extraction method of textile spindle speed

Abstract

The invention discloses a method for dynamic detection and feature extraction of textile spindle speed, relates to the technical field of textile spindle speed detection, and is used to solve the problem of poor detection accuracy of textile spindle speed. The invention installs a marker with high contrast and periodic texture on the upper end of a yarn tube, and optimizes its geometric shape, material and installation angle, so that the image is stable and clear during high-speed rotation, the detection accuracy of weak twist phenomenon is improved, the exposure time and sampling rate are dynamically adjusted, and images can be stably captured at different rotation speeds, and the blur and noise problems in dynamic scenes in traditional methods are overcome. Dynamic features are effectively extracted by denoising, edge detection and rotation feature analysis of the collected image, and weak twist phenomenon is accurately judged by marker position matching and stability evaluation. In addition, based on the image analysis result, the textile detection strategy is adjusted in combination with the generated signal, and instant feedback and adjustment are provided, thereby improving the quality control efficiency and equipment response speed in the production process.

Description

Technical Field

[0001] The present invention relates to the technical field of textile spindle speed detection, and more specifically, to a textile spindle speed dynamic detection and feature extraction method. Background Art

[0002] The spinning spindle speed is a dynamic parameter in the spinning process, which affects the formation of yarn twist and the stability of quality. During the spinning process, the synchronous rotation of the bobbin and the spindle twists the yarn to form a certain twist. The change of spindle speed directly affects the yarn tension, which in turn affects the strength and uniformity of the yarn, resulting in the appearance of weak twist.

[0003] Weak twist is a common abnormal phenomenon in the spinning process, which refers to the insufficient synchronization of the rotation speed between the yarn tube and the spindle, resulting in low yarn twist. Yarn twist is a key factor affecting the quality of textiles. Insufficient twist can lead to uneven yarn tension, reduced yarn strength and loose textile structure, thus affecting the quality and market competitiveness of the product.

[0004] The shortcomings of the existing technology are as follows: the spinning process is carried out under high-speed rotation. During the image acquisition process of spinning, image blur and data loss are prone to occur due to the asynchronous acquisition of the stroboscope and the rotation speed of the yarn tube, which affects the reliability of weak twist judgment. The traditional image processing and dynamic feature extraction are not accurate enough to effectively identify the synchronization difference between the yarn tube and the spindle, resulting in too many defective products and affecting the benefits. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, there is a solution as follows to solve the problem of poor accuracy in detecting the spinning spindle speed in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] The method for dynamic detection and feature extraction of textile spindle speed comprises the following steps:

[0008] Install a marker with high contrast and periodic texture on the top of the bobbin and optimize the marker geometry, material and installation angle;

[0009] Set up the camera and stroboscope, dynamically adjust the exposure time and sampling rate, and collect images at different rotation speeds;

[0010] De-noising, edge detection and rotation feature analysis are performed on the collected images. Dynamic features are extracted through marker position matching and stability assessment. The quality of the markers is comprehensively assessed to determine whether weak twist occurs.

[0011] The image extraction information generated during the evaluation process of the weak twist phenomenon is obtained and analyzed, and the textile detection strategy is adjusted according to the different signals generated by the analysis.

[0012] In a preferred embodiment, a marker with high contrast and periodic texture is installed on the upper end of the bobbin, and the marker geometry, material and installation angle are optimized. The specific steps are as follows:

[0013] Add a cap-shaped mark on the top of the bobbin, the shape of which is designed to be a triangle or a QR code, and add periodic stripes or textures on the surface of the mark, and calculate the edge sharpness, reflectivity, and coverage of the cap-shaped mark;

[0014] Use a high-frame-rate camera and stroboscope combination to dynamically shoot the mark, record the mark clarity at different rotation speeds, and adjust the stroboscope flashing frequency according to the mark's dynamic blur and light reflection intensity;

[0015] The overall quality of the marking is evaluated by comprehensively considering edge sharpness, reflectivity, coverage and dynamic blur, and a marking scoring model is established to determine the optimal marking parameters and layout scheme.

[0016] In a preferred embodiment, a camera and a stroboscope are set, the exposure time and the sampling rate are dynamically adjusted, and image acquisition is performed at different rotation speeds. The specific steps include:

[0017] The stroboscope frequency calculation and synchronization initialization are performed, the rotation speed data of the bobbin is obtained through the sensor, the optimal flashing frequency of the stroboscope is calculated according to the bobbin rotation speed, and the flashing time of the stroboscope is dynamically adjusted to make each flashing interval consistent with the rotation of the marker;

[0018] According to the stroboscope frequency, set the camera sampling rate to synchronize with the stroboscope, adjust the camera exposure time and aperture size, and detect inter-frame jitter;

[0019] Perform clarity detection on the real-time collected images and evaluate the edge blurriness of the mark. If the blurriness exceeds the blurriness threshold, adjust the stroboscope flashing frequency or camera sampling rate and recalibrate the synchronization;

[0020] When the image edge blur is greater than the image edge blur threshold, the stroboscope frequency is increased or the exposure time is adjusted;

[0021] When the image edge blurriness is less than or equal to the image edge blurriness threshold, the current frequency and exposure settings remain unchanged.

[0022] In a preferred embodiment, the collected image is subjected to denoising, edge detection and rotation feature analysis, dynamic features are extracted through mark position matching and stability evaluation, and the quality of the mark is comprehensively evaluated to determine whether weak twist occurs. The specific steps are as follows:

[0023] Use median filtering to denoise, replacing each pixel value with the median value within the pixel neighborhood;

[0024] Extract the marked edge information in the image through the Canny edge detection algorithm;

[0025] By analyzing the rotation features of the markers in the image sequence, the dynamic behavior of the markers is extracted. The dynamic feature behaviors of the markers include the rotation speed and the rotation angle changes.

[0026] Based on the temporal continuity of the image sequence, the rotation speed of the marker is calculated, and the stability of the rotation state is determined according to the rate of change of the rotation angle;

[0027] The dynamic blur is calculated using the gradient information of the image edge, the local mean shift algorithm is used to evaluate the blur of each frame, and the dynamic blur value is compared with the threshold.

[0028] Analyze the tracking stability of the marker by changing the marker position between consecutive frame images;

[0029] The quality of the marker is comprehensively evaluated by combining the rotation speed, ambiguity, and tracking stability of the marker to obtain the final marker quality score and determine the weak twist phenomenon.

[0030] In a preferred embodiment, the quality of the mark is comprehensively evaluated to determine whether weak twist occurs, including the following steps:

[0031] The marking quality scores were compared with the threshold value for weak twist phenomenon;

[0032] If the mark quality score is less than the weak twist phenomenon threshold, it is determined that the weak twist phenomenon occurs and the mark is analyzed.

[0033] In a preferred embodiment, obtaining and analyzing image extraction information generated during the evaluation of the weak twist phenomenon includes the following steps:

[0034] Acquire image extraction information generated during the evaluation process of the weak twist phenomenon, wherein the image extraction information includes dynamic feature information and image scale information;

[0035] The dynamic feature information includes a rotation dynamic feature index, and the image scale information includes an image clarity index;

[0036] The obtained rotation dynamic characteristic index and image clarity index are combined to generate a textile determination coefficient;

[0037] The rotation dynamic characteristic index, image clarity index and textile determination coefficient are proportional.

[0038] In a preferred embodiment, adjusting the textile detection strategy according to the different signals generated by the analysis includes the following steps:

[0039] comparing the generated prediction adjustment coefficient with the set textile state determination threshold;

[0040] If the predicted adjustment coefficient is greater than or equal to the textile state determination threshold, a textile detection stability signal is generated, indicating that the textile process is normal and no additional intervention is required;

[0041] If the predicted adjustment coefficient is less than the textile state determination threshold, a textile detection state abnormality signal is generated to remind the production personnel or system operators that measures need to be taken to intervene in the textile process.

[0042] The technical effects and advantages of the textile spindle speed dynamic detection and feature extraction method of the present invention are as follows:

[0043] The present invention achieves efficient detection of weak twist phenomenon in the textile process by installing a marker with high contrast and periodic texture on the upper end of the yarn tube and combining it with an accurate image acquisition and processing method. By optimizing the geometric shape, material and installation angle of the marker, it is ensured that the marker has a stable and clear image during high-speed rotation, thereby improving the detection accuracy. The camera and the stroboscope work together to dynamically adjust the exposure time and sampling rate, and can stably capture images at different rotation speeds, thus overcoming the blur and noise problems that are prone to occur in traditional methods in dynamic scenes. The collected images are denoised, edge detected and analyzed for rotation features, and dynamic features can be effectively extracted. The presence of weak twist phenomenon can be accurately determined through marker position matching and stability evaluation. In addition, based on the analysis of image extraction information, the textile detection strategy is adjusted in combination with the generated signal, thereby achieving instant feedback and adjustment of targeted problems, improving the quality control efficiency in the production process, and greatly improving the detection accuracy and response speed of textile equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 It is a schematic flow chart of the method for dynamic detection and feature extraction of textile spindle speed of the present invention. DETAILED DESCRIPTION

[0045] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0046] In order to achieve the above objectives, Figure 1A structural schematic diagram of a method for dynamic detection and feature extraction of textile spindle speed according to the present invention is given, which specifically comprises the following steps:

[0047] Install a marker with high contrast and periodic texture on the top of the bobbin and optimize the marker geometry, material and installation angle;

[0048] Set up the camera and stroboscope, dynamically adjust the exposure time and sampling rate, and collect images at different rotation speeds;

[0049] De-noising, edge detection and rotation feature analysis are performed on the collected images. Dynamic features are extracted through marker position matching and stability assessment. The quality of the markers is comprehensively assessed to determine whether weak twist occurs.

[0050] The image extraction information generated during the evaluation process of the weak twist phenomenon is obtained and analyzed, and the textile detection strategy is adjusted according to the different signals generated by the analysis.

[0051] Step 1: Optimize the design and installation of textile markers. Add a cap-shaped marker to the top of the bobbin. Combine the camera and stroboscope to optimize the marker design and installation layout to ensure the recognition accuracy of the marker in the dynamic detection of textile spindle speed. The specific steps are as follows:

[0052] Add a cap-shaped mark on the top of the bobbin. Its shape is designed as a high-contrast asymmetric geometric pattern (such as a triangle or a QR code), and add periodic stripes or textures on the surface of the mark to produce stable dynamic features when rotating, which is easy for the camera to capture and mark the edge sharpness. The formula is: ,in, is the brightness distribution of the marker, and is the edge interval, and the edge sharpness is the edge gradient change rate per unit radian:

[0053] Use materials with high reflectivity (such as coated plastics or reflective coatings) to make the mark to reduce the interference of ambient light changes. Apply a fluorescent coating on the surface of the mark to improve the recognition ability in dark environments. The reflectivity of the mark is defined as the intensity of the light reflected by the mark. With the incident light intensity Ratio of: , reflectivity Directly affects the visibility of the mark in complex lighting environments;

[0054] Arrange and install the markers, arrange multiple markers symmetrically on the upper end of the bobbin to ensure that at least one marker is within the camera's field of view at any rotation angle, adjust the angle between the marker and the bobbin surface to maintain stable optical properties during rotation, reduce visual distortion caused by rotation, and mathematically model the layout position of each marker to optimize the coverage of the marker layout;

[0055] The coverage rate C is defined as the probability of being within the camera's field of view, and the calculation formula is: ,in, is the field of view coverage angle of the i-th marker, and n is the number of markers;

[0056] Use a high frame rate camera and stroboscope combination to dynamically shoot the mark to ensure that the dynamic features of the mark are clearly visible during rotation, that is, test the mark clarity at different rotation speeds, analyze key indicators such as mark blur, light reflection intensity, optimize the stroboscope flashing frequency, keep it consistent with the bobbin rotation speed, reduce motion blur, and dynamic blur is the rate of change of boundary displacement, and the formula is: ,in, is the marker boundary displacement function, is the static marker boundary position, T is the sampling time;

[0057] For markers that perform poorly in the test (such as low recognition rate or insufficient reflectivity), dynamically adjust their geometric shape or material parameters, and iteratively optimize the placement and quantity of the markers based on the data analysis results, so that the recognition rate of the markers in the camera and stroboscope combined shooting system reaches the target value (such as ≥95%);

[0058] The comprehensive quality of the marking is evaluated by integrating parameters such as edge sharpness, reflectivity, coverage, and dynamic blur, and a marking scoring model is established to output the optimal marking parameters and layout scheme to the subsequent image processing steps. The comprehensive scoring model is: ,in, , , , is the weight coefficient, which can be adjusted according to actual needs.

[0059] Step 2: Perform image acquisition and stroboscope synchronous control. After the mark design and installation optimization are completed, the high frame rate camera and stroboscope are synchronously controlled to ensure the clarity and stability of image acquisition and provide high-quality raw data for subsequent image processing and dynamic feature extraction. The specific steps are as follows:

[0060] Initialize the stroboscope frequency calculation and synchronization to obtain the bobbin speed data (Unit: rpm), provided by the sensor or the initial setting value, the optimal flashing frequency of the stroboscope is calculated according to the bobbin speed , to match the dynamic period of the marker and set the dynamic adjustment stroboscope flash time , to ensure that each flashing interval is consistent with the rotation of the marker, the formula is: , the flashing frequency of the stroboscope and time interval It is used for strobe control and frequency adjustment;

[0061] Set the camera sampling rate according to the stroboscope frequency , ensure synchronization with the stroboscope, and dynamically adjust the camera exposure time and aperture size to optimize the image acquisition quality under light conditions. The formula is: , where k is the sampling redundancy coefficient;

[0062] The exposure time formula is: ,in, For the ideal light intensity, is the ambient light intensity, is the exposure correction factor;

[0063] The exposure time is directly related to the stroboscopic interval. It is necessary to ensure that the mark collected in each frame is stable and of moderate brightness. The sampling redundancy factor is added to the calculation of the sampling rate to ensure the complete recording of the mark. Even if there is a slight fluctuation in the rotation speed, no frame is lost.

[0064] Perform clarity detection on the real-time collected image and evaluate the edge blurriness of the mark. If the blurriness exceeds the blurriness threshold, adjust the stroboscope flashing frequency or camera sampling rate and recalibrate the synchronization. The formula is: ,in, is the image edge blur, For the image at pixel points The brightness value of, N is the total number of pixels in the image;

[0065] when When the image edge blur is greater than the image edge blur threshold, increase the stroboscope frequency or adjust the exposure time;

[0066] when When the image edge blurriness is less than or equal to the image edge blurriness threshold, the current frequency and exposure settings remain unchanged;

[0067] The brightness gradient change in the fuzziness detection formula reflects the sharpness of the edge in the image. Clearly marked edges help improve the recognition rate.

[0068] When the bobbin rotates at high speed, the marked image may produce inter-frame jitter due to equipment vibration or synchronization error, which will lead to instability of the marked position in the continuous image. Through inter-frame position analysis and jitter correction algorithm, dynamic deviation in the image sequence is eliminated to ensure data consistency. That is, through real-time acquisition of multiple frames of images, the position stability of the mark is analyzed, inter-frame jitter is detected, and image registration algorithm (such as optical flow method) is used to correct jitter to ensure the consistency of the image sequence: ,in, To mark the position stability, is the position coordinate of the marker at time t, when When the inter-frame jitter correction is started, is the marker position stability threshold;

[0069] In the formula of marker position stability, the position change rate reflects the dynamic stability of the marker on the time axis. The stable image sequence is crucial for subsequent weak twist detection and dynamic feature analysis.

[0070] This step solves the image blur and position offset problems of the bobbin dynamic mark under high-speed rotation through technical means such as synchronous control of the stroboscope and camera, clarity detection, and jitter correction, and ultimately outputs high-quality image sequences and accurate acquisition parameters.

[0071] Step 3: Perform dynamic feature extraction and mark recognition. After image acquisition and synchronization control are completed, dynamic feature extraction and mark recognition are performed on the acquired image sequence. The specific steps are as follows:

[0072] In the image sequence, due to the possible presence of environmental noise or sensor errors, denoising processing is required. Gaussian filtering or median filtering can be used to eliminate noise in the image to ensure the accuracy of subsequent feature extraction. The specific processing steps are as follows:

[0073] Use median filtering to denoise, replacing each pixel value with the median value in the pixel neighborhood;

[0074] Through edge detection algorithms, such as Canny edge detection, the edge information of the markers in the image is extracted to further improve the recognizability of the image. Edge detection helps to determine the exact position and contour of the marker. The formula of the Canny edge detection algorithm is: ,in, For the image The brightness function of the position, is the second-order partial derivative of the image;

[0075] By analyzing the rotation features of the markers in the image sequence, the dynamic behavior of the markers, such as rotation speed, rotation angle change, etc., can be extracted. Based on the temporal continuity of the image sequence, the rotation speed of the marker can be calculated. , and judge the stability of the rotation state according to the rate of change of the rotation angle. The formula is: ,in, For time interval The angle change of the inner marker rotation;

[0076] Usually, in image processing, the image brightness function is used to represent the brightness (or intensity) value of each pixel in the image. The brightness value is usually a grayscale value, which reflects the brightness of a certain pixel. In a color image, it may represent the brightness values ​​of the red, green, and blue channels respectively. In textile inspection, the image brightness function reflects the light intensity information of each pixel in the image and is used to capture the characteristics of textile markings. If the image clarity is high, the brightness function will provide a clear, contrasty marking image;

[0077] The evaluation of dynamic blur is used to determine whether the marker is over-blurred during high-speed rotation. According to the definition in step 1, the gradient information of the image edge is used to evaluate the blur. The local mean shift (LMD) algorithm can be used to evaluate the blur of each frame of the image and compare the blur value with the threshold. The formula for dynamic blur is: ,in, is the dynamic displacement function of the marker, is a static position;

[0078] The dynamic displacement function is the displacement of the mark in the rotating system at a certain time point t. Specifically, the mark may be a small physical mark or feature point, which changes position with rotation or other dynamic changes during the textile production process. The dynamic displacement of the mark is a mathematical function that describes the motion trajectory of the mark over time. It may reflect the rotation dynamics of the bobbin, the changing speed, or the displacement anomaly caused by process problems (such as weak twist). For example, if the mark is a small object placed on the bobbin, the dynamic displacement function may represent the displacement change of the mark in a certain direction (such as axial or radial) during the rotation process. This function helps to analyze the stability and irregularity of the rotation process;

[0079] The mark in each frame image is matched with features, and the template matching technology is used to extract the position information of the mark by comparing it with the preset mark template. The template matching formula is: ,in, is the image to be recognized, is the template image, For the matching degree,

[0080] By analyzing the marker position changes between consecutive frame images, the tracking stability of the marker is analyzed to ensure the consistency of the marker position in the image sequence, and the marker position is estimated in real time. The formula of the optical flow method is: , where v is the motion vector of the pixel, is the gradient of the image;

[0081] Combined with multiple dynamic features such as rotation speed, ambiguity, tracking stability, etc., the quality of the marker is comprehensively evaluated. The scores of each feature are combined through the weighted fusion method to obtain the final marker quality score. The weighted average method or more complex fuzzy logic method can be used for fusion, as follows:

[0082] The marking quality scoring formula is: in, Score the marker tracking stability, is the weight coefficient of each feature;

[0083] According to the dynamic features extracted in step 3, analyze whether the dynamic behavior of the marker meets the criteria of weak twist phenomenon. For example, the slow and stable rotation speed of the marker, high fuzziness, unstable rotation angle change, etc. can all be regarded as indicators of weak twist phenomenon;

[0084] The criteria for determining the occurrence of weak twist can be defined as: ,in, is the preset threshold value of weak twist phenomenon;

[0085] After the dynamic feature extraction and marker recognition phases are completed, once it is determined that some weak twisting occurs, it is necessary to conduct a comprehensive analysis of the multi-dimensional features of the marker, such as rotational features (rotation speed), ambiguity, and tracking stability, so as to determine whether there will indeed be weak twisting in the future, and trigger corresponding alarms or adjustment operations based on the judgment results;

[0086] If the weak twist phenomenon is limited to a very small part of the textile, and these parts will not significantly affect the final quality of the textile during the overall production process, then such phenomena may be ignored, especially in mass production, where local irregularities may not affect the functionality or appearance of the final product. If the weak twist phenomenon occurs in multiple parts or over a continuous period of time, and the quality problems of these parts will significantly affect the strength, softness or other functional indicators of the final product, then it should not be ignored, even if these phenomena appear to be relatively minor at the initial stage, so further testing is required to determine;

[0087] Use the extracted features (such as rotation speed, fuzziness, tracking stability, etc.) to comprehensively evaluate the weak twist phenomenon that occurs, evaluate and analyze various dynamic features, and obtain image extraction information generated during the evaluation process of the weak twist phenomenon. The image extraction information includes dynamic feature information and image scale information;

[0088] The dynamic feature information includes the rotation dynamic feature index and is calibrated as XZD, and the image scale information includes the image clarity index and is calibrated as QXD;

[0089] The rotation dynamic characteristic index is used to measure the dynamic behavior characteristics of the yarn tube during the rotation process, especially the change of its rotation rate, stability and possible nonlinear dynamic phenomena (such as oscillation, sudden acceleration or deceleration, etc.). It combines the time domain characteristics and frequency domain characteristics of the rotation process to fully capture the subtle changes and complex dynamics of the rotation process, especially the behavior patterns related to the weak twist phenomenon;

[0090] By monitoring the changes in the rotation rate, especially frequent fluctuations or abnormal speed changes, the stability of the rotation process is reflected. If the rotation rate changes dramatically or irregularly, it may indicate that there is an abnormal phenomenon in the rotation process, such as weak twist:

[0091] In the evaluation of weak twist, the rotational dynamic characteristic index acts as a dynamic behavior monitoring tool, which can reveal potential abnormal changes in the tube rotation process. Weak twist usually manifests as sudden changes in the tube rotation rate or unstable behavior, which can be quantified and detected by the rotational dynamic characteristic index.

[0092] Weak twist phenomenon is usually accompanied by unstable rotation of the yarn tube, such as a sudden increase or decrease in the rotation rate, or frequent fluctuations in the rotation rate. The rotation dynamic characteristic index can capture these abnormal dynamics through time-frequency analysis and nonlinear dynamic characteristics, thereby discovering weak twist phenomenon in advance; the rotation dynamic characteristic index can also be used in combination with the image clarity index to jointly judge the occurrence of weak twist phenomenon. When the rotation is unstable or the image blur increases, it is usually accompanied by the occurrence of weak twist phenomenon. By comprehensively considering the two, the accuracy and reliability of weak twist phenomenon evaluation can be improved;

[0093] The logic for obtaining the rotational dynamic characteristic index is as follows:

[0094] The rotation speed data XZ(t) is collected by the rotation sensor and the data is divided into time windows ; Perform short-time Fourier transform on the collected rotation speed signal to extract the frequency domain characteristics of the rotation signal. The formula is: , where is a window function, which is used to limit the signal components near time t. is a complex exponential function, representing the frequency component, is a time variable, used to represent the instantaneous value of the signal, and f is a frequency variable, representing the frequency in the frequency domain;

[0095] For rotation speed data The Lyapunov exponent is calculated using the formula: , the rotational dynamic characteristic index is calculated based on the frequency domain characteristics of STFT and the Lyapunov index, and the formula is as follows: ;

[0096] It should be noted that the Lyapunov index is used to measure the sensitivity and chaotic characteristics of the rotation speed. The index describes the stability of the speed change. The larger the Lyapunov index, the more unstable the system is and the weak twist phenomenon may occur.

[0097] The image clarity index (QXD) measures the clarity, sharpness and visibility of the target object (in this case, the mark on the bobbin) in the image. It evaluates the overall quality and detail of the image by analyzing the edge sharpness, contrast, noise level and texture characteristics in the image. A higher clarity image means that the edges of the mark are sharper, the contrast is stronger and it is less susceptible to noise interference;

[0098] The clarity of an image is closely related to the sharpness of its edges. A clear image will show distinct edges, while a blurred image will make the edges unclear and the transition blurry. Sharpness is usually quantified by edge gradient, second-order derivative and other methods. A clear mark means that the edge changes quickly and can be accurately identified.

[0099] The image clarity index serves as a key visual quality inspection tool in the evaluation process of weak twist phenomenon, helping to evaluate whether the dynamic characteristics of the bobbin mark can be clearly captured, which is crucial for the accurate identification and determination of weak twist phenomenon. Weak twist phenomenon is usually accompanied by the instability of bobbin rotation. The fluctuation of rotation rate will directly affect the clarity of the image, especially at high-speed rotation, the image may be blurred or distorted. The image clarity index monitors the image quality in real time to ensure that the image is clear enough to capture the slight changes of the mark when extracting dynamic features;

[0100] The marker will produce dynamic blur during high-speed rotation, especially when weak twist occurs, the rotation of the marker may be unstable, causing the marker to be blurred or distorted in the image. The image clarity index can help detect these blurred areas in real time and guide the system to adjust the shooting parameters (such as exposure time, aperture, etc.) to optimize the image quality and ensure that the marker always remains clear, so as to more accurately evaluate the weak twist phenomenon;

[0101] The logic for obtaining the image clarity index is as follows:

[0102] Perform wavelet transform on the marker image captured by the camera to extract image features at different frequencies in a multi-scale manner;

[0103] The edge information of the image is extracted through Canny edge detection, with special attention paid to the sharpness of the marked edge. The edge strength and distribution characteristics are used to evaluate the image clarity, and the expression is: ,in, is the wavelet transform coefficient, is the wavelet basis function;

[0104] The image clarity of the edge information of the image is obtained, and the image gradient value is calculated as follows: , where represents the gradient of brightness, and Respectively represent the rate of change of the image in the horizontal and vertical directions, and use the high-order gradient fuzziness metric to calculate the nonlinear fuzziness value of the image. The calculation expression is: , where , Respectively represent the position of the i-th pixel in the horizontal and vertical directions, and N is the total number of pixels;

[0105] Calculate the image clarity index, the calculation expression is: .

[0106] It should be noted that wavelet transform can capture local details, texture and edge information in images; high-order gradient blur metric is used to introduce the nonlinear blur degree of the image. This blur metric captures the high-order blur information in the image through the second-order derivative. The stronger the edge blur, the worse the image quality.

[0107] The dynamic feature information and image scale information are combined to generate the textile determination coefficient, that is, the rotation dynamic feature index and image clarity index are obtained to generate the textile determination coefficient. The expression is: , where , is the preset proportional coefficient of the rotation dynamic characteristic index and the image clarity index, and , Both are greater than 0.

[0108] The specific method of jointly generating a textile determination coefficient may involve a variety of algorithms and models, which depends on the actual situation and application requirements. In this embodiment, a weighted summation method can be used to combine the rotational dynamic characteristic index and the image clarity index to generate a comprehensive textile determination coefficient. This textile determination coefficient can be used as an input for a comprehensive evaluation of the weak twist phenomenon and is used to determine the final weak twist phenomenon.

[0109] It should be noted that the size of the preset proportional coefficient is a specific value obtained by quantifying each parameter. In order to facilitate subsequent comparison, the size of the coefficient depends on the amount of sample data and the preset proportional coefficient initially set by technical personnel in this field for each group of sample data. It is not unique, as long as it does not affect the proportional relationship between the parameter and the quantized value. For example, the rotational dynamic characteristic index is proportional to the textile determination coefficient. The rotational dynamic characteristic index and the image clarity index are normalized to have the same dimension and range. This can be achieved by subtracting the mean from the original data and dividing it by the standard deviation, or mapping the data to the range of [0, 1].

[0110] The larger the rotation dynamic feature index and the larger the image clarity index, the larger the textile determination coefficient generated by the combination, indicating that the rotation stability of the bobbin is good when rotating at high speed, the change law of the marked dynamic feature is clear, there is no obvious abnormal fluctuation or instability, the rotation process of the textile is stable, the image acquisition quality is good, and the detailed information of the textile can be accurately obtained during the detection process, thereby obtaining an accurate quality assessment, which usually indicates that the quality of the textile is high and meets the production standards;

[0111] The smaller the rotation dynamic feature index and the image clarity index, the smaller the textile determination coefficient generated by the combination, indicating that there may be unstable factors in the rotation of the bobbin, such as large fluctuations in the rotation speed and irregular changes in the dynamic features of the mark, resulting in blurred or distorted marks in the image. This indicates that the rotation stability of the bobbin is poor, and there may be process or equipment problems, resulting in the inability to effectively identify the dynamic features or quality problems of the textile. This usually means that there are quality problems with the textile, such as insufficient strength, appearance defects, irregular fabrics, etc.

[0112] The generated textile determination coefficient is compared with a preset textile state determination threshold value to generate a textile detection stability signal and a textile detection state abnormality signal;

[0113] After obtaining the textile determination coefficient, the textile determination coefficient is compared with the textile state determination threshold;

[0114] If the textile determination coefficient is greater than or equal to the textile state determination threshold, a textile detection stability signal is generated, indicating that the performance of the textile is in line with expectations in terms of rotation dynamic characteristics and image clarity, the dynamic characteristics of the mark during rotation are clear and stable, and the image acquisition quality is high, indicating that the detection can accurately identify the state and quality of the textile, the textile detection can operate normally during the working process, the equipment has no malfunction or instability, the collected image information is reliable, the textile production line has no obvious quality problems during the manufacturing process, and the product is in good condition without additional intervention;

[0115] If the textile determination coefficient is less than the textile state determination threshold, a textile detection state abnormality signal is generated, indicating that the equipment may have a fault or instability, such as an unstable rotation system or a problem with the image acquisition device (such as insufficient light source, inaccurate focus, vibration interference, etc.), resulting in poor image quality and inability to accurately obtain the dynamic characteristics of the textile. The generation of abnormal signals reminds production personnel or system operators to take measures to intervene and repair. Specific measures include:

[0116] Equipment inspection and adjustment: Check whether the rotating equipment, image acquisition equipment, and stroboscope have any faults or unstable operation, and make necessary repairs or adjustments:

[0117] Optimize production parameters: If it is determined to be a process problem, it is necessary to adjust production parameters such as rotation speed, tension, etc. to ensure the stability of the bobbin rotation process;

[0118] Enhance image acquisition quality: Check the settings of light source, exposure, focus, etc. to ensure that there are no interferences during image acquisition and improve image clarity.

[0119] Material and process adjustments: If the problem lies with the textile itself (such as weak twist or uneven quality), you can try to adjust the textile parameters of the raw materials and optimize the textile process to avoid similar problems from happening again.

[0120] In summary, the stable signal of textile detection indicates that the quality of the current textile meets the standards, the equipment and process are stable, the production and detection systems are in good working condition, and normal production and detection operations can continue. The abnormal signal of textile detection status indicates that there are quality problems with the textile or abnormalities in the production process, and corresponding intervention measures need to be taken, such as checking equipment, adjusting production parameters, optimizing image acquisition, improving material processes, etc., to restore the normal operation of the production line and improve product quality;

[0121] It should be noted that the threshold information in this embodiment is pre-set by professionals and will not be explained in detail here. Some parameter English letters in the embodiments have the same situation, but different meanings are explained when used, which will not be explained one by one here.

[0122] The present invention achieves efficient detection of weak twist phenomenon in the textile process by installing a marker with high contrast and periodic texture on the upper end of the yarn tube and combining it with an accurate image acquisition and processing method. By optimizing the geometric shape, material and installation angle of the marker, it is ensured that the marker has a stable and clear image during high-speed rotation, thereby improving the detection accuracy. The camera and the stroboscope work together to dynamically adjust the exposure time and sampling rate, and can stably capture images at different rotation speeds, thus overcoming the blur and noise problems that are prone to occur in traditional methods in dynamic scenes. The collected images are denoised, edge detected and analyzed for rotation features, and dynamic features can be effectively extracted. The presence of weak twist phenomenon can be accurately determined through marker position matching and stability evaluation. In addition, based on the analysis of image extraction information, the textile detection strategy is adjusted in combination with the generated signal, thereby achieving instant feedback and adjustment of targeted problems, improving the quality control efficiency in the production process, and greatly improving the detection accuracy and response speed of textile equipment.

[0123] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0124] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0125] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0126] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0127] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0128] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for dynamic detection and feature extraction of textile spindle speed, characterized in that: The steps include: Install a marker with high contrast and periodic texture on the top of the bobbin and optimize the marker geometry, material and installation angle; Set up the camera and stroboscope, dynamically adjust the exposure time and sampling rate, and collect images at different rotation speeds; De-noising, edge detection and rotation feature analysis are performed on the collected images. Dynamic features are extracted through marker position matching and stability assessment. The quality of the markers is comprehensively assessed to determine whether weak twist occurs. Obtain and analyze the image extraction information generated during the evaluation process of weak twist phenomenon, and adjust the textile detection strategy according to the different signals generated by the analysis; The collected images are subjected to denoising, edge detection and rotation feature analysis. Dynamic features are extracted through marker position matching and stability assessment. The quality of the markers is comprehensively assessed to determine whether weak twisting occurs. The specific steps are as follows: Use median filtering to denoise, replacing each pixel value with the median value within the pixel neighborhood; Extract the marked edge information in the image through the Canny edge detection algorithm; Analyze the rotation features of the markers in the image sequence and extract the dynamic features of the markers, the dynamic features of the markers include the rotation speed and the rotation angle change; Based on the temporal continuity of the image sequence, the rotation speed of the marker is calculated, and the stability of the rotation state is determined according to the rate of change of the rotation angle; The dynamic blur is calculated using the gradient information of the image edge, the local mean shift algorithm is used to evaluate the blur of each frame, and the dynamic blur value is compared with the threshold; The tracking stability of the marker is determined by the change in the marker position between consecutive frame images; The quality of the marker is comprehensively evaluated by combining the rotation speed, ambiguity, and tracking stability of the marker to obtain the final marker quality score and determine the weak twist phenomenon.

2. The method for dynamic detection and feature extraction of textile spindle speed according to claim 1, characterized in that: Install a marker with high contrast and periodic texture on the top of the bobbin and optimize the marker geometry, material and installation angle. The specific steps are as follows: Add a cap-shaped mark on the top of the bobbin, the shape of which is designed to be a triangle or a QR code, and add periodic stripes or textures on the surface of the mark, and calculate the edge sharpness, reflectivity, and coverage of the cap-shaped mark; Use a high-frame-rate camera and stroboscope combination to dynamically shoot the mark, record the mark clarity at different rotation speeds, and adjust the stroboscope flashing frequency according to the mark's dynamic blur and light reflection intensity; The overall quality of the marking is evaluated by comprehensively considering edge sharpness, reflectivity, coverage and dynamic blur, and a marking scoring model is established to determine the optimal marking parameters and layout scheme.

3. The method for dynamic detection and feature extraction of textile spindle speed according to claim 2, characterized in that: Set up the camera and stroboscope, dynamically adjust the exposure time and sampling rate, and collect images at different rotation speeds. The specific steps include: The stroboscope frequency calculation and synchronization initialization are performed, the rotation speed data of the bobbin is obtained through the sensor, the optimal flashing frequency of the stroboscope is calculated according to the bobbin rotation speed, and the flashing time of the stroboscope is dynamically adjusted to make each flashing interval consistent with the rotation of the marker; According to the stroboscope frequency, set the camera sampling rate to synchronize with the stroboscope, adjust the camera exposure time and aperture size, and detect inter-frame jitter; Perform clarity detection on the real-time collected images and evaluate the edge blurriness of the mark. If the blurriness exceeds the blurriness threshold, adjust the stroboscope flashing frequency or camera sampling rate and recalibrate the synchronization; When the image edge blur is greater than the image edge blur threshold, the stroboscope frequency is increased or the exposure time is adjusted; When the image edge blurriness is less than or equal to the image edge blurriness threshold, the current frequency and exposure settings remain unchanged.

4. The method for dynamic detection and feature extraction of textile spindle speed according to claim 3, characterized in that: And conduct a comprehensive assessment of the quality of the mark to determine whether weak twist occurs, including the following steps: The marking quality scores were compared with the threshold value for weak twist phenomenon; If the mark quality score is less than the weak twist phenomenon threshold, it is determined that the weak twist phenomenon occurs and the mark is analyzed.

5. The method for dynamic detection and feature extraction of textile spindle speed according to claim 4, characterized in that: Obtaining image extraction information generated during the evaluation of the weak twist phenomenon and analyzing it includes the following steps: Acquire image extraction information generated during the evaluation process of the weak twist phenomenon, wherein the image extraction information includes dynamic feature information and image scale information; The dynamic feature information includes a rotation dynamic feature index, and the image scale information includes an image clarity index; The obtained rotation dynamic characteristic index and image clarity index are combined to generate a textile determination coefficient; The rotation dynamic characteristic index, image clarity index and textile determination coefficient are proportional.

6. The method for dynamic detection and feature extraction of textile spindle speed according to claim 5, characterized in that: The textile detection strategy is adjusted according to the different signals generated by the analysis, including the following steps: comparing the generated prediction adjustment coefficient with the set textile state determination threshold; If the predicted adjustment coefficient is greater than or equal to the textile state determination threshold, a textile detection stability signal is generated, indicating that the textile process is normal and no additional intervention is required; If the predicted adjustment coefficient is less than the textile state determination threshold, a textile detection state abnormality signal is generated to remind the production personnel or system operators that measures need to be taken to intervene in the textile process.

Images