Textile surface defect detection method and system

Abstract

The invention relates to the technical field of textile detection, in particular to a textile surface defect detection method and a textile surface defect detection system, which are used for carrying out online detection on a textile of which the surface is coated with a thermosensitive coating, and the method comprises the following steps: applying preset periodic intensity modulation to a detection light source used for online detection, generating a reference sequence serving as a synchronism analysis benchmark; collecting a series of brightness values of the suspected defect points within a preset time period under the periodic intensity modulation aiming at the suspected defect points, which are preliminarily identified, on the surface of the textile, so as to form a brightness response sequence representing the optical response characteristics of the suspected defect points; calculating a time synchronization degree between the brightness response sequence and the reference sequence; by distinguishing the temporary thermal effect artifacts and the real physical defects, the method has the advantages that the temporary optical artifacts and the real physical defects caused by the thermal effect can be effectively distinguished, the false alarm rate is reduced, and the detection accuracy is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of textile detection, and in particular to a method and system for detecting textile surface defects. Background Art

[0002] In the modern textile industry, especially in areas with extremely stringent product quality requirements, the use of automated optical inspection systems for online detection of defects on textile surfaces has become a key link in ensuring product qualification and production efficiency. Such systems typically rely on high-intensity lighting and high-resolution cameras to capture subtle images of the surface of fabrics moving at high speed on the production line, and use image analysis to identify defects such as broken yarns, stains, and holes. However, with the development of functional textiles, the special physical properties of some new materials have brought new challenges to traditional optical inspection methods. In particular, for textiles with functional coatings on their surfaces, the coating materials may be sensitive to physical quantities in the inspection environment (such as temperature), thereby generating interference signals during the inspection process that are difficult to distinguish from real defects, resulting in a large number of misjudgments.

[0003] In order to identify possible minor structural defects in the fabric substrate that may affect product performance, such as broken yarns, sparse or dense threads, as well as defects in the coating itself, multiple sets of high-brightness linear LED light sources are installed in the inspection darkroom. These light sources continuously and intensely illuminate the fabric surface from specific angles, ensuring that the line scan camera can capture sufficiently clear surface images with a high signal-to-noise ratio. However, as continuous production time increased, the system began to frequently issue alarms, reporting that a large number of tiny, irregularly shaped "dark spots" or "bright spots" defects were found on the fabric surface. When the operator paused the production line for manual re-inspection, he was unable to detect any physical abnormalities in the static state, and these so-called defects disappeared without a trace.

[0004] The root of the problem lies in the long-term, continuous exposure to high-intensity LED light, which causes the fabric surface to absorb significant heat. Due to subtle variations in density and heat capacity at the microscopic level within the fabric substrate (such as at yarn interlacing points), numerous "micro-hotspots" (micro-hotspots) form on the fabric surface, where temperatures are slightly higher than those of the surrounding areas. The functional coating on this fabric surface is temperature-sensitive, causing the optical reflectivity of the coating above these micro-hotspots to temporarily and slightly change, resulting in blurred "dark spots" or "bright spots" in the line scan camera image. The size, shape, and contrast of these temporary thermal artifacts are virtually indistinguishable from real, permanent defects caused by coating process instabilities, such as a small, partially spread coating droplet or a tiny coating miss caused by an air bubble. This makes it impossible for the inspection system to distinguish, from a single static image feature, whether the underlying physical cause is a temporary thermal effect or a permanent material defect, resulting in a high false alarm rate and a high risk of missed detections.

[0005] In view of the above problems, the existing technology is in urgent need of improvement. Summary of the Invention

[0006] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a method and system for detecting surface defects of textiles.

[0007] In a first aspect, the present invention provides a method for detecting surface defects of textiles, which is used for online detection of textiles with a heat-sensitive coating on the surface, and the method comprises the following steps: Applying a preset periodic intensity modulation to the detection light source used for online detection to generate a reference sequence as a benchmark for synchronization analysis; For the suspected defect point initially identified on the textile surface, under the periodic intensity modulation, a series of brightness values ​​of the suspected defect point within a preset time period are collected to form a brightness response sequence representing its optical response characteristics; calculating a degree of time synchronization between the brightness response sequence and the reference sequence; Based on the time synchronization degree, it is determined whether the suspected defect point is a real physical defect or a temporary optical artifact caused by thermal effect.

[0008] The core innovation of this application lies in applying a preset periodic intensity modulation to the detection light source and analyzing the degree of temporal synchronization between the brightness response sequence of the suspected defect point and the modulation synchronization analysis benchmark reference sequence, thereby utilizing the dynamic optical response difference between real physical defects and temporary optical artifacts caused by thermal effects under periodic illumination, solving the problem in the existing technology of difficulty in distinguishing between these two types of defects, and achieving filtering of false alarms caused by thermal artifacts.

[0009] In a second aspect, a textile surface defect detection system is provided for online detection of textiles with a heat-sensitive coating on the surface, the system comprising: A modulation module, configured to apply a preset periodic intensity modulation to the detection light source used for online detection, so as to generate a reference sequence as a benchmark for synchronization analysis; an acquisition module for acquiring, for a suspected defect point initially identified on the textile surface, a series of brightness values ​​of the suspected defect point within a preset time period under the periodic intensity modulation, so as to form a brightness response sequence representing its optical response characteristics; a calculation module, configured to calculate a degree of time synchronization between the brightness response sequence and the reference sequence; A determination module is configured to determine, based on the degree of time synchronization, whether the suspected defect point is a real physical defect or a temporary optical artifact caused by a thermal effect.

[0010] Compared with the prior art, the present invention has the following beneficial effects: By applying periodic intensity modulation to the detection light source and analyzing the time synchronization between the brightness response of the suspected defect point and the modulation signal, and combining frequency analysis based on geometric features when necessary, it is possible to distinguish temporary thermal effect artifacts from real physical defects. This has the advantages of being able to effectively distinguish temporary optical artifacts caused by thermal effects from real physical defects, reduce false alarm rates, and improve detection accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 Flow chart of the method of the present invention.

[0012] Figure 2 Schematic diagram of the system structure of the present invention.

[0013] In the figure: 201, modulation module; 202, acquisition module; 203, calculation module; 204, determination module. DETAILED DESCRIPTION

[0014] The embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and are not to be construed as limiting the present invention.

[0015] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of the technical features being referred to. Thus, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the specified features. In the description of the present invention, "plurality" means two or more, unless otherwise specifically defined.

[0016] Traditional automated optical inspection systems, when performing online inspection of textiles coated with heat-sensitive coatings, struggle to distinguish between real physical defects and temporary optical artifacts caused by the thermal effects of the inspection light source. High-intensity illumination causes localized temperature rises in the fabric microstructure, which in turn temporarily alters the optical properties of the heat-sensitive coating, creating artifacts in the line camera image that resemble the visual characteristics of real defects. This similarity makes it difficult to accurately determine the physical cause of the signal using only static image features, resulting in a high rate of false positives in the inspection system.

[0017] For example, consider a production line inspecting a composite textile with a heat-sensitive coating. To ensure clear images of the fabric in high-speed motion, the system continuously illuminates it with a high-intensity LED light source. This illumination causes a localized temperature rise in the fabric's surface microstructure. The heat-sensitive coating, being sensitive to temperature, temporarily alters its optical properties, creating temporary optical artifacts in the image that mimic the visual characteristics of real defects. These artifacts are difficult to distinguish from permanent defects in static images, leading to frequent false alarms from the inspection system.

[0018] If these issues are not addressed, the inspection system will generate numerous false alarms, requiring frequent manual rechecks and reducing production efficiency. High false alarm rates can lead to operator fatigue and increase the risk of missing real defects. The inability to reliably distinguish between temporary artifacts and permanent defects will impair the system's quality control capabilities and lead to unreliable inspection results.

[0019] To this end, this application Figure 1 A textile surface defect detection method is shown, which is used for online detection of textiles with a heat-sensitive coating on the surface. The method includes the following steps: S101, applying a preset periodic intensity modulation to a detection light source used for online detection to generate a reference sequence serving as a benchmark for synchronization analysis; S102. For a suspected defect point initially identified on the textile surface, collect a series of brightness values ​​of the suspected defect point within a preset time period under periodic intensity modulation to form a brightness response sequence representing its optical response characteristics; S103, calculating the degree of time synchronization between the brightness response sequence and the reference sequence; S104: Based on the degree of time synchronization, determine whether the suspected defect point is a real physical defect or a temporary optical artifact caused by thermal effects.

[0020] The application of a preset periodic intensity modulation to the detection light source refers to the variation of the luminous intensity of the light source used to illuminate the textile surface according to a predetermined time period. For example, the intensity variation can be in the form of a sine wave, square wave, pulse sequence, etc. This can be achieved by controlling the light source driving current or voltage, such as controlling the brightness of the LED light source through pulse width modulation (PWM). The main purpose is to make the optical response changes on the textile surface caused by the thermal effect of the light source exhibit characteristics related to the periodic modulation, thereby providing time dimension information for the subsequent distinction between real defects and thermal artifacts. The generation of a reference sequence as a benchmark for synchronization analysis refers to the generation of a sequence that is synchronized with the preset periodic intensity modulation law applied to the detection light source in time. The sequence can be ideal modulation waveform data or actual monitored light source intensity change data. The control signal output by the signal generator can be used as a reference. It is mainly used as a standard to measure whether the brightness response sequence of the suspected defect point is synchronized with the light source modulation. Among them, collecting a series of brightness values ​​of suspected defect points within a preset time period refers to using an image sensor or photodetector to continuously record the light signal intensity of the suspected defect point position that has been initially identified over a period of time while detecting the periodic intensity modulation of the light source, and converting it into a digital brightness value sequence. This can be achieved by using a high-speed linear array camera or point sensor to continuously sample specific pixels or areas. Its main purpose is to obtain dynamic optical response data of the suspected defect point under periodic illumination changes. Among them, constituting a brightness response sequence that characterizes its optical response characteristics refers to arranging a series of collected brightness values ​​in chronological order to form a sequence that can reflect the brightness change of the suspected defect point over time within a preset time period. The morphology, amplitude, phase and other characteristics of this sequence are related to the response characteristics of the suspected defect point to periodic illumination modulation, and it is mainly used as input data for subsequent time synchronization degree calculation. Calculating the degree of time synchronization between the brightness response sequence and the reference sequence refers to using signal processing or statistical methods to quantify the correlation or consistency between the brightness response sequence and the reference sequence in the time dimension. For example, indicators such as the mutual correlation coefficient, phase difference, and frequency component matching degree can be calculated. This can be achieved by using a digital signal processor (DSP) or a general-purpose processor to execute relevant algorithms. Its main purpose is to determine whether the brightness change of the suspected defect point is synchronized with the periodic modulation of the detection light source.Among them, based on the degree of time synchronization, judging whether the suspected defect point is a real physical defect or a temporary optical artifact caused by thermal effects means comparing the calculated time synchronization value with a preset discrimination threshold to draw a conclusion about the nature of the suspected defect point. If the synchronization degree is high, it tends to be judged as a thermal artifact; if the synchronization degree is low, it tends to be judged as a real physical defect. It is mainly to utilize the dynamic response difference of the suspected defect point to periodic light modulation to achieve the distinction and filtering of real defects and thermal artifacts.

[0021] The solution of this application overcomes the limitations of relying solely on static image features for detection by incorporating temporal analysis. Specifically, a preset periodic intensity modulation is applied to the detection light source used for online detection, which is equivalent to applying an external stimulus with a specific temporal pattern to the textile surface. Simultaneously, a reference sequence is generated based on this modulation pattern, representing the ideal timing of the stimulus signal. Next, for the initially identified suspected defect point, a series of brightness values ​​are continuously acquired over a period of time under the periodic modulation of the light source, forming a brightness response sequence that represents the dynamic optical response of the point. This brightness response sequence contains information about the actual response of the suspected defect point to the periodic illumination stimulus. Because textiles coated with thermosensitive coatings are sensitive to the thermal effects of illumination, the brightness changes of temporary optical artifacts caused by this thermal effect tend to follow the periodic modulation of the light source, exhibiting a high degree of temporal synchronization with the reference sequence. However, real physical defects are structural and their optical properties are relatively stable. Even under periodic illumination, their brightness changes (if any) are generally not highly synchronized with the light source modulation. Therefore, by calculating the degree of temporal synchronization between the brightness response sequence and the reference sequence, the correlation between the brightness variation of the suspected defect and the light source modulation can be quantified. Finally, based on the calculated temporal synchronization, a discrimination criterion, such as a synchronization threshold, is set. Synchronization above the threshold is identified as a temporary optical artifact caused by thermal effects, while synchronization below the threshold is identified as a true physical defect. This entire process forms a complete dynamic response analysis chain, enabling the system to identify the nature of the defect based on the differences in the dynamic behavior of the suspected defect in response to periodic illumination.

[0022] In one specific embodiment, the detection light source used for online inspection can be an LED array. A preset periodic intensity modulation is applied to the LED array, for example, by controlling the LED drive current using pulse width modulation (PWM) technology, causing its brightness to vary periodically at a fixed frequency (e.g., 100 Hz) and waveform (e.g., a square wave or sine wave). Synchronized with the PWM control signal, a digital reference sequence is generated that reflects the expected variation in light source intensity. As the textile passes through the inspection area, a line scan camera continuously images the fabric surface. When the image processing algorithm initially identifies a suspected defect, the system locks onto the corresponding pixel location in the consecutive image frames. Over a preset time period (e.g., 100 milliseconds), a series of brightness values ​​at that pixel location are extracted from these consecutive image frames to form a brightness response sequence for the suspected defect. This brightness response sequence reflects the temporal variation in brightness of the defect under the periodic light source modulation. To calculate the degree of temporal synchronization, a cross-correlation analysis can be performed between the brightness response sequence and the reference sequence to calculate the cross-correlation function. The peak value and position of the cross-correlation function represent the degree of temporal synchronization between the two sequences. If the cross-correlation function exhibits a peak at zero or a fixed time lag, it indicates a high degree of temporal synchronization between the brightness response sequence and the reference sequence. The cross-correlation peak is compared with a preset threshold. If the peak exceeds the threshold, the suspected defect is determined to be a temporary optical artifact caused by thermal effects; if the peak is below the threshold, it is determined to be a real physical defect.

[0023] As an embodiment of the present invention, when the time synchronization degree is lower than a preset synchronization threshold, the method further includes: Obtaining a geometric characteristic frequency determined by the physical dimensions of the three-dimensional periodic structure of the textile and the running speed of the textile; Analyze the frequency composition of the brightness response sequence to determine its dominant frequency; Based on the matching relationship between the dominant frequency and the geometric characteristic frequency, the suspected defect point is again determined to be a real physical defect or a temporary optical artifact.

[0024] Among them, the geometric characteristic frequency refers to the natural frequency of the textile periodic structure in the image sequence, which is determined by the physical size of the three-dimensional periodic structure of the textile and the running speed of the textile. It reflects the periodic changes in the textile's own structure in motion. The frequency composition of the brightness response sequence refers to the energy distribution or intensity distribution of the sequence at different frequencies obtained by frequency analysis of the brightness response sequence. The dominant frequency refers to the frequency component with the most prominent energy or intensity in the frequency composition of the brightness response sequence, which represents the most significant periodic pattern in the brightness change. The matching relationship refers to whether there is a preset correlation or consistency between the dominant frequency and the geometric characteristic frequency, for example, whether the frequency values ​​of the two are close to or equal within the allowable error range.

[0025] The solution of the present application introduces a secondary judgment mechanism based on frequency analysis when the degree of time synchronization is lower than a preset synchronization threshold. First, the geometric characteristic frequency determined by the physical size of the three-dimensional periodic structure of the textile and the running speed of the textile is obtained. This frequency represents the natural frequency response generated by the periodic structure of the textile itself during movement. Next, the frequency composition of the brightness response sequence of the suspected defect point is analyzed to determine its dominant frequency, which reflects the main periodic component of the brightness change of the suspected defect point. Finally, based on the matching relationship between the dominant frequency and the geometric characteristic frequency, the type of the suspected defect point is determined again. If the dominant frequency matches the geometric characteristic frequency, it indicates that the brightness change period of the suspected defect point is consistent with the period of the textile structure, and it is very likely to be a temporary optical artifact caused by the periodic structure of the textile. On the contrary, if the two do not match, it indicates that the brightness change of the suspected defect point has nothing to do with the textile structure, and it is more likely to be a real physical defect. This secondary judgment based on frequency characteristics, combined with the preliminary judgment based on time synchronization, forms a multi-dimensional defect judgment method, which effectively makes up for the shortcomings of a single judgment dimension. Especially when time synchronization is not obvious, it can more accurately distinguish between artifacts caused by textile structure and real physical defects, thereby improving the accuracy of defect detection.

[0026] As an embodiment of the present invention, the steps of analyzing the frequency composition of the brightness response sequence to determine its dominant frequency, and re-determining whether the suspected defect point is a real physical defect or a temporary optical artifact based on the matching relationship between the dominant frequency and the geometric characteristic frequency include: Obtaining the frequency composition of the brightness response sequence; A frequency interval is defined with the geometric characteristic frequency as the center; Based on the frequency structure of the brightness response sequence, the energy accumulation value in the frequency interval is calculated; Based on the prominence of the energy accumulation value, the suspected defect point is determined to be a real physical defect.

[0027] Obtaining the frequency composition of the brightness response sequence refers to converting the brightness response sequence from the time domain to the frequency domain through a frequency transformation, such as a Fourier transform, to obtain its energy or amplitude distribution at different frequencies. This serves as the basis for subsequent frequency analysis, aiming to transform the characteristics of the brightness response sequence from the time dimension to the frequency dimension for observation and analysis. Defining a frequency interval centered on the geometric characteristic frequency refers to defining a specific frequency range around the geometric characteristic frequency, determined by the physical dimensions of the textile's three-dimensional periodic structure and the textile's operating speed. This frequency interval limits the frequency range of interest, as real physical defects often produce a more significant frequency response near the geometric characteristic frequency. The purpose is to focus the analysis on this critical region, eliminate interference from other frequency components, and improve analysis accuracy. Calculating the energy accumulation within the frequency interval based on the frequency composition of the brightness response sequence refers to summing or integrating the energy of the brightness response sequence in the frequency domain within the defined frequency interval. The energy accumulation value reflects the degree of energy concentration of the brightness response sequence near the geometric characteristic frequency, aiming to quantify the response intensity of suspected defects within this critical frequency range. Judging a suspected defect point as a real physical defect based on the prominence of the energy accumulation value means comparing the calculated energy accumulation value with a certain benchmark or threshold to evaluate its significance. If the energy accumulation value is significantly higher than the background or preset threshold, the suspected defect point is considered to be more likely to be a real physical defect. Its purpose is to make a judgment on the physical properties of the suspected defect point based on the quantified energy concentration.

[0028] The solution of this application further refines the frequency analysis method based on the initial distinction between thermal artifacts and real defects based on time synchronization, and a secondary judgment based on the matching relationship between the dominant frequency and the geometric characteristic frequency when time synchronization is low. The solution recognizes that simple dominant frequency matching may not be sufficient for all situations, especially when structural noise produces energy concentration near the geometric characteristic frequency. Therefore, the solution obtains the frequency composition of the brightness response sequence and defines a frequency interval centered on the geometric characteristic frequency, accurately locking the focus of analysis on the critical frequency range related to the periodic structure of the textile. Then, by calculating the energy accumulation value within this frequency interval, the solution can more comprehensively evaluate the total energy intensity of the suspected defect point within this critical frequency range, rather than focusing solely on a single dominant frequency. Finally, judging based on the prominence of the energy accumulation value can effectively distinguish artifacts or noise whose dominant frequency is close to the geometric characteristic frequency but whose energy concentration is not high, because real physical defects generally exhibit stronger and more concentrated energy responses near the geometric characteristic frequency. This approach, combined with the aforementioned time synchronization analysis and dominant frequency matching, forms a multi-stage, multi-dimensional judgment process, which significantly enhances the system's ability to distinguish between real defects and temporary optical artifacts, thereby improving the accuracy and reliability of defect detection.

[0029] As an embodiment of the present invention, the step of determining whether a suspected defect point is a real physical defect based on the prominence of the energy accumulation value includes: Obtaining the frequency composition of the suspected defect point in one or more adjacent areas on the textile; Determining a local background energy baseline within the frequency interval based on the frequency composition of one or more adjacent regions; The energy accumulation value is compared with the local background energy benchmark, and based on the comparison result, the suspected defect point is determined to be a real physical defect.

[0030] Among them, one or more neighboring areas of a suspected defect point on a textile refer to one or more areas selected around the suspected defect point for analyzing its background texture characteristics, which can be achieved by selecting a rectangular area of ​​fixed size, a circular area, or other shape areas selected according to preset rules; frequency composition refers to the spectrum information obtained by frequency analysis of the image area, which characterizes the energy distribution of the area at different spatial frequencies, which can be achieved by methods such as two-dimensional Fourier transform; frequency interval refers to a frequency range defined with the geometric characteristic frequency as the center; local background energy benchmark refers to the typical energy level of the textile background texture in the neighboring area of ​​the suspected defect point within the frequency interval, which can be achieved by calculating the average energy value, median of the neighboring area within the frequency interval, or a value determined by other statistical methods; energy accumulation value refers to the sum of the energy of the suspected defect point itself within the frequency interval or a representative value; comparison result refers to the relationship between the energy accumulation value and the local background energy benchmark, which can be judged by whether the energy accumulation value exceeds the benchmark plus a threshold.

[0031] The solution of this application obtains the frequency composition of the area adjacent to a suspected defect point to understand the frequency characteristics of the textile background surrounding the suspected defect point. As a periodic structure, the texture of a textile generates energy concentrations at specific frequencies, which constitute background noise. By analyzing the frequency composition of the adjacent areas, the intensity and distribution of this background noise can be estimated. Based on the frequency composition of one or more adjacent areas, a local background energy baseline is determined within the frequency interval. This local background energy baseline represents the typical energy level of the textile background within a specific frequency range. The energy accumulation value is compared with the local background energy baseline, and based on the comparison result, the suspected defect point is determined to be a real physical defect. If the energy accumulation value of the suspected defect point is significantly higher than the local background energy baseline, the energy accumulation can be assumed to be caused by the defect rather than the textile's inherent texture. Conversely, if the energy accumulation value is close to the local background energy baseline, the energy accumulation can be assumed to be caused by the textile's inherent texture, thus eliminating false positives. This approach, which incorporates a local background energy baseline and performs comparison, improves upon the method of determining the suspected defect point based solely on its own energy accumulation value. It takes into account the fluctuation of the local background texture, making the determination process more adaptable to complex textile texture backgrounds. Combined with the preliminary judgment based on time synchronization and geometric characteristic frequency matching in the previous scheme, this scheme further refines the judgment logic in the frequency domain. By introducing local background reference, it effectively improves the ability to distinguish real defects and schlieren noise in complex backgrounds, forming a more robust and accurate defect recognition system.

[0032] As an embodiment of the present invention, the step of obtaining the frequency composition of one or more adjacent areas of a suspected defect point on a textile includes: Determine the directional characteristics of the textile background surrounding the suspected defect; selecting one or more adjacent regions on the textile based on the directional characteristics; The frequency corresponding to the selected one or more adjacent regions is obtained.

[0033] Directional features refer to the spatial distribution of the main directions or angles of a textile's texture. Image processing techniques, such as Fourier transforms or Gabor filter bank analysis, can be used to extract texture directional information from local image regions. A neighboring region refers to one or more image regions that are spatially adjacent or close to a suspected defect point. The shape and size of the region can be customized based on the application and texture characteristics. Frequency composition refers to the energy distribution of an image region in the frequency domain. This can be obtained by performing a frequency transform on the image region.

[0034] The solution of the present application determines the directional characteristics of the textile background around the suspected defect point and selects the adjacent area based on the directional characteristics to obtain a frequency composition that is more representative of the actual local background texture. This is because the texture of the textile often has a directionality, such as the direction of the longitude and latitude lines. If the selected adjacent area does not match the directional characteristics, its frequency composition may not accurately reflect the background texture around the suspected defect point. By considering the directional characteristics to select the adjacent area, it can be ensured that the texture of the selected area is more consistent with the background texture around the suspected defect point. Based on the frequency composition of these more representative adjacent areas, a local background energy benchmark within the frequency range can be more accurately determined. Comparing the energy accumulation value of the suspected defect point with this more accurate local background energy benchmark can more effectively eliminate the interference of the textile's own texture on defect judgment. It is precisely because the adjacent area with the same directionality as the background texture is selected that the obtained background energy benchmark is more reliable, which enables the subsequent comparison and judgment process to more accurately distinguish between real defects and artifacts caused by texture, thereby improving the accuracy of defect detection.

[0035] As an embodiment of the present invention, the step of determining the directional characteristics of the textile background around the suspected defect point includes: Perform a two-dimensional frequency transform on the background image around the suspected defect point to obtain the corresponding spectrum; Directional characteristics are determined based on the distribution orientation of energy in the spectrum.

[0036] Among them, two-dimensional frequency transform refers to a mathematical method for converting an image from the spatial domain to the frequency domain, which can be achieved by two-dimensional Fourier transform, two-dimensional discrete cosine transform, etc. Its purpose is to reveal the components of the image at different frequencies and directions; spectrum refers to the frequency domain representation obtained after two-dimensional frequency transform, which is usually a complex matrix. Its amplitude spectrum reflects the intensity of different frequency components in the image, and the phase spectrum reflects the position information of these components. Its purpose is to visualize the frequency and direction information of the image; the distribution orientation of energy in the spectrum refers to the main directionality of the region with larger amplitude (energy) in the spectrum on the frequency domain plane. For images with periodic textures, their energy will be concentrated in a specific direction perpendicular to the texture direction. Its purpose is to infer the texture direction of the original image by analyzing the direction of energy concentration; directional characteristics refer to the main direction or arrangement direction of the textile background texture in space, such as the direction of warp or weft yarn. Its purpose is to characterize the structural characteristics of the textile background and provide a basis for the subsequent selection of adjacent areas.

[0037] The solution of this application performs a two-dimensional frequency transform on the background image surrounding the suspected defect, converting the image from the spatial domain to the frequency domain and obtaining the corresponding spectrum. This process converts complex spatial texture information into easily analyzable frequency and directional information. Subsequently, directional features are determined based on the energy distribution orientation in the spectrum. Because textiles typically have a periodic texture structure, their spectral energy is concentrated in a specific direction perpendicular to the texture direction. By analyzing the direction of energy concentration in the spectrum, the texture direction of the original image can be directly and accurately extracted. For example, if the textile texture is horizontal, its spectral energy will be concentrated in the vertical direction. Compared to directly analyzing the image texture in the spatial domain, this frequency-domain-based analysis method effectively filters out the effects of noise and illumination variations, and more robustly extracts directional information from periodic textures. After determining the directional features, one or more adjacent regions on the textile are selected based on these directional features, and the frequency composition corresponding to these selected adjacent regions is then determined. Accurate directional features enable the subsequently selected adjacent regions to better represent the background texture in that direction, thereby more accurately obtaining the background frequency composition and, in turn, more accurately determining the local background energy baseline, ultimately improving the accuracy of defect detection. It is precisely because of the use of frequency domain analysis to determine the directional characteristics that the subsequent adjacent area selection and background energy benchmark determination are more reliable, thereby effectively solving the problem of adjacent area selection deviation caused by inaccurate directional feature judgment.

[0038] As an embodiment of the present invention, the step of determining the directional characteristics based on the distribution orientation of energy in the spectrum includes: Integrate the energy of the spectrum in the angular direction; According to the energy integration result, determine the angle direction with the maximum energy; The angular direction with the maximum energy is taken as the directional feature.

[0039] The spectrum refers to two-dimensional data obtained by performing a two-dimensional frequency transform on an image, representing the energy distribution of the image at different frequencies and directions. This can be implemented using algorithms such as the Fast Fourier Transform (FFT). The angular energy integration of the spectrum refers to the accumulation of spectral energy along each angular direction from the center of the spectrum (zero frequency point) in the spectral space. This can be achieved by dividing the spectrum into multiple angular sectors and calculating the sum of the energy within each sector. The energy integration result refers to the distribution data obtained after the angular energy integration, representing the accumulated energy values ​​in different angular directions. This can be represented by a one-dimensional array or curve, where each element or point corresponds to an angular direction and its corresponding energy integration value. The angular direction with the highest energy refers to the angle corresponding to the maximum energy value in the energy integration result. This can be determined by searching for the maximum value in the energy integration result and its index (corresponding angle). The directional feature refers to a parameter used to describe the main orientation of the image texture or structure. It can be represented by an angle value or a directional vector.

[0040] The solution of the present application integrates the energy in the acquired spectrum in the angular direction, accumulating the energy information scattered in the two-dimensional spectrum along different angular directions, thereby obtaining a one-dimensional curve reflecting the distribution of energy with angle. This integration process effectively smooths the local fluctuations and noise in the spectrum, making the distribution trend of energy in different directions clearer. Then, by analyzing the energy integration result, the angular direction where the energy value reaches the maximum is directly determined. This angular direction with the maximum energy represents the dominant direction with the most concentrated energy in the spectrum, corresponding to the main direction of the texture or structure in the image. Finally, this angular direction with the maximum energy is directly output as the directional feature of the image. The entire process transforms complex two-dimensional spectrum analysis into simple one-dimensional curve analysis, providing a robust and efficient way to accurately extract the directional information of the image. The accurately extracted directional feature can guide the subsequent steps to more accurately select adjacent areas with similar background textures to the suspected defect point, thereby more accurately estimating the local background energy baseline, thereby improving the accuracy of the final defect judgment and effectively distinguishing between real physical defects and temporary optical artifacts.

[0041] As an embodiment of the present invention, the step of determining a local background energy reference within a frequency interval includes: sorting the energy values ​​of one or more adjacent regions within a frequency interval; According to the sorting results, a specific percentile value in the sorted results is selected as the local background energy benchmark.

[0042] Among them, the specific quantile value refers to the value at a specific position after a set of data is sorted from small to large, such as the percentile. It can be achieved by calculating the value at a certain percentage position in the sorted data set. Its purpose is to represent a certain position in the data distribution and to effectively exclude the influence of extreme values; the local background energy benchmark refers to a reference value used to measure the energy level of the background area around the suspected defect point within a specific frequency range. It can be determined by statistical analysis of the energy values ​​of the adjacent areas (such as calculating the mean, median, quantile value, etc.). Its purpose is to provide a reliable background reference for determining whether the suspected defect point is a real physical defect.

[0043] The solution of this application sorts the energy values ​​of one or more adjacent regions within a frequency interval, revealing the overall distribution of these energy values ​​rather than simply taking an average. After sorting, the energy values ​​can be clearly seen from low to high, providing a better understanding of background energy concentration areas and outliers. Based on this, a specific quantile within the sorted results is selected as the local background energy baseline. The selection of quantiles effectively eliminates the influence of outliers. For example, if the average is taken directly, a few extremely high energy values ​​may significantly increase the background energy baseline, causing some real defects to be misclassified as background. Selecting a lower quantile, however, allows for a more robust estimation of the background energy level, reducing the likelihood of misclassification. Furthermore, selecting an appropriate quantile can reflect the concentration of background energy. For example, if the background energy distribution is relatively uniform, the differences between different quantiles may be small. However, if the background energy is concentrated in a few areas, a lower quantile may better represent the true background level. Sorting the energy values ​​of adjacent regions and selecting an appropriate quantile makes the estimation of the local background energy baseline more accurate, thereby improving the accuracy and robustness of defect detection. This more accurate benchmark determination method, combined with the framework for making decisions based on local background energy benchmarks, makes the entire decision process more robust and reliable.

[0044] As an embodiment of the present invention, the step of selecting a specific quantile value in the sorting result as the local background energy benchmark includes: Analyze the distribution trend of energy values ​​in the sorting results; According to the distribution trend, a percentile value that can represent the background energy concentration area is determined; The quantile value is used as the local background energy benchmark.

[0045] Analyzing the distribution trend of energy values ​​in the sorting results refers to examining the overall characteristics of the sorted energy value sequence through statistical or data analysis methods, such as the degree of concentration, dispersion, and presence of multiple peaks of the energy values. This can be achieved by calculating a histogram of the energy values, density estimation, or calculating statistical moments (such as mean, standard deviation, skewness, and kurtosis). Determining a quantile value that can represent the background energy concentration area refers to adaptively selecting a quantile based on the analyzed energy value distribution trend. The energy value corresponding to this quantile can effectively represent the main concentration level of the background energy. The selection of the quantile value can be dynamically adjusted based on the peak position of the distribution, the width of the distribution, or preset rules (for example, if the distribution is biased towards low energy, a lower quantile value is selected; if the distribution is wide, a quantile value that can cover the main background energy range is selected).

[0046] The solution of the present application first analyzes the distribution trend of the results of the energy value sorting of the adjacent areas, which enables the system to understand the actual distribution form of the background energy, such as whether it is concentrated or dispersed, and whether there are abnormally high or low values. Based on the analysis of the distribution trend, the system can adaptively determine a more representative percentile value. This percentile value is no longer a fixed percentage, but is dynamically adjusted according to the actual data distribution, so as to more accurately capture the concentration level of the background energy. This adaptively determined percentile value is used as the local background energy benchmark for comparison with the energy accumulation value of the suspected defect point. This method overcomes the benchmark deviation problem that may be caused by fixed percentile values ​​when the background energy distribution is uneven or there is interference. In the defect detection framework based on time and frequency analysis, especially based on the method of selecting percentile values ​​as the background benchmark through energy sorting of adjacent areas, this solution optimizes the percentile selection process to make the estimation of the background energy benchmark more accurate. This more precise benchmark enables more effective differentiation of the prominent energy accumulation values ​​of real physical defects from the energy of background noise or artifacts during subsequent comparative judgments, thereby improving the accuracy and robustness of the entire detection method and effectively resolving the problem of misjudgment caused by temporary optical artifacts caused by thermal effects.

[0047] like Figure 2 A textile surface defect detection system is shown, which is used for online detection of textiles with a heat-sensitive coating on the surface. The system includes: The modulation module 201 is used to apply a preset periodic intensity modulation to the detection light source used for online detection to generate a reference sequence as a basis for synchronization analysis; The acquisition module 202 is configured to acquire a series of brightness values ​​of a suspected defect point initially identified on the textile surface within a preset time period under periodic intensity modulation to form a brightness response sequence representing its optical response characteristics; A calculation module 203 is used to calculate the degree of time synchronization between the brightness response sequence and the reference sequence; The determination module 204 is configured to determine, based on the degree of time synchronization, whether the suspected defect point is a real physical defect or a temporary optical artifact caused by thermal effects.

[0048] The solution of the present application realizes the systematic detection of surface defects of heat-sensitive coated textiles by decomposing the detection process into four functional modules: modulation, acquisition, calculation and judgment. The modulation module 201 applies periodic intensity changes to the detection light source to provide a time reference for subsequent analysis. The acquisition module 202 synchronously acquires the brightness change data of suspected defect points under light source modulation. The calculation module 203 analyzes the time synchronization relationship between the collected brightness change data and the light source modulation reference. The judgment module 204 distinguishes between temporary optical artifacts caused by the thermal effect of the light source and real physical defects based on the degree of synchronization. This modular collaborative working mode enables the system to identify defects from dynamic response characteristics rather than static image features, effectively responding to the detection challenges brought by heat-sensitive coatings and improving the accuracy and robustness of detection.

[0049] The basic principles, main features, and advantages of the present invention are shown and described above. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. Various changes and improvements are possible without departing from the spirit and scope of the present invention, and such changes and improvements fall within the scope of the invention as claimed.

Claims

1. A method for detecting surface defects of textiles, for online detection of textiles with a heat-sensitive coating on the surface, characterized in that: The method comprises the following steps: Applying a preset periodic intensity modulation to the detection light source used for online detection to generate a reference sequence as a benchmark for synchronization analysis; For a suspected defect point initially identified on the textile surface, collecting a series of brightness values ​​of the suspected defect point within a preset time period under the periodic intensity modulation to form a brightness response sequence representing its optical response characteristics; calculating a degree of time synchronization between the brightness response sequence and the reference sequence; Based on the time synchronization degree, it is determined whether the suspected defect point is a real physical defect or a temporary optical artifact caused by thermal effect.

2. A textile surface defect detection method according to claim 1, characterized in that: When the time synchronization degree is lower than a preset synchronization threshold, the method further includes: obtaining a geometric characteristic frequency determined by the physical dimensions of the three-dimensional periodic structure of the textile and the running speed of the textile; analyzing the frequency composition of the brightness response sequence to determine its dominant frequency; Based on the matching relationship between the dominant frequency and the geometric characteristic frequency, it is again determined whether the suspected defect point is a real physical defect or a temporary optical artifact.

3. A textile surface defect detection method according to claim 2, characterized in that: The step of analyzing the frequency composition of the brightness response sequence to determine its dominant frequency, and re-determining whether the suspected defect point is a real physical defect or a temporary optical artifact based on a matching relationship between the dominant frequency and the geometric characteristic frequency includes: Acquiring the frequency composition of the brightness response sequence; Defining a frequency interval with the geometric characteristic frequency as the center; Calculating an energy accumulation value within the frequency interval based on the frequency composition of the brightness response sequence; According to the prominence of the energy accumulation value, the suspected defect point is determined to be a real physical defect.

4. A textile surface defect detection method according to claim 3, characterized in that: The step of determining the suspected defect point as a real physical defect based on the prominence of the energy accumulation value includes: Obtaining a frequency composition of one or more adjacent regions of the suspected defect point on the textile; Determining a local background energy benchmark within the frequency interval based on the frequency composition of the one or more neighboring regions; The energy accumulation value is compared with the local background energy benchmark, and based on the comparison result, the suspected defect point is determined to be a real physical defect.

5. A textile surface defect detection method according to claim 4, characterized in that: The step of obtaining the frequency composition of one or more adjacent areas of the suspected defect point on the textile comprises: Determining the directional characteristics of the textile background surrounding the suspected defect point; selecting the one or more adjacent regions on the textile according to the directional characteristics; The frequency corresponding to the selected one or more adjacent regions is obtained.

6. A textile surface defect detection method according to claim 5, characterized in that: The step of determining the directional characteristics of the textile background around the suspected defect point includes: Performing a two-dimensional frequency transformation on the background image around the suspected defect point to obtain a corresponding frequency spectrum; The directional characteristics are determined based on the distribution orientation of energy in the spectrum.

7. A textile surface defect detection method according to claim 6, characterized in that: The step of determining the directional characteristics based on the distribution orientation of energy in the spectrum comprises: performing an energy integration of the spectrum in an angular direction; Determining the angular direction with the maximum energy according to the energy integration result; The angular direction at which the energy is maximum is used as the directional feature.

8. A textile surface defect detection method according to claim 4, characterized in that: The step of determining a local background energy reference within the frequency interval comprises: sorting energy values ​​of the one or more adjacent regions within the frequency interval; According to the sorting result, a specific percentile value in the sorting result is selected as the local background energy benchmark.

9. A textile surface defect detection method according to claim 8, characterized in that: The step of selecting a specific quantile value in the sorting result as the local background energy benchmark includes: Analyzing the distribution trend of energy values ​​in the sorting results; Determining a percentile value representing an area where background energy is concentrated based on the distribution trend; The percentile value is used as the local background energy benchmark.

10. A textile surface defect detection system for online detection of textiles with a heat-sensitive coating on the surface, characterized in that: The system includes: A modulation module, configured to apply a preset periodic intensity modulation to the detection light source used for online detection, so as to generate a reference sequence as a benchmark for synchronization analysis; an acquisition module for acquiring, for a suspected defect point initially identified on the textile surface, a series of brightness values ​​of the suspected defect point within a preset time period under the periodic intensity modulation, so as to form a brightness response sequence representing its optical response characteristics; a calculation module, configured to calculate a degree of time synchronization between the brightness response sequence and the reference sequence; A determination module is configured to determine, based on the degree of time synchronization, whether the suspected defect point is a real physical defect or a temporary optical artifact caused by a thermal effect.

Images