CN121275903A - Automatic detection method and system for surface flaws of printed fabric based on computer signal processing - Google Patents
Automatic detection method and system for surface flaws of printed fabric based on computer signal processingInfo
- Publication number
- CN121275903A CN121275903A CN202511831449.7A CN202511831449A CN121275903A CN 121275903 A CN121275903 A CN 121275903A CN 202511831449 A CN202511831449 A CN 202511831449A CN 121275903 A CN121275903 A CN 121275903A
- Authority
- CN
- China
- Prior art keywords
- fabric
- signal
- abnormal
- defect
- features
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Landscapes
- Investigating Or Analyzing Materials By The Use Of Ultrasonic Waves (AREA)
Abstract
The invention provides a method and a system for automatically detecting surface flaws of printed fabric based on computer signal processing, and relates to the technical field of data processing, wherein the method comprises the steps of arranging an ultrasonic transducer array above a fabric travelling path, and respectively and fixedly installing a photoelectric position sensor at the starting end and the tail end of the path; the method comprises the steps of determining a virtual reference direction based on fabric edge position information detected by two photoelectric position sensors in real time, analyzing the virtual reference direction into a plurality of continuous analysis intervals, evaluating tension and flatness properties of the fabric in each analysis interval to obtain dynamic adjustment values, adjusting emission parameters of an ultrasonic transducer array in real time based on the dynamic adjustment values, emitting regulated and controlled ultrasonic waves with specific frequency to the surface of the fabric, and receiving reflected echoes to obtain original acoustic signal data of the fabric. The invention realizes high adaptability automatic detection of the defects on the surface and the near surface of the printed fabric.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a method and a system for automatically detecting surface flaws of printed fabric based on computer signal processing.
Background
The detection of the surface flaws of the fabric is a quality control core link, but the traditional detection mode has a short plate, the detection rate of manual detection is as high as 8-25%, the detection amount of single shift detection is less than 800 linear meters, the requirement of a high-speed production line is difficult to match, the main stream AI visual detection system is limited by an optical imaging principle although the automation is realized, and the problems in the detection of complex patterns and flexible fabrics are outstanding.
When an enterprise of a certain main-camp digital printing home textile fabric adopts a 500-ten-thousand-pixel visual detection system to detect cotton-flax printed cloth containing gradual change patterns, once because 0.6mm overprinting deviation is not recognized, bedding sets are returned by customers due to unqualified quality, typical technical defects are exposed in cases, visual detection relies on surface light reflection characteristic extraction information, optical signal differences of defects such as normal texture fluctuation and overprinting deviation of gradual change patterns cannot be effectively decoupled, imaging distortion can be caused by micro-wrinkles generated by tension change in fabric travelling, detection precision of tiny and recessive defects is reduced, and the full-breadth and high-precision detection requirements of high-end printed fabrics are difficult to meet.
Disclosure of Invention
The invention aims to solve the technical problem of providing a method and a system for automatically detecting surface flaws of printed fabric based on computer signal processing, so as to realize high-adaptability automatic detection of the surface flaws and near-surface flaws of the printed fabric.
In order to solve the technical problems, the technical scheme of the invention is as follows:
In a first aspect, a method for automatically detecting surface flaws of printed fabric based on computer signal processing, the method comprising:
The method comprises the steps of setting an ultrasonic transducer array above a fabric travelling path, respectively and fixedly installing a photoelectric position sensor at the initial end and the tail end of the path, determining a virtual reference direction based on fabric edge position information detected by the two photoelectric position sensors in real time, analyzing the virtual reference direction into a plurality of continuous analysis intervals, evaluating tension and flatness properties of the fabric in each analysis interval to obtain dynamic adjustment values, real-time adjusting emission parameters of the ultrasonic transducer array based on the dynamic adjustment values, emitting regulated and controlled specific frequency ultrasonic waves to the surface of the fabric, and receiving reflected echoes to obtain original acoustic signal data of the fabric;
The method comprises the steps of preprocessing original acoustic signal data to obtain a standardized material characteristic signal, and comparing the standardized material characteristic signal with a preset reference material signal corresponding to the flawless printed fabric to obtain a signal difference result;
Carrying out multidimensional feature analysis on the signal difference result, distinguishing signal fluctuation caused by normal change of printing pattern color and material and abnormal signal feature caused by flaws from the difference, and positioning potential flaw areas through a polygon area calculation algorithm;
Based on the abnormal signal characteristics corresponding to the potential flaw area, classifying flaw types and judging the severity degree, and obtaining a judging result containing flaw types and position information;
And converting the judging result into a sorting control instruction, and separating the defective fabric section from the qualified product according to the sorting control instruction.
Further, an ultrasonic transducer array is arranged above a fabric travelling path, a photoelectric position sensor is fixedly arranged at the beginning and the tail end of the path respectively, a virtual reference direction is determined based on fabric edge position information detected by the two photoelectric position sensors in real time, the virtual reference direction is analyzed into a plurality of continuous analysis intervals, tension and flatness attributes of the fabric in each analysis interval are evaluated to obtain dynamic adjustment values, the emission parameters of the ultrasonic transducer array are adjusted in real time based on the dynamic adjustment values, regulated and controlled specific frequency ultrasonic waves are emitted to the surface of the fabric, and reflected echoes are received to obtain original acoustic signal data of the fabric, and the method comprises the following steps:
acquiring position information of the edge of the fabric in real time through two photoelectric position sensors arranged at the beginning end and the tail end of the fabric travelling path;
Determining a dynamically changing virtual reference direction based on the location information;
Analyzing the virtual reference direction into a plurality of continuous analysis intervals, and evaluating the tension and flatness properties of the fabric in each continuous analysis interval to obtain a corresponding dynamic adjustment value;
And transmitting ultrasonic waves to the surface of the fabric by using the adjusted ultrasonic transducer array and receiving reflected echoes to obtain original acoustic signal data.
Further, the method comprises the steps of preprocessing original acoustic signal data to obtain a standardized material characteristic signal, comparing the standardized material characteristic signal with a preset reference material signal corresponding to the flawless printed fabric to obtain a signal difference result, and comprises the following steps:
noise reduction and filtering are carried out on the original acoustic signal data, and a purified acoustic signal is obtained;
carrying out amplitude normalization processing on the purified acoustic signals to obtain standardized material characteristic signals;
The standardized material characteristic signals are compared with the preset reference material signals of the flawless printed fabric point by point to obtain a comparison result;
And calculating the difference degree between the signals based on the comparison result to obtain a signal difference result representing abnormal fabric state.
Further, the multi-dimensional feature analysis is performed on the signal difference result, the signal fluctuation caused by the normal change of the color and the material of the printed pattern and the abnormal signal feature caused by the flaw are distinguished from the difference, and the potential flaw area is positioned by the polygon area calculation algorithm, which comprises the following steps:
carrying out multi-dimensional feature extraction of a time domain and a frequency domain on the signal difference result to generate a feature vector set;
Based on the feature vector set, distinguishing signal fluctuation features caused by normal changes of colors and materials of the printed patterns and abnormal signal features caused by flaws through feature classification;
Performing spatial clustering on the identified abnormal signal characteristics to form abnormal characteristic distribution data;
And determining the boundary of the aggregation area of the abnormal features through a polygon area calculation algorithm based on the abnormal feature distribution data, and positioning the potential flaw area according to the boundary of the aggregation area.
Further, determining an aggregate region boundary of the abnormal feature by a polygon area calculation algorithm based on the abnormal feature distribution data, locating a potential flaw region according to the aggregate region boundary, including:
Extracting a coordinate point set of the abnormal signal characteristics from the abnormal characteristic distribution data;
Performing convex hull calculation on the coordinate point set to generate a minimum convex polygon containing all abnormal characteristic points;
calculating the area of the minimum convex polygon, and comparing the area of the minimum convex polygon with a preset area threshold;
Determining the minimum convex polygon boundary with the area exceeding a preset area threshold as an aggregation area boundary of the abnormal feature;
and positioning potential flaw areas according to the mapping positions of the aggregation area boundaries on the surface of the fabric.
Further, based on the abnormal signal characteristics corresponding to the potential flaw area, classifying flaw types and judging the severity, and obtaining a judging result containing flaw types and position information, wherein the judging result comprises the following steps:
Extracting spectral features and time domain amplitude features for classification from abnormal signal features corresponding to potential flaw areas;
Matching the extracted frequency spectrum features and time domain amplitude features with a preset flaw feature rule base to obtain a preliminary flaw classification result;
determining the severity level of the flaw according to a preset severity evaluation rule based on the intensity distribution and the spatial distribution characteristics of the abnormal signal characteristics;
And carrying out fusion treatment on the primary classification result and the severity level to obtain a judgment result containing the specific type and severity of the flaw.
Further, converting the determination result into a sorting control instruction, separating the defective fabric section from the qualified product according to the sorting control instruction, including:
receiving a judging result, and analyzing the position information, the category and the severity of the flaw from the judging result;
Calculating the expected time for the flaw to reach the sorting station by combining the real-time detected fabric travelling speed based on the analyzed flaw position information;
Based on the analyzed flaw category and severity, generating a corresponding sorting control instruction according to a preset sorting rule;
And when the calculated expected time arrives, executing the generated sorting control instruction, and triggering the sorting mechanism to separate the fabric section containing the defects from the qualified products.
In a second aspect, a system for automatically detecting surface flaws of printed fabric based on computer signal processing includes:
The system comprises an acquisition module, a virtual reference direction analysis module, a dynamic adjustment module, a control module and a control module, wherein the acquisition module is used for setting an ultrasonic transducer array above a fabric travelling path, respectively and fixedly installing a photoelectric position sensor at the initial end and the tail end of the path;
The comparison module is used for preprocessing the original acoustic signal data to obtain a standardized material characteristic signal, and comparing the standardized material characteristic signal with a preset reference material signal corresponding to the flawless printed fabric to obtain a signal difference result;
The calculation module is used for carrying out multidimensional feature analysis on the signal difference result, distinguishing signal fluctuation caused by normal change of the color and the material of the printed pattern from the difference and abnormal signal features caused by flaws, and positioning potential flaw areas through a polygonal area calculation algorithm;
The judging module is used for classifying the flaw types and judging the severity degree based on the abnormal signal characteristics corresponding to the potential flaw areas to obtain judging results containing flaw types and position information;
and the processing module is used for converting the judging result into a sorting control instruction and separating the defective fabric section from the qualified product according to the sorting control instruction.
In a third aspect, a computing device includes:
One or more processors;
And a storage means for storing one or more programs that, when executed by the one or more processors, cause the one or more processors to implement the method.
In a fourth aspect, a computer readable storage medium has a program stored therein, which when executed by a processor, implements the method.
The scheme of the invention at least comprises the following beneficial effects:
Because the ultrasonic transducer array is combined with the photoelectric position sensor to dynamically generate a virtual reference direction and adjust emission parameters in real time to adapt to fabric tension and flatness change, acoustic signals are preprocessed and compared with reference signals, normal signal fluctuation and flaw abnormal characteristics are distinguished through multi-dimensional feature analysis, potential flaw areas are positioned by utilizing a polygonal area calculation algorithm, flaw classification judgment and sorting control are completed based on feature matching, the problems of low manual detection efficiency and high omission ratio are solved, visual detection is limited by optical characteristics, such as imaging distortion caused by complicated printing pattern texture interference and fabric wrinkles, and technical defects of difficult identification of hidden or tiny flaws are overcome, so that the high-precision and high-adaptability automatic detection effect on the surface and near-surface flaws of the printed fabric is achieved, accurate positioning, classification and automatic sorting of flaws are realized, and quality control requirements of a high-speed production line are met.
Drawings
Fig. 1 is a schematic flow chart of a method for automatically detecting surface flaws of printed fabric based on computer signal processing according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of an automatic detection system for surface flaws of printed fabric based on computer signal processing according to an embodiment of the present invention.
Detailed Description
Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.
As shown in fig. 1, an embodiment of the present invention proposes a method for automatically detecting surface flaws of printed fabric based on computer signal processing, the method comprising the steps of:
The method comprises the steps of 1, setting an ultrasonic transducer array above a fabric travelling path, respectively and fixedly installing a photoelectric position sensor at the initial end and the tail end of the path, determining a virtual reference direction based on fabric edge position information detected by the two photoelectric position sensors in real time, analyzing the virtual reference direction into a plurality of continuous analysis intervals, evaluating the tension and flatness properties of the fabric in each analysis interval to obtain a dynamic adjustment value, real-time adjusting the emission parameters of the ultrasonic transducer array based on the dynamic adjustment value, emitting regulated and controlled specific frequency ultrasonic waves to the surface of the fabric, and receiving reflected echoes to obtain original acoustic signal data of the fabric;
Step 2, preprocessing the original acoustic signal data to obtain a standardized material characteristic signal, and comparing the standardized material characteristic signal with a preset reference material signal corresponding to the flawless printed fabric to obtain a signal difference result;
step 3, carrying out multidimensional feature analysis on the signal difference result, distinguishing signal fluctuation caused by normal change of printing pattern colors and materials and abnormal signal features caused by flaws from the difference, and positioning potential flaw areas through a polygon area calculation algorithm;
Step 4, based on the abnormal signal characteristics corresponding to the potential flaw area, classifying flaw types and judging the severity degree, and obtaining a judging result containing flaw types and position information;
and 5, converting the judging result into a sorting control instruction, and separating the defective fabric section from the qualified product according to the sorting control instruction.
According to the embodiment of the invention, the ultrasonic transducer array and the photoelectric position sensors at the two ends are deployed on the fabric travelling path, the virtual reference direction is determined by detecting the edge position in real time, the analysis interval is analyzed, the ultrasonic emission parameters are dynamically regulated by combining the fabric tension and flatness evaluation results to obtain the original acoustic signals, the signals are preprocessed and compared with the flawless reference signals, the normal signal fluctuation and the flawed abnormal characteristics are distinguished through multidimensional feature analysis, the potential flawed areas are positioned by means of a polygonal area calculation algorithm, and finally the flawed classification judgment is completed based on the abnormal characteristics and converted into the sorting control instruction, so that the problems of low manual detection efficiency and high omission ratio in the prior art are solved, the technical defects that the visual detection is interfered by the color material of the printed pattern, the imaging is difficult to identify the hidden or tiny flaws of the fabric are overcome, the high-precision and high-adaptability automatic detection effect on the surface and near-surface flaws of the printed fabric is further achieved, the accurate positioning, the type and severity judgment of the flawed sections of the printed fabric are automatically sorted, and the requirements of high-efficiency quality control of the printed fabric production line are met.
In a preferred embodiment of the present invention, the step 1 may include:
The method comprises the step 1.1 of acquiring position information of the edge of the fabric in real time through two photoelectric position sensors arranged at the beginning end and the tail end of a fabric travelling path, wherein the two photoelectric position sensors are fixedly arranged at corresponding positions of a frame of a production line respectively at the beginning end and the tail end of the fabric travelling path, so that detection vision lines of the two sensors are vertically aligned with the edge area of the fabric. The two photoelectric position sensors continuously emit infrared detection light beams, when the fabric passes through the detection light beams in the advancing process, the sensors can capture coordinate information of the fabric edge in real time according to position changes of the reflection light beams, and the edge position data are recorded once every 50 milliseconds, so that dynamic deviation conditions of the fabric in the advancing process can be tracked in real time.
The step 1.2 of determining a dynamically changing virtual reference direction based on the position information specifically comprises the steps of calculating a connecting line between two points based on edge coordinate points captured by two sensors and taking the connecting line as an initial reference direction, updating edge position data of the two sensors once every 100 milliseconds by a system control unit because the fabric possibly generates edge deviation due to tension fluctuation in the running process, recalculating the connecting line of the two points, dynamically correcting the initial reference direction, and finally forming the virtual reference direction which is adjusted in real time along with the change of the edge position of the fabric.
Step 1.3, analyzing the virtual reference direction into a plurality of continuous analysis intervals; the method comprises the steps of uniformly analyzing a determined dynamic virtual reference direction into a plurality of continuous analysis intervals according to the width direction of the fabric, wherein the width of each interval is set to be 10 cm, so that each interval can cover a local area of the surface of the fabric and no overlapping exists between the intervals. For each continuous analysis interval, calculating fluctuation amplitude of the edge position by comparing edge position data at different moments in the interval, dividing tension grades according to the fluctuation amplitude, judging tension to be stable and corresponding to a tension coefficient of 1.0 if the fluctuation amplitude is less than or equal to 5 mm, judging tension to be slightly unstable and corresponding to a tension coefficient of 1.2 if the fluctuation amplitude is between 5 and 10mm, judging tension to be severely unstable and corresponding to a tension coefficient of 1.5 if the fluctuation amplitude is greater than 10mm, forming a tension evaluation result, simultaneously pre-transmitting low-power ultrasonic waves to each interval, judging flatness according to amplitude consistency of reflected echoes, dividing flatness grades, judging flatness according to the amplitude difference of 15% or less of the echoes, corresponding to a flatness coefficient of 1.0 if the amplitude difference is between 15% and 30%, judging slight unevenness and corresponding to a flatness coefficient of 1.2 if the amplitude difference is greater than 30%, judging serious unevenness and corresponding to a flatness coefficient of 1.5, then setting weights of the tension coefficient and the flatness coefficient to be 50%, calculating dynamic adjustment values of each interval, and multiplying the dynamic adjustment values by 50% of the flatness coefficient by 50%.
The method comprises the steps of 1.4, adjusting the transmitting parameters of an ultrasonic transducer array in real time according to the dynamic adjusting value, transmitting ultrasonic waves to the surface of the fabric by using the adjusted ultrasonic transducer array and receiving reflected echoes to obtain original acoustic signal data, and specifically comprises the steps of transmitting the obtained dynamic adjusting value of each analysis interval to a driving module of the ultrasonic transducer array, and adjusting the transmitting parameters of the corresponding interval in the transducer array in real time by the driving module according to the dynamic adjusting value, wherein the transmitting frequency is adjusted in an adaptive manner between 200kHz and 500kHz according to the adjusting value, and if the dynamic adjusting value of a certain interval shows large tension and poor flatness, the transmitting power of the transducer corresponding to the interval is increased by 10% to 20%, and meanwhile, the pulse width is adjusted to 10 microseconds to enhance the signal penetrability. After the adjustment is completed, the ultrasonic transducer array transmits ultrasonic waves with specific frequency to the surface of the fabric area by area according to the adjustment parameters of each area, and the transducer array simultaneously receives echo signals reflected from the surface of the fabric and converts the echo signals into electric signals, so that original acoustic signal data capable of reflecting the surface and near-surface states of the fabric are finally formed.
In the embodiment of the invention, because the technical means that photoelectric position sensors are arranged at the beginning end and the tail end of the fabric travelling path to acquire edge position information in real time, a dynamic virtual reference direction is determined based on the information and is analyzed into a continuous analysis interval, the tension and the flatness of the fabric in each interval are evaluated to obtain a dynamic adjustment value, and then the transmission parameters of an ultrasonic transducer array are adjusted in real time according to the dynamic adjustment value to acquire original acoustic signal data, the technical problems that the edge is easy to deviate, the tension and the flatness of different areas are uneven during fabric travelling, the signal acquisition is unstable when the ultrasonic transmission parameters are fixed, and the accuracy of the original acoustic signal is influenced are solved, and the purposes that the transmission parameters of the ultrasonic transducer array can be dynamically adapted to the real-time state of the fabric and the actual condition of the fabric is ensured to be accurately reflected by the original acoustic signal data are achieved.
In a preferred embodiment of the present invention, the step 2 may include:
The method comprises the steps of 2.1, carrying out noise reduction and filtering treatment on original acoustic signal data to obtain purified acoustic signals, specifically, carrying out noise reduction treatment on the original acoustic signal data, acquiring interference noise in the production line environment in real time by a system, including mechanical noise generated by equipment operation, workshop airflow noise and the like, generating offset signals with the same frequency and amplitude as the interference noise but opposite phases through a self-adaptive noise elimination technology, superposing the offset signals with the original acoustic signal data to offset environment interference components, then carrying out filtering treatment, selecting a band-pass filter with a corresponding frequency range according to the ultrasonic frequency range emitted by an ultrasonic transducer array, retaining effective signals in the frequency range in the original acoustic signals, filtering high-frequency interference signals higher than 500kHz and low-frequency clutter signals lower than 200kHz, and obtaining purified acoustic signals only containing surface state information of fabrics after noise reduction and filtering treatment.
Step 2.2, carrying out amplitude normalization processing on the purified acoustic signals to obtain normalized material characteristic signals, wherein the normalized material characteristic signals comprise the steps of firstly extracting amplitude data of all signal points in the obtained purified acoustic signals, finding out the maximum amplitude and the minimum amplitude of the obtained purified acoustic signals, then calculating the difference value between the amplitude of each signal point and the minimum amplitude of each signal point according to each signal point in the purified acoustic signals, simultaneously calculating the difference value between the maximum amplitude and the minimum amplitude, dividing the amplitude difference value of each signal point by the difference value between the maximum amplitude and the minimum amplitude to obtain normalized amplitude of each signal point, enabling the amplitude of all signal points to be in a uniform range of 0-1, and eliminating signal amplitude differences caused by fine adjustment of ultrasonic emission parameters in different detection periods and different surface material areas after the amplitude normalization processing.
The method comprises the steps of 2.3, comparing standardized material characteristic signals with reference material signals of preset flawless printed fabrics point by point, wherein the reference material signals of the preset flawless printed fabrics are firstly obtained, the reference signals are required to be made of flawless fabrics which are the same as the fabrics to be detected and have the same printed patterns, the standardized material characteristic signals are generated after noise reduction filtering treatment and amplitude normalization treatment under the conditions of the same production line speed and ultrasonic emission parameters as the current detection, the standardized material characteristic signals are aligned with the reference material signals in time sequence and correspond to the same position point in the travelling direction of the fabrics, amplitude comparison is carried out on the position point corresponding to each of the two signals, and whether the difference is in the preset normal fluctuation range or not is recorded to form a comparison result containing comparison conditions of all the position points.
Step 2.4, calculating the difference degree between signals based on the comparison result to obtain a signal difference result representing the state abnormality of the fabric, wherein the method specifically comprises the steps of firstly counting the number of abnormal position points, of which the amplitude difference value exceeds a normal fluctuation range, in all the position points based on the obtained comparison result, calculating the proportion of the number of the abnormal position points to the number of the total position points, simultaneously calculating the average value of the absolute values of the amplitude difference values of all the position points, combining the average value with the proportion of the abnormal position points, calculating to obtain the difference degree comprehensively reflecting the overall difference degree of the two signals through a weight distribution rule built-in the system, and if the difference degree exceeds a preset difference degree threshold value, setting the value to be 0.08 according to the repeated detection data of the flawless fabric, judging that the state abnormality exists in the current fabric, and integrating the difference degree and the corresponding distribution information of the abnormal position points to obtain the signal difference result representing the state abnormality of the fabric.
In the embodiment of the invention, noise reduction and filtering processing are carried out on original acoustic signal data to remove environmental noise and equipment interference, amplitude normalization processing is carried out on the purified acoustic signal to unify signal dimensions, standardized material characteristic signals are compared with preset flawless printed fabric reference material signals point by point, and then the difference degree between the signals is calculated based on the comparison result, so that the technical problems that the original acoustic signal is easy to be interfered to cause distortion, the comparison accuracy is influenced by inconsistent signal amplitude under different detection scenes or fabric batches, and normal fluctuation and abnormal difference of the signal cannot be clearly distinguished are solved, and further, the purposes of acquiring pure and standardized material characteristic signals, ensuring the comparison accuracy of the signals and effectively extracting the signal difference result representing abnormal state of the fabric are achieved.
In a preferred embodiment of the present invention, the step 3 may include:
Step 3.1, carrying out multi-dimensional feature extraction on a signal difference result to generate a feature vector set, wherein the method specifically comprises the steps of firstly carrying out time domain feature extraction on the obtained signal difference result, counting the peak value size, duration time, amplitude change rate of the signal difference result corresponding to each abnormal position point in a time dimension and the time interval between adjacent abnormal position points, wherein features can reflect the fluctuation rule of the signal difference in the time layer, then carrying out frequency domain feature extraction, decomposing the signal difference result into components with different frequencies through a signal conversion means, extracting the feature frequencies, frequency spectrum energy distribution, frequency bandwidth and energy duty ratio of each frequency component, wherein the features can reflect the distribution characteristics of the signal difference in the frequency layer, finally integrating the time domain features corresponding to each abnormal position point with the frequency domain features to form feature vectors comprising parameters such as peak value, duration time, feature frequency, energy duty ratio and the like, and jointly forming the feature vector set of all abnormal position points.
Step 3.2, distinguishing signal fluctuation characteristics caused by normal changes of printing pattern colors and materials and abnormal signal characteristics caused by flaws based on a characteristic vector set, wherein the characteristic classification rule system is constructed firstly, the rule system judges that the signal fluctuation characteristics caused by normal changes of printing pattern colors and materials on the flawless fabric are obtained through collecting the signal fluctuation characteristics corresponding to the changes of different printing patterns and materials on the flawless fabric in advance, and abnormal signal characteristics corresponding to common flaws such as holes, overprinting deviations and staining are used as training data, stable classification rules are formed after training and optimization for many times, the generated characteristic vector set is input into the classification rule system, the rule system can compare the characteristic vector to be detected with normal signal fluctuation characteristic templates and abnormal signal characteristic templates in the classification rule system one by one, the matching degree is calculated, if the matching degree of a certain characteristic vector and the normal signal fluctuation characteristic templates is higher than a preset threshold, the signal fluctuation characteristics caused by normal changes of the printing pattern colors and the materials are judged, and the abnormal signal characteristics caused by flaws are judged if the matching degree of the abnormal signal characteristic templates is higher than the preset threshold, and therefore the two types of characteristics are distinguished.
And 3.3, performing spatial clustering processing on the identified abnormal signal features to form abnormal feature distribution data, wherein the method specifically comprises the steps of firstly determining a spatial clustering basis, taking detection coordinates of the surface of a fabric as a reference, taking spatial position coordinates corresponding to each abnormal signal feature as a clustering reference basis, setting a clustering distance threshold, determining the threshold according to the travelling speed and ultrasonic detection precision of the fabric, ensuring that adjacent abnormal signal features generated by the same flaw can be classified into the same class, grouping all the identified abnormal signal features by adopting a clustering algorithm, dividing the abnormal signal features with the distance between the spatial position coordinates smaller than a clustering distance threshold into a cluster, recording the spatial coordinates, corresponding difference degree values and feature vector parameters of all the abnormal signal features for each cluster, and sorting information according to the cluster to form the abnormal feature distribution data comprising the spatial distribution and feature information of each cluster.
Step 3.4, determining the boundary of the aggregation area of the abnormal features through a polygon area calculation algorithm based on the abnormal feature distribution data; extracting space coordinate points of all abnormal signal characteristics in each cluster from abnormal characteristic distribution data, calculating the coordinate points by adopting a convex hull algorithm to generate minimum convex polygons containing all coordinate points in the cluster, wherein the boundaries of the convex polygons are preliminary abnormal characteristic aggregation area boundaries, calculating the area of each minimum convex polygon, comparing the area of each minimum convex polygon with a preset area threshold, and if the area of a certain minimum convex polygon exceeds the preset area threshold, confirming that the boundary of the convex polygon is an effective abnormal characteristic aggregation area boundary; if the coordinates of the boundary of the effective aggregation area are not exceeded, judging that the coordinates of the boundary of the effective aggregation area are not exceeded, eliminating the invalid aggregation caused by the detection error, and finally mapping the coordinates of the boundary of the effective aggregation area to the actual position of the surface material according to the relative position relation between the detection equipment and the surface material, wherein the mapping position is the potential flaw area.
In the embodiment of the invention, the technical means that the multi-dimensional feature extraction of the time domain and the frequency domain is carried out on the signal difference result to generate the feature vector set, the signal fluctuation feature caused by the normal change of the printing pattern color and the material and the abnormal signal feature caused by the flaw are distinguished based on the feature vector set, the abnormal signal feature is spatially clustered to form abnormal feature distribution data, and the boundary of the abnormal feature aggregation area is determined and the potential flaw area is positioned based on the distribution data through the polygon area calculation algorithm is adopted, so that the technical problems that the normal change of the printing pattern and the flaw signal difference are difficult to effectively decouple in the traditional detection mode, the flaw feature space aggregation range cannot be accurately locked, and the potential flaw area is further positioned and fuzzy are solved, and the technical effects that the normal signal fluctuation and the flaw abnormal feature are accurately distinguished, the flaw feature aggregation range is clearly defined, and the potential flaw area is accurately positioned are achieved.
In a preferred embodiment of the present invention, the step 3.4 may include:
Step 3.41, extracting a coordinate point set of the abnormal signal features from the abnormal feature distribution data, wherein the abnormal feature distribution data comprises spatial coordinate information corresponding to each abnormal signal feature after spatial clustering is carried out on the abnormal signal features before, and the coordinates are associated with actual positions in the fabric advancing process, such as the transverse direction of the coordinates corresponds to the width direction of the fabric and the longitudinal direction of the fabric, the spatial coordinate values corresponding to all the abnormal signal features are extracted from the abnormal feature distribution data one by one, the coordinate values are orderly arranged according to detection, and a complete abnormal signal feature coordinate point set is formed, so that each coordinate point can accurately correspond to a specific position on the surface of the fabric.
Step 3.42, performing convex hull calculation on the coordinate point set to generate a minimum convex polygon containing all abnormal characteristic points, wherein the method specifically comprises the steps of leading the obtained coordinate point set into a convex hull calculation flow, analyzing the relative position relation between each coordinate point and all other coordinate points in the flow, screening out the outermost coordinate points capable of forming a closed graph around all coordinate points, namely connecting any two adjacent outermost coordinate points without passing through other abnormal coordinate points, and sequentially connecting the outermost coordinate points in a anticlockwise order to form a closed graph completely containing all abnormal signal characteristic coordinate points, wherein the graph is the minimum convex polygon containing all abnormal characteristic points.
Calculating the area of the minimum convex polygon, and comparing the area of the minimum convex polygon with a preset area threshold, wherein the method specifically comprises the steps of obtaining the actual area of the convex polygon through a geometric calculation method according to the coordinates of each vertex of the minimum convex polygon, adopting a unit consistent with the size of the fabric during calculation, ensuring that the area value can directly reflect the size of a corresponding fabric area, then calling the preset area threshold, repeatedly detecting a large number of flawless printed fabrics, counting the areas of the convex polygon corresponding to abnormal point aggregation generated by small errors of equipment and environmental interference in the detection, taking the maximum value of the areas, adding 10% of allowance, determining the maximum value, removing the small abnormal aggregation caused by non-flaw factors, comparing the calculated minimum convex polygon area with the preset area threshold, and recording the size relation of the minimum convex polygon area and the preset area threshold.
And 3.44, determining the minimum convex polygon boundary with the area exceeding the preset area threshold as an abnormal feature aggregation area boundary, wherein the method specifically comprises the steps of judging according to a comparison result, if the area of a certain minimum convex polygon is larger than the preset area threshold, indicating that abnormal signal feature aggregation in the convex polygon is not caused by detection errors or tiny interference but caused by flaws actually existing on the surface of the fabric, determining the outline formed by each side of the minimum convex polygon as an abnormal feature aggregation area boundary, and if the area of the minimum convex polygon is smaller than or equal to the preset area threshold, judging that abnormal points in the area are invalid aggregation.
Step 3.45, positioning a potential flaw area according to the mapping position of the boundary of the aggregation area on the surface of the fabric, wherein the specific method comprises the steps of firstly obtaining the relative position parameters of detection equipment and the fabric, including the installation height of an ultrasonic detection assembly, the distance between a photoelectric position sensor and the edge of the fabric and the initial position of the detection area on the travelling path of the fabric, converting the coordinate value of the boundary of the determined abnormal characteristic aggregation area into the actual physical position of the surface of the fabric according to the parameters, such as determining the centimeter of the boundary from the left edge in the width direction of the fabric and the meter of the detection initial point in the travelling direction, forming the specific mapping range of the boundary of the aggregation area on the surface of the fabric, and clearly indicating the corresponding area on the surface of the fabric according to the mapping range, wherein the corresponding area is the potential flaw area.
In the embodiment of the invention, the technical means that the coordinate point set of the abnormal signal characteristic is extracted from the abnormal characteristic distribution data, the coordinate point set is subjected to convex hull calculation to generate the minimum convex polygon containing all abnormal characteristic points, the minimum convex polygon area is calculated and compared with the preset area threshold, the minimum convex polygon boundary with the area exceeding the threshold is determined as the abnormal characteristic gathering area boundary, and then the potential defect area is positioned according to the mapping position of the boundary on the surface of the fabric is adopted, so that the technical problems that the gathering area boundary is defined in a fuzzy way due to the dispersion of the abnormal characteristic points, the tiny abnormal point gathering generated by the detection error is easily misjudged as a defect, and the abnormal characteristic area cannot accurately correspond to the actual position of the fabric are overcome, and the technical effects that the effective gathering range of the abnormal characteristic is clearly defined, the detection error interference is effectively eliminated, and the potential defect area which is completely matched with the actual surface position of the fabric is accurately positioned are achieved.
In a preferred embodiment of the present invention, the step 4 may include:
And 4.1, extracting the frequency spectrum characteristic and the time domain amplitude characteristic for classification from the abnormal signal characteristic corresponding to the potential flaw area, wherein the frequency spectrum characteristic and the time domain amplitude characteristic of the abnormal signal characteristic in the located potential flaw area are extracted, the frequency spectrum characteristic comprises the energy duty ratio of the abnormal signal in different frequency bands, the offset of the characteristic frequency and the distribution range of high-frequency components, and the time domain amplitude characteristic comprises the maximum amplitude, the average amplitude and the steepness of amplitude change of the abnormal signal and the duration of the abnormal signal. During the extraction process, all abnormal signal points in the potential flaw area need to be focused, so that the characteristics can completely reflect the abnormal signal characteristics of the area.
And 4.2, matching the extracted frequency spectrum features and time domain amplitude features with a preset flaw feature rule base to obtain a preliminary classification result of flaws, wherein the preset flaw feature rule base stores feature templates of various common flaws, the feature templates comprise that frequency spectrum energy corresponding to overprinting deviation is concentrated in a specific low frequency band, time domain amplitude fluctuation is gentle, high frequency components corresponding to holes are high, time domain amplitude mutation is obvious, characteristic frequency offset corresponding to staining is small but long in duration, and the like, the extracted frequency spectrum features and time domain amplitude features are compared with the features of each flaw template in the rule base one by one, the overall matching degree is calculated, and if the matching degree of a flaw template of a certain type is highest and exceeds a preset matching threshold, a potential flaw area is preliminarily judged as flaws of a corresponding category to form a preliminary classification result.
And 4.3, determining the severity level of the flaw according to a preset severity evaluation rule based on the intensity distribution and the spatial distribution characteristics of the abnormal signal characteristics, wherein the method specifically comprises the steps of counting the maximum amplitude and the average amplitude of the abnormal signal in the potential flaw area based on the intensity distribution of the abnormal signal characteristics, if the amplitude is larger, indicating that the flaw has stronger interference on the signal, simultaneously analyzing the spatial distribution characteristics, including the area size of the potential flaw area and the distribution density of abnormal signal points in the area, wherein the preset severity evaluation rule combines the intensity with the spatial characteristics, for example, the area exceeds 5 square centimeters and the average amplitude reaches more than 30% of a reference value, the area is 2-5 square centimeters and the average amplitude is 15% -30% of the average amplitude is medium, and the area is less than 2 square centimeters or the average amplitude is lower than 15% of the average amplitude is slight, and accordingly determining the severity level of the flaw.
And 4.4, carrying out fusion processing on the preliminary classification result and the severity level to obtain a judging result comprising specific types and severity levels of flaws, wherein the method specifically comprises the steps of integrating the obtained preliminary classification result with the determined severity level, checking the matching rationality of the classification result and the grade during integration, for example, confirming whether the severity level of overprinting deviation is consistent with the typical influence level of the flaws, and avoiding logic contradiction. Finally, a determination result comprising specific types of flaws, severity levels and corresponding position information of potential flaw areas is formed.
In the embodiment of the invention, the technical means of extracting the frequency spectrum characteristic and the time domain amplitude characteristic from the abnormal signal characteristic corresponding to the potential flaw area, matching the extracted characteristic with the preset flaw characteristic rule base to obtain the preliminary classification result, determining the severity level according to the preset rule based on the intensity distribution and the space distribution characteristic of the abnormal signal characteristic, and then fusing the preliminary classification result with the severity level are overcome, so that the technical problems that the classification erroneous judgment is caused by the difficulty in accurately extracting the flaw classification characteristic in the traditional detection, the quality control standard is not uniform due to the lack of assessment on the flaw severity level, and the classification result and the severity level are disjointed to form a complete judgment basis are solved, and the technical effects of identifying the specific flaw classification, dividing the flaw severity level and forming the complete judgment result containing the flaw classification and the severity level are achieved.
In a preferred embodiment of the present invention, the step 5 may include:
And 5.1, receiving a judging result, and analyzing the position information, the type and the severity of the flaws from the judging result, wherein the receiving and outputting of the judging result comprises the specific position, the corresponding flaw type and the severity of the potential flaw area on the surface of the fabric are clearly recorded in the judging result, the judging result is subjected to information analysis, the position information is required to be disassembled into specific coordinates of the width direction and the length direction of the fabric, such as the centimeter of the width direction from the left edge of the fabric and the meter of the length direction from the detection starting point, the flaw type is required to be clearly distinguished into overprinting deviation, hole breakage and staining type, and the severity is required to be determined into three grades of slight, medium or serious.
And 5.2, calculating the expected time for the flaw to reach the sorting station based on the analyzed flaw position information and the real-time detected fabric travelling speed, wherein the method specifically comprises the steps of firstly obtaining the coordinates in the length direction in the analyzed flaw position information, determining the linear distance between the current flaw on the fabric travelling path and the sorting station, calculating the distance according to the actual layout parameters of the production line and the real-time position of the flaw, simultaneously, detecting the travelling speed of the fabric in real time through a speed sensor arranged beside a driving roller of the production line, updating the speed data every 10 milliseconds to cope with the tiny fluctuation of the speed of the production line, dividing the calculated distance from the flaw to the sorting station by the real-time detected fabric travelling speed, and obtaining the time required for the flaw to move from the current position to the sorting station, wherein the time is the expected time for the flaw to reach the sorting station.
Step 5.3, based on the analyzed flaw types and severity, generating corresponding sorting control instructions according to preset sorting rules, wherein the preset sorting rules are prepared according to quality control standards of textile industry and production requirements of enterprises, such as sorting rules corresponding to serious-grade holes or overprinting deviations, are formed by directly removing the fabric sections to unqualified product channels, medium-grade staining rules are marked and then are temporarily stored in a to-be-rechecked area, slight-grade small-area flaws are recorded and then are continuously conveyed along with qualified products, the analyzed flaw types and severity are combined, corresponding processing modes are determined according to the preset sorting rules, and then the processing modes are converted into control instructions identifiable by a sorting mechanism, such as instructions for controlling actions of pneumatic push rods, instructions for switching directions of sorting conveying belts or instructions for starting marking devices, so that the instruction content is ensured to be completely matched with the processing modes.
And 5.4, executing the generated sorting control instruction when the calculated expected time arrives, triggering the sorting mechanism to separate the fabric section containing the defects from the qualified products, specifically comprising the steps of starting a timer and comparing the current time with the calculated expected time in real time while generating the sorting control instruction, and immediately sending the corresponding sorting control instruction to the sorting mechanism when the time displayed by the timer reaches the expected time. After the sorting mechanism receives the instruction, the sorting mechanism executes the action according to the instruction requirement, for example, if the instruction is to remove unqualified products, the pneumatic push rod stretches out within the preset time to push the fabric section containing flaws to the unqualified product recovery channel, if the instruction is to be re-inspected, the marking device prints marks at the corresponding positions of the fabric section, after the action is executed, the sorting mechanism feeds back an execution completion signal, the system confirms that the flaw fabric section is separated from the qualified product conveying path, then the timer is reset, the sorting task of the next flaw is prepared to be processed, and the continuity of the whole sorting process is ensured without affecting the normal travelling speed of the production line.
In the embodiment of the invention, the position information, the type and the severity of the flaws are received and judged, the expected time for the flaws to arrive at the sorting station is calculated by combining the real-time detected fabric travelling speed, the sorting control instruction is generated according to the preset sorting rule according to the flaw type and the severity, and the technical means of triggering the sorting mechanism to separate the flaw fabric sections is executed when the expected time arrives, so that the technical problems that the flaw position judgment deviation is large, the sorting time is delayed, the similar flaw treatment is inconsistent due to the non-uniform sorting rule, and the real-time sorting requirement of a high-speed production line cannot be met are solved, the precise control flaw sorting time is further achieved, the flaw fabric sections with different types and different severity are separated in a targeted manner according to the standardized rule, the high-efficiency and accurate distinction between the flaw fabric sections and qualified products are realized, and the real-time quality control and automatic sorting requirement of the printed fabric high-speed production line is met.
As shown in fig. 2, the embodiment of the invention further provides an automatic detection system for surface flaws of printed fabric based on computer signal processing, which comprises:
The system comprises an acquisition module, a virtual reference direction analysis module, a dynamic adjustment module, a control module and a control module, wherein the acquisition module is used for setting an ultrasonic transducer array above a fabric travelling path, respectively and fixedly installing a photoelectric position sensor at the initial end and the tail end of the path;
The comparison module is used for preprocessing the original acoustic signal data to obtain a standardized material characteristic signal, and comparing the standardized material characteristic signal with a preset reference material signal corresponding to the flawless printed fabric to obtain a signal difference result;
The calculation module is used for carrying out multidimensional feature analysis on the signal difference result, distinguishing signal fluctuation caused by normal change of the color and the material of the printed pattern from the difference and abnormal signal features caused by flaws, and positioning potential flaw areas through a polygonal area calculation algorithm;
The judging module is used for classifying the flaw types and judging the severity degree based on the abnormal signal characteristics corresponding to the potential flaw areas to obtain judging results containing flaw types and position information;
and the processing module is used for converting the judging result into a sorting control instruction and separating the defective fabric section from the qualified product according to the sorting control instruction.
While the foregoing is directed to the preferred embodiments of the present invention, it will be appreciated by those skilled in the art that various modifications and adaptations can be made without departing from the principles of the present invention, and such modifications and adaptations are intended to be comprehended within the scope of the present invention.
Claims (10)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511831449.7A CN121275903A (en) | 2025-12-08 | 2025-12-08 | Automatic detection method and system for surface flaws of printed fabric based on computer signal processing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202511831449.7A CN121275903A (en) | 2025-12-08 | 2025-12-08 | Automatic detection method and system for surface flaws of printed fabric based on computer signal processing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN121275903A true CN121275903A (en) | 2026-01-06 |
Family
ID=98232608
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202511831449.7A Pending CN121275903A (en) | 2025-12-08 | 2025-12-08 | Automatic detection method and system for surface flaws of printed fabric based on computer signal processing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN121275903A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5445029A (en) * | 1993-11-08 | 1995-08-29 | General Electric Co. | Calibration and flaw detection method for ultrasonic inspection of acoustically noisy materials |
| US20110244594A1 (en) * | 2010-03-31 | 2011-10-06 | Fujifilm Corporation | Target substance detecting method and target substance detecting apparatus |
| CN119516242A (en) * | 2024-09-30 | 2025-02-25 | 科大讯飞华南人工智能研究院（广州）有限公司 | A method for detecting textile blending ratio defects |
| CN120315293A (en) * | 2025-06-17 | 2025-07-15 | 天津市新宇彩板有限公司 | A steel plate tension dynamic optimization method and system based on model predictive control |
| CN120490306A (en) * | 2025-07-17 | 2025-08-15 | 中国测试技术研究院 | Calibration method of ultrasonic flaw detector calibrating device |
-
2025
- 2025-12-08 CN CN202511831449.7A patent/CN121275903A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5445029A (en) * | 1993-11-08 | 1995-08-29 | General Electric Co. | Calibration and flaw detection method for ultrasonic inspection of acoustically noisy materials |
| US20110244594A1 (en) * | 2010-03-31 | 2011-10-06 | Fujifilm Corporation | Target substance detecting method and target substance detecting apparatus |
| CN119516242A (en) * | 2024-09-30 | 2025-02-25 | 科大讯飞华南人工智能研究院（广州）有限公司 | A method for detecting textile blending ratio defects |
| CN120315293A (en) * | 2025-06-17 | 2025-07-15 | 天津市新宇彩板有限公司 | A steel plate tension dynamic optimization method and system based on model predictive control |
| CN120490306A (en) * | 2025-07-17 | 2025-08-15 | 中国测试技术研究院 | Calibration method of ultrasonic flaw detector calibrating device |
Non-Patent Citations (4)
| Title |
|---|
| CUADRADO， M 等: "Model updating of uncertain parameters of carbon/epoxy composite plates from experimental data", JOURNAL OF SOUND AND VIBRATION, vol. 455, 1 September 2019 (2019-09-01), pages 380 - 401 * |
| 任佳军: "基于深度学习的复合材料自动铺缠表面缺陷检测研究", 中国优秀硕士学位论文全文数据库工程科技I辑, no. 07, 15 July 2025 (2025-07-15), pages 020 - 740 * |
| 肖琨 等: "厚壁管道内部缺陷的超声柔性探头仿真分析", 武汉理工大学学报(交通科学与工程版), no. 02, 15 April 2013 (2013-04-15), pages 196 - 199 * |
| 邱海飞 等: "纱线张力自动控制与检测试验平台", 丝绸, vol. 54, no. 06, 30 June 2017 (2017-06-30), pages 27 - 32 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114757949B (en) | Wire and cable defect detection method and system based on computer vision | |
| US12469123B2 (en) | Inspecting sheet goods using deep learning | |
| CA3053219C (en) | Real-time, full web image processing method and system for web manufacturing supervision | |
| CN109886298A (en) | A welding seam quality detection method based on convolutional neural network | |
| Nandini et al. | A review on applications of machine vision systems in industries | |
| WO2022144722A1 (en) | Machine learning-based generation of rule-based classification recipes for inspection system | |
| CN118379287A (en) | Image recognition-based printed product quality monitoring and analyzing method and system | |
| CN118443799A (en) | Online defect detection method and system for calcium silicate decorative plate | |
| CN116342966A (en) | Rail inspection method, device, equipment and medium based on deep learning | |
| Reis et al. | Wavelet texture analysis of on-line acquired images for paper formation assessment and monitoring | |
| CN117147699B (en) | Medical non-woven fabric detection method and system | |
| JP3913517B2 (en) | Defect detection method | |
| Oni et al. | Patterned fabric defect detection and classification (fddc) techniques: A review | |
| Sun et al. | Online detection of eggshell cracks using adaptive light‐source intensity | |
| CN117314829A (en) | Industrial part quality inspection method and system based on computer vision | |
| CN121275903A (en) | Automatic detection method and system for surface flaws of printed fabric based on computer signal processing | |
| CN118371734B (en) | Additive manufacturing defect detection system and method based on deep learning | |
| CN118469993B (en) | Intelligent automatic picking system based on Internet of things | |
| CN114742768A (en) | Steel rail shallow layer peeling damage detection method and detection system | |
| CN120672732A (en) | Automobile wire harness precise fastener surface flaw detection system based on visual recognition | |
| Li et al. | Wood chip crack detection based on linear scale-space differential | |
| CN119022811A (en) | A film thickness measurement system for enameled flat wire | |
| Chan et al. | Automation technology for fabric inspection system | |
| Zhou et al. | Segmentation of main weld seam area based on MGLNS-Retinex image enhancement algorithm | |
| Suleymanova | ANALYSIS AND SELECTION OF TECHNICAL MEANS IN ORDER TO DETECTING DEFECTS IN FABRICS |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |