Method and system for measuring color of dyed polyester fabric based on computer vision

Abstract

The invention provides a dyed polyester fabric color measuring method and system based on computer vision, and relates to the technical field of image processing.The method includes the steps that temperature change data under the current production condition are collected, a preset temperature interval is set, and uniform dynamic temperature change is applied in the interval; the thermal imaging characteristics of the dyed polyester fabric in the heating and cooling processes can be accurately captured, and then key characteristics such as temperature gradient and thermal diffusion rate are extracted; the method not only improves the accuracy of color difference measurement, but also realizes scientific evaluation of color performance by calculating the combined analysis of the humidity correction coefficient and the color difference deviation coefficient, effectively eliminates the interference of environmental factors, and ensures the accuracy of color measurement in the production process.

Description

Technical Field

[0001] The invention relates to the technical field of image processing, and in particular to a method and system for measuring the color of dyed polyester fabric based on computer vision. Background Art

[0002] With the continuous advancement of science and technology and the rapid development of the textile industry, the color measurement and control technology of dyed polyester fabrics has also been evolving. The traditional dyeing process usually relies on manual experience and subjective judgment, and lacks objective and accurate measurement methods. In recent years, the development of computer vision technology has brought new opportunities for the color measurement of textiles. Automated color recognition systems based on image processing and machine learning have been gradually applied to the dyeing industry, which can provide fast and accurate color difference analysis by analyzing the color information in the image. However, the existing technology still faces many challenges in practical applications, such as changes in the measurement environment, the influence of temperature and humidity on color performance, etc., which limits the reliability and consistency of color difference measurement results.

[0003] In the prior art, the publication number is CN202310027735.2, and the name is an image-based dyed polyester fabric color measurement method. Under a standard D65 light source environment, the original raw format response value and predicted CIELAB value of an unknown polyester fabric sample are obtained, and the original raw format response value and true CIELAB value of M kinds of dyed polyester fabric samples are obtained at the same time, and P dyed polyester fabric samples with the highest similarity to the unknown polyester fabric sample are selected, and their true CIELAB values ​​are averaged with the predicted CIELAB values ​​of the unknown polyester fabric sample, and the average value is used as the true CIELAB value of the unknown polyester fabric sample. The method of the present invention has high accuracy and good stability.

[0004] Existing color measurement technologies for dyed polyester fabrics have poor adaptability in dynamic environments, especially in production processes where temperature and humidity change frequently. Traditional color measurement methods often fail to reflect the actual color performance of the fabric in real time. Specifically, temperature changes can cause significant visual differences in the color of dyed polyester fabrics, especially the thermal imaging characteristics during heating and cooling, which are often not fully considered. In addition, existing technologies mostly focus on color measurement under static conditions and fail to effectively integrate the relationship between dynamic temperature changes and color performance. Therefore, in a changing production environment, the computer vision measurement results of existing methods may have color measurement errors and cannot accurately evaluate the color consistency and quality of dyed polyester fabrics.

[0005] The above information disclosed in the above Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not form the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0006] The object of the present invention is to provide a method and system for measuring the color of dyed polyester fabric based on computer vision to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] The color measurement method of dyed polyester fabric based on computer vision includes the following steps:

[0009] Step S1: collecting temperature change data and humidity change data of the dyed polyester fabric sample under the current production conditions, and setting the temperature range determined by the temperature change data as a predetermined temperature interval; setting the humidity range determined by the humidity change data as a predetermined humidity interval;

[0010] Step S2: applying a uniform dynamic temperature change treatment to the dyed polyester fabric sample within a predetermined temperature range, and capturing the dynamic thermal imaging characteristics and RGB values ​​of the dyed polyester fabric sample in the heating or cooling process in real time, and extracting the key features corresponding to each heating or cooling process from these dynamic thermal imaging characteristics through computer vision technology, the key features including the average temperature gradient and the average heat diffusion rate;

[0011] Step S3: During the heating process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next heating processes, a first measurement error rate indicating the degree of visual measurement color difference error of the current heating process is calculated;

[0012] During the cooling process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next cooling processes, a second measurement error rate indicating the degree of visual measurement color difference error of the current cooling process is calculated;

[0013] Step S4: Divide the predetermined humidity interval into a number of continuous humidity zones, collect the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone, and analyze the RGB value measurement error data to generate a humidity correction coefficient, which is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample;

[0014] Step S5: determining whether the dyed polyester fabric sample is in a heating process at the current moment; if it is in a heating process, combining and analyzing the first measurement error rate and the humidity correction coefficient to generate a first measurement error calibration index;

[0015] If it is a cooling process, the second measurement error rate and the humidity correction coefficient are combined and analyzed to generate a second measurement error calibration index;

[0016] The first measurement error calibration index and the second measurement error calibration index can both be used to generate different levels of adjustment strategies for error calibration of the RGB value of the current dyed polyester fabric sample;

[0017] Step S6: extracting the RGB value of the current dyed polyester fabric sample using computer vision technology, and performing error calibration on the RGB value of the current dyed polyester fabric sample according to different levels of adjustment strategies.

[0018] A computer vision-based dyed polyester fabric color measurement system, the system is used to perform the computer vision-based dyed polyester fabric color measurement method, comprising:

[0019] Predetermined interval setting module: used to collect temperature change data and humidity change data of dyed polyester fabric samples under current production conditions, and set the temperature range determined by the temperature change data as the predetermined temperature interval; set the humidity range determined by the humidity change data as the predetermined humidity interval;

[0020] Key feature extraction module: used to apply uniform dynamic temperature change processing to the dyed polyester fabric sample within a predetermined temperature range, and to capture the dynamic thermal imaging characteristics and RGB values ​​of the dyed polyester fabric sample in the heating or cooling process in real time, and to extract the key features corresponding to each heating or cooling process from these dynamic thermal imaging characteristics through computer vision technology. The key features include the average temperature gradient and the average thermal diffusion rate;

[0021] A measurement error rate calculation module: used for calculating a first measurement error rate representing the degree of visual measurement color difference error of the current heating process according to the visual measurement color difference error of the dyed polyester fabric sample between the current, previous and next heating processes for each key feature during the heating process;

[0022] During the cooling process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next cooling processes, a second measurement error rate indicating the degree of visual measurement color difference error of the current cooling process is calculated;

[0023] Humidity correction coefficient generation module: used to divide the predetermined humidity interval into several continuous humidity zones, collect the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone, and analyze these RGB value measurement error data to generate a humidity correction coefficient. The humidity correction coefficient is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample;

[0024] The measurement error calibration index generation module is used to determine whether the dyed polyester fabric sample is in a heating process at the current moment. If it is a heating process, the first measurement error rate and the humidity correction coefficient are combined and analyzed to generate a first measurement error calibration index;

[0025] If it is a cooling process, the second measurement error rate and the humidity correction coefficient are combined and analyzed to generate a second measurement error calibration index;

[0026] The first measurement error calibration index and the second measurement error calibration index can both be used to generate different levels of adjustment strategies for error calibration of the RGB value of the current dyed polyester fabric sample;

[0027] Error calibration module: used to extract the RGB value of the current dyed polyester fabric sample using computer vision technology, and perform error calibration on the RGB value of the current dyed polyester fabric sample according to different levels of adjustment strategies.

[0028] Compared with the prior art, the beneficial effects of the present invention are as follows: by collecting temperature change data under current production conditions and setting a predetermined temperature range, uniform dynamic temperature changes are applied within the range, so that the thermal imaging characteristics of dyed polyester fabrics during the heating and cooling processes can be accurately captured, and key features such as temperature gradient and thermal diffusion rate can be extracted; this method not only improves the accuracy of color difference measurement, but also realizes a scientific evaluation of color performance by calculating the combined analysis of the humidity correction coefficient and the color difference deviation coefficient, effectively eliminating the interference of environmental factors, and ensuring the accuracy of color measurement in the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0030] Figure 2 This is a block diagram of the overall system module of the present invention. DETAILED DESCRIPTION

[0031] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0032] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0033] Embodiment 1:

[0034] See also Figure 1 , the present invention provides a technical solution:

[0035] The color measurement method of dyed polyester fabric based on computer vision includes the following steps:

[0036] Step S1: collecting temperature change data and humidity change data of the dyed polyester fabric sample under the current production conditions, and setting the temperature range determined by the temperature change data as a predetermined temperature interval; setting the humidity range determined by the humidity change data as a predetermined humidity interval;

[0037] Step S2: applying a uniform dynamic temperature change treatment to the dyed polyester fabric sample within a predetermined temperature range, and capturing the dynamic thermal imaging characteristics and RGB values ​​of the dyed polyester fabric sample in the heating or cooling process in real time, and extracting the key features corresponding to each heating or cooling process from these dynamic thermal imaging characteristics through computer vision technology, the key features including the average temperature gradient and the average heat diffusion rate;

[0038] Step S3: During the heating process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next heating processes, a first measurement error rate indicating the degree of visual measurement color difference error of the current heating process is calculated;

[0039] The steps to determine the color difference error of visual measurement are as follows:

[0040] 1.11) Determine the standard color:

[0041] Reference standards: Standard colors usually come from authoritative color databases or color cards, such as Pantone, Munsell, or CIELAB standard color cards; these standards provide precise color definitions for comparison with samples.

[0042] Laboratory measurement: Standard colors are obtained through laboratory measurement, using calibrated colorimeter and other equipment to measure the target color under controlled conditions.

[0043] 1.12) Calculation of color difference error in visual measurement:

[0044] Color difference calculation: The color data extracted from the dyed polyester fabric sample using computer vision technology is compared with the standard color data. The comparison process is as follows:

[0045] Determine the start time t0 and end time t1 of the heating or cooling process. Computer vision technology extracts color data at t0 and t1 respectively, and calculates the difference between the color data extracted at these two times to obtain ΔRGB i =(ΔR i ,ΔG i ,ΔB i );

[0046] Using the CIELAB color space, the color difference is quantified by calculating the ΔE value, and the measurement error of the RGB value corresponding to the i-th heating process is expressed as ΔRGB i =(ΔR i ,ΔG i ,ΔB i );The RGB value measurement error corresponding to the jth cooling process is expressed as ΔRGB j =(ΔR j ,ΔG j ,ΔB j );

[0047] Data Analysis: Analyze color difference data to determine the degree of match between sample color and standard color.

[0048] During the cooling process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next cooling processes, a second measurement error rate indicating the degree of visual measurement color difference error of the current cooling process is calculated;

[0049] Step S4: Divide the predetermined humidity interval into a number of continuous humidity zones, collect the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone, and analyze the RGB value measurement error data to generate a humidity correction coefficient, which is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample;

[0050] Step S5: determining whether the dyed polyester fabric sample is in a heating process at the current moment; if it is in a heating process, combining and analyzing the first measurement error rate and the humidity correction coefficient to generate a first measurement error calibration index;

[0051] If it is a cooling process, the second measurement error rate and the humidity correction coefficient are combined and analyzed to generate a second measurement error calibration index;

[0052] The first measurement error calibration index and the second measurement error calibration index can evaluate the degree of color measurement error of the dyed polyester fabric sample according to the current key characteristics and humidity partition of the dyed polyester fabric sample;

[0053] The first measurement error calibration index and the second measurement error calibration index can both be used to generate different levels of adjustment strategies for error calibration of the RGB value of the current dyed polyester fabric sample;

[0054] Step S6: extracting the RGB value of the current dyed polyester fabric sample using computer vision technology, and performing error calibration on the RGB value of the current dyed polyester fabric sample according to different levels of adjustment strategies.

[0055] Further description, obtaining the predetermined temperature range and the predetermined humidity range specifically includes:

[0056] Arrange multiple temperature sensors on the surface of the dyed polyester fabric sample to ensure that the sensors are evenly distributed and cover the entire sample area;

[0057] Start the temperature sensor to record the real-time temperature change data of the dyed polyester fabric sample under the current production conditions;

[0058] Monitor the collected temperature data in real time and record the temperature change curve;

[0059] Analyze the temperature variation curve, determine its maximum and minimum values, and set a predetermined temperature range based on these values;

[0060] And set the predetermined temperature range as [T init , T final ], where T init , T final are the minimum and maximum temperature in the temperature change curve respectively;

[0061] Record the real-time humidity change data of dyed polyester fabric samples under current production conditions; and record the humidity change curve;

[0062] Analyze the humidity change curve, determine its maximum and minimum values, and set the predetermined humidity range as [SD init , S.D.final ], where SD init , S.D. final They are the minimum and maximum humidity in the humidity change curve respectively.

[0063] Further explanation: key features are extracted from dynamic thermal imaging characteristics, including:

[0064] 1.1) For dynamic temperature change processing:

[0065] Based on the predetermined temperature interval [T init , T final ], the heating rate of the heating process is defined as CSv;

[0066]

[0067] The cooling rate of the cooling process is defined as CJv;

[0068]

[0069] Wherein, ΔT1 and ΔT2 are the temperature change values ​​of the heating process and the cooling process under the unit time t, respectively; the unit time t in this embodiment is in minutes;

[0070] 1.11) The number of each heating process is marked to form a heating process number sequence set {1, 2, ..., i, ..., n}, where i represents the index of the i-th heating process, and the number of each heating process corresponds to the corresponding heating interval, and n represents the total number of heating processes; and when i = 1,

[0071] T1 represents the temperature value after the first heating process; and T n =T init +n×ΔT1=T final ;

[0072] 1.12) Each cooling process is marked to form a cooling process frequency sequence set {1, 2, ..., j, ..., m}, where j represents the index of the jth cooling process, and the number of each cooling process corresponds to the corresponding cooling interval, and m represents the total number of cooling processes; and when j = 1,

[0073] T1 represents the temperature value after the first cooling process; and T m =T final -m×ΔT2=T init ;

[0074] 1.2) For dynamic thermal imaging feature capture:

[0075] The acquisition frequency of thermal imaging is defined as f. During the heating or cooling process, the thermal imaging of the dyed polyester fabric sample is captured at the acquisition frequency f.

[0076] In this embodiment, the thermal imaging acquisition frequency f = 30 frames / second;

[0077] 1.3) Extraction of key features corresponding to each heating or cooling process:

[0078] Perform median filtering to denoise the captured thermal images;

[0079] I filtered (x,y)=median(I(x+i′,y+j′)),i′,j′∈{-1,0,1}

[0080] Where I(x,y) is the intensity value of the original thermal image at the pixel position (x,y); the pixel position (x,y) represents the pixel at the coordinates (x,y);

[0081] I filtered (x, y) is the intensity value of the filtered thermal image at the pixel position (x, y);

[0082] Median is a median operation that calculates the median of the pixel values ​​in the neighborhood of the central pixel;

[0083] i′, j′ represent the pixel offsets to the left and right and up and down respectively from the current pixel position (x, y), and form a 3x3 window to perform median filtering;

[0084] For the calculation of the average temperature gradient: determine the start time t0 and end time t1 of each heating or cooling process, and calculate the temperature gradient at these two times and average them; specifically include:

[0085] For each time t0 and t1, the temperature gradient is calculated:

[0086]

[0087] in, is the temperature gradient vector of the pixel at coordinate (x, y) at time t0;

[0088] is the temperature gradient vector of the pixel at coordinate (x, y) at time t1;

[0089] represent the partial derivatives of temperature with respect to the x and y coordinates respectively;

[0090] Average the temperature gradients at t0 and t1:

[0091]

[0092] Consider the global average temperature gradient to calculate the spatially average temperature gradient:

[0093]

[0094] For the calculation of the average thermal diffusion rate: determine the change in temperature at the two times t0 and t1;

[0095] For each time t0 and t1 of the heating or cooling process, calculate the temperature Laplace operator:

[0096]

[0097] Based on the Laplace operator, calculate the heat diffusion rate at (x,y) and find the average:

[0098]

[0099] Among them RKv ( x,y ) is the thermal diffusion rate at coordinate (x,y), are the Laplace operators corresponding to the temperature at time t0 and t1, respectively, representing the sum of the second-order derivatives of temperature in space; α is the thermal diffusion coefficient; represents the physical properties of the material; M and N are the width and height of the thermal imaging, respectively;

[0100] Calculate the global average heat diffusion rate to take all points in space into account. The calculation formula is as follows:

[0101]

[0102] is the average heat diffusion rate;

[0103] Through the above steps, the average temperature gradient and the average thermal diffusion rate can be calculated for evaluating the thermal characteristics of materials or structures under different temperature change conditions.

[0104] The measurement error of the RGB value of the dyed polyester fabric sample corresponding to the i-th heating process is defined as ΔRGB i =(ΔR i ,ΔG i ,ΔB i );and ΔRGB i The corresponding key features are expressed in a mapping manner as follows:

[0105]

[0106] where f i,R ,f i,G ,f i,BMap the corresponding features to ΔRGB respectively i The functions of the components in ;

[0107] The measurement error of the RGB value of the dyed polyester fabric sample corresponding to the jth cooling process is defined as ΔRGB j =(ΔR j ,ΔG j ,ΔB j ), and ΔRGB j The corresponding key features are expressed in a mapping manner as follows:

[0108]

[0109] where f j,R ,f j,G ,f j,B Map the corresponding features to ΔRGB respectively j Functions of the components in .

[0110] To further illustrate, the first measurement error rate and the second measurement error rate specifically include:

[0111] The first measurement error rate corresponding to the i-th heating process is defined as The calculation formula is as follows:

[0112]

[0113] Among them, ΔRGB i-1 and ΔRGB i+1 are the RGB value measurement errors corresponding to the previous heating process and the next heating process respectively; Respectively represent the measurement error rates corresponding to the changes in the R, G, and B values ​​during the i-th heating process;

[0114] Using the Sigmoid function, The output value is limited to the range (0,1), and the specific form is:

[0115]

[0116] in is the adjusted first measurement error rate;

[0117] The closer the value is to 1, the higher the rate of change of the RGB measurement error in the i-th heating process;

[0118] The closer the value is to 0, the lower the change rate of the RGB measurement error in the i-th heating process;

[0119] The second measurement error rate corresponding to the jth cooling process is defined as The calculation formula is as follows:

[0120]

[0121] Among them, ΔRGB j-1 and ΔRGB j+1 They are the RGB value measurement errors corresponding to the previous cooling process and the next cooling process respectively; Respectively represent the measurement error rates corresponding to the changes in the R, G, and B values ​​during the jth cooling process;

[0122] Using the Sigmoid function, The output value is limited to the range (0,1), and the specific form is:

[0123]

[0124] in is the adjusted second measurement error rate;

[0125] The closer the value is to 1, the higher the rate of change of the RGB measurement error in the jth cooling process;

[0126] The closer the value is to 0, the lower the change rate of the RGB measurement error in the jth cooling process.

[0127] Further explanation: the predetermined humidity interval is divided into several continuous humidity zones, the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone is collected, and the RGB value measurement error data are analyzed to generate a humidity correction coefficient. The humidity correction coefficient is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample, specifically including:

[0128] The preset humidity interval [SD init , S.D. final ] is divided into H consecutive humidity partitions of the same length, and these humidity partitions are recorded as {1,2,…,h,…,H}, where h represents the index of the hth humidity partition and H is the total number of humidity partitions;

[0129] When h = 1, the initial value of the humidity zone is SD init ; When h=H, the maximum value of the humidity zone is D final

[0130] The number of times the RGB value measurement error data is collected in each humidity partition is recorded as {1, 2, ..., u, ..., U}, where u represents the index of the u-th number of times the RGB value measurement error is collected, and U is the total number of times collected;

[0131] Define the humidity correction coefficient of the hth humidity zone as HC h , the calculation formula is as follows:

[0132]

[0133] Setting HC h The valid value range of is (0,1);

[0134] When HC h The closer it is to 1, the greater the change rate of the RGB measurement error of the hth humidity partition;

[0135] When HC h The closer it is to 0, the smaller the change rate of the RGB measurement error of the hth humidity partition is.

[0136] Further description, the acquisition of the first measurement error calibration index and the second measurement error calibration index specifically includes:

[0137] The temperature of the dyed polyester fabric sample at the current moment is compared with the temperature at the previous moment. If the temperature value increases, it is a heating process, and the temperature at the current moment is matched with each temperature change interval corresponding to the heating process frequency sequence set {1,2,…,i,…,n} to determine the index i of the heating process frequency; the previous moment is determined by the time length corresponding to each heating process or cooling process;

[0138] If the temperature value decreases, it is a cooling process. The temperature value at the current moment is matched with each corresponding temperature change interval in the cooling process number sequence set {1, 2, ..., j, ..., m} to determine the index j of the cooling process number;

[0139] In the i-th heating process, the first measurement error calibration index is defined as ΔEc i , the calculation formula is as follows:

[0140]

[0141] Where a1 and a2 are weight coefficients of the corresponding parameters, and a1+a2=1, and the value ranges of a1 and a2 are both in the interval (0,1);

[0142] μ1 is the first adjustment factor, 0.11≤μ1≤0.65, μ1 is used to ensure ΔEc i The valid value range of is the interval (0,1);

[0143] set up and HC h The benchmark judgment thresholds are q1 i and q2 h ; and q1 iand q2 h The value ranges are both set to (0.1, 0.9); Set ΔEc i The effective value range is (0, 1);

[0144] In the i-th heating process, when the first measurement error calibration index of the dyed polyester fabric sample exceeds 30% of the total RGB value measurement error of the current heating process, the value at this time is taken as the value of q1 i ;

[0145] In the h-th humidity zone, when the RGB value measurement error of the dyed polyester fabric sample exceeds 15%, the HC h value at this time is taken as the value of q2 h ;

[0146] When ΔEc i ≥0.5, and HC h ≥q2 h , both a1 and a2 are set to 0.5, a1 = a2, indicating that and HC h are of equal importance. At this time and HC h both have a great influence on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity zone h corresponding to the current i-th heating process, a primary adjustment strategy is performed on the RGB value of the dyed polyester fabric sample;

[0147] When ΔEc i ≥0.5, and and HC h There is any one less than the corresponding reference judgment threshold, then the parameter weight less than the corresponding reference judgment threshold is set to qz1′, and 0 < qz1′ < 0.5; The specific value of qz1′ is determined by the expert group through experimental data; At this time or HC h has a great influence on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity zone h corresponding to the current i-th heating process, a secondary adjustment strategy is performed on the RGB value of the dyed polyester fabric sample;

[0148] When ΔEc i <0.5, and and HC h both are less than the corresponding reference judgment threshold, both a1 and a2 are set to 0.5, a1 = a2, indicating that and HC h are of equal importance. At this time and HC hAll of them have little effect on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity partition h corresponding to the current i-th heating process, a three-level adjustment strategy is implemented for the RGB value of the dyed polyester fabric sample;

[0149] The adjustment ranges of the first-level adjustment strategy, the second-level adjustment strategy, and the third-level adjustment strategy decrease in turn;

[0150] In the jth cooling process, the second measurement error calibration index is defined as ΔEc j , the calculation formula is as follows:

[0151]

[0152] Where b1 and b2 are weight coefficients of corresponding parameters, and b1+b2=1, and the value range of b1 and b2 is the interval (0,1);

[0153] μ2 is the second adjustment factor, 0.12≤μ2≤0.71, μ2 is used to ensure ΔEc j The valid value range of is the interval (0,1); in this embodiment, μ2=0.2;

[0154] set up The benchmark judgment thresholds are q1 j ; and q1 j The value range is set to (0.1, 0.9); set ΔEc j The valid value range of is (0,1);

[0155] In the jth cooling process, when the RGB value measurement error corresponding to the color difference deviation of the dyed polyester fabric sample exceeds 30%, The value is q1 j The value of

[0156] When ΔEc j ≥0.5, and HC h ≥q2 h When b1 and b2 are both set to 0.5, b1=b2, indicating and HC h are equally important, then and HC h All of them have a great influence on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity partition h corresponding to the current j-th cooling process, a four-level adjustment strategy is implemented for the RGB value of the dyed polyester fabric sample;

[0157] When ΔEc j ≥0.5, and and HC hIf any one of them is less than the corresponding reference judgment threshold, the parameter weight less than the corresponding reference judgment threshold is set to qz2′, and 0 < qz2′ < 0.5; the specific value of qz2′ is determined by the expert group based on experimental data; at this time or HC h It has a great influence on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity zone h corresponding to the current j-th cooling process, a five-level adjustment strategy is carried out for the RGB values of the dyed polyester fabric sample;

[0158] When ΔEc j < 0.5, and and HC h both are less than the corresponding reference judgment threshold, both b1 and b2 are set to 0.5, b1 = b2, indicating that and HC h have the same degree of importance. At this time and HC h both have little influence on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity zone h corresponding to the current j-th cooling process, a six-level adjustment strategy is carried out for the RGB values of the dyed polyester fabric sample;

[0159] Among them, the adjustment ranges of the four-level adjustment strategy, the five-level adjustment strategy, and the six-level adjustment strategy decrease in turn.

[0160] Furthermore, the RGB values of the current dyed polyester fabric sample are extracted by using computer vision technology, and the error calibration of the RGB values of the current dyed polyester fabric sample is carried out according to different levels of adjustment strategies, specifically including:

[0161] Collect the image of the dyed polyester fabric sample by using a high-resolution camera;

[0162] Use computer vision algorithms to extract the initial RGB values (R init , G init , B init ) of the current dyed polyester fabric sample;

[0163] The computer vision algorithm adopts any one of the image processing algorithm (Image Processing Algorithms), the edge detection algorithm (Edge Detection Algorithms), and the color clustering algorithm (Color Clustering Algorithms);

[0164] Ensure that the RGB values are within the range of 0 to 255 and perform initial normalization processing;

[0165] Determine whether the dyed polyester fabric sample is currently in a heating process or a cooling process, and determine the key features corresponding to the heating process or the cooling process, as well as the humidity zone in which the dyed polyester fabric sample is currently located; the current heating process or the cooling process is respectively recorded as i1 and j1, and i1∈{1,2,…,n}; j1∈{1,2,…,m}; the current humidity zone is recorded as h1, and h1∈{1,2,…,H};

[0166] Define the humidity correction coefficient of the current h1 humidity zone as HC h1 ;

[0167] In the i1th heating process, the first measurement error calibration index is defined as ΔEc i1 ;

[0168] In the i1th heating process, the first measurement error rate after adjustment is defined as

[0169] In the j1th cooling process, the second measurement error calibration index is defined as ΔEc j1 ;

[0170] In the j1th cooling process, the second measurement error rate after adjustment is defined as

[0171] Define the initial RGB value (R init ,G init ,B init ) value adjustment formula is as follows:

[0172]

[0173] Among them (R adj ,G adj ,B adj ) is the RGB value after fine-tuning; when ΔRGB ve >0, indicating that during the ve-th temperature change trend, the RGB value measurement error is increasing; at this time, it is necessary to determine the RGB value corresponding to the RGB value and the initial RGB value (R init ,G init ,B init ) and judge the result according to the size. init ,G init ,B init ) for fine-tuning; the size judgment results refer to the following content for the same judgment method:

[0174] η1, η2, η3 are correction factors, and the value range of η1, η2, η3 is (0,1);

[0175] The ve-th temperature change trend process represents the i1-th heating process or the j1-th cooling process;

[0176] When ΔRGB ve <0, indicating that during the ve-th temperature change trend, the RGB value measurement error is decreasing; at this time, if the RGB value measurement error corresponds to an RGB value greater than the initial RGB value (R init ,G init ,B init ) value, the initial RGB value (R init ,G init ,B init ) value; F adj is the fine-tuning factor; if the RGB value measurement error corresponds to an RGB value less than the initial RGB value (R init ,G init ,B init ) value, then through F adj To increase the initial RGB value (R init ,G init ,B init ) value;

[0177] Reduce the initial RGB value (R init ,G init ,B init ) value corresponding to F adj for

[0178] Increase the initial RGB value (R init ,G init ,B init ) value corresponding to F adj for

[0179] If ΔRGB ve =0, then F adj =1;

[0180] When ΔEc i1 ≥0.5, and HC h1 ≥q2 h1 When , the η1, η2, and η3 corresponding to the first-level adjustment strategy are

[0181] When ΔEc i1 ≥0.5, and and HC h1 When any one of them is less than the corresponding benchmark judgment threshold, the η1, η2, and η3 corresponding to the secondary adjustment strategy are

[0182] When ΔEci1 <0.5, and and HC h1 When all are less than the corresponding benchmark judgment threshold, the η1, η2, and η3 corresponding to the three-level adjustment strategy are

[0183] When ΔEc j1 ≥0.5, and HC h1 ≥q2 h1 When , the η1, η2, and η3 corresponding to the four-level adjustment strategy are

[0184] When ΔEc j1 ≥0.5, and and HC h1 When any one of them is less than the corresponding benchmark judgment threshold, the η1, η2, and η3 corresponding to the five-level adjustment strategy are

[0185] When ΔEc j1 <0.5, and and HC h1 When all are less than the corresponding benchmark judgment threshold, the η1, η2, and η3 corresponding to the six-level adjustment strategy are

[0186] Make sure the RGB value (R adj ,G adj ,B adj ) is still in the range of 0 to 255.

[0187] Embodiment 2:

[0188] See also Figure 2 , a dyed polyester fabric color measurement system based on computer vision, the system is used to perform the dyed polyester fabric color measurement method based on computer vision, comprising:

[0189] Predetermined interval setting module: used to collect temperature change data and humidity change data of dyed polyester fabric samples under current production conditions, and set the temperature range determined by the temperature change data as the predetermined temperature interval; set the humidity range determined by the humidity change data as the predetermined humidity interval;

[0190] Key feature extraction module: used to apply uniform dynamic temperature change processing to the dyed polyester fabric sample within a predetermined temperature range, and to capture the dynamic thermal imaging characteristics and RGB values ​​of the dyed polyester fabric sample in the heating or cooling process in real time, and to extract the key features corresponding to each heating or cooling process from these dynamic thermal imaging characteristics through computer vision technology. The key features include the average temperature gradient and the average thermal diffusion rate;

[0191] A measurement error rate calculation module: used for calculating the first measurement error rate of the dyed polyester fabric sample between the current heating process and the previous and next heating processes for each key feature during the heating process;

[0192] During the cooling process, for each key feature, calculating the second measurement error rate of the dyed polyester fabric sample between the current cooling process and the previous and next cooling processes;

[0193] Humidity correction coefficient generation module: used to divide the predetermined humidity interval into several continuous humidity zones, collect the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone, and analyze these RGB value measurement error data to generate a humidity correction coefficient. The humidity correction coefficient is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample;

[0194] The measurement error calibration index generation module is used to determine whether the dyed polyester fabric sample is in a heating process at the current moment. If it is a heating process, the first measurement error rate and the humidity correction coefficient are combined and analyzed to generate a first measurement error calibration index;

[0195] If it is a cooling process, the second measurement error rate and the humidity correction coefficient are combined and analyzed to generate a second measurement error calibration index;

[0196] The first measurement error calibration index and the second measurement error calibration index can evaluate the degree of color measurement error of the dyed polyester fabric sample according to the current key characteristics and humidity partition of the dyed polyester fabric sample;

[0197] The first measurement error calibration index and the second measurement error calibration index can both be used to generate different levels of adjustment strategies for error calibration of the RGB value of the current dyed polyester fabric sample;

[0198] Error calibration module: used to extract the RGB value of the current dyed polyester fabric sample using computer vision technology, and perform error calibration on the RGB value of the current dyed polyester fabric sample according to different levels of adjustment strategies.

[0199] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0200] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0201] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0202] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A color measurement method for dyed polyester fabric based on computer vision, characterized in that: The specific steps include: Step S1: collecting temperature change data and humidity change data of the dyed polyester fabric sample under the current production conditions, and setting the temperature range determined by the temperature change data as a predetermined temperature interval; setting the humidity range determined by the humidity change data as a predetermined humidity interval; Step S2: applying a uniform dynamic temperature change treatment to the dyed polyester fabric sample within a predetermined temperature range, and capturing the dynamic thermal imaging characteristics and RGB values ​​of the dyed polyester fabric sample in the heating or cooling process in real time, and extracting the key features corresponding to each heating or cooling process from these dynamic thermal imaging characteristics through computer vision technology, the key features including the average temperature gradient and the average heat diffusion rate; Step S3: During the heating process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next heating processes, a first measurement error rate indicating the degree of visual measurement color difference error of the current heating process is calculated; During the cooling process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next cooling processes, a second measurement error rate indicating the degree of visual measurement color difference error of the current cooling process is calculated; Step S4: Divide the predetermined humidity interval into a number of continuous humidity zones, collect the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone, and analyze the RGB value measurement error data to generate a humidity correction coefficient, which is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample; Step S5: determining whether the dyed polyester fabric sample is in a heating process at the current moment; if it is in a heating process, combining and analyzing the first measurement error rate and the humidity correction coefficient to generate a first measurement error calibration index; If it is a cooling process, the second measurement error rate and the humidity correction coefficient are combined and analyzed to generate a second measurement error calibration index; The first measurement error calibration index and the second measurement error calibration index can both be used to generate different levels of adjustment strategies for error calibration of the RGB value of the current dyed polyester fabric sample; Step S6: extracting the RGB value of the current dyed polyester fabric sample using computer vision technology, and performing error calibration on the RGB value of the current dyed polyester fabric sample according to different levels of adjustment strategies.

2. The method for color measurement of dyed polyester fabric based on computer vision according to claim 1, characterized in that: The acquisition of the predetermined temperature range and the predetermined humidity range specifically includes: Record the real-time temperature change data of the dyed polyester fabric sample under the current production conditions; and record the temperature change curve; Analyze the temperature variation curve, determine its maximum and minimum values, and set a predetermined temperature range based on these values; And set the predetermined temperature range as [T init , T final ], where T init , T final are the minimum and maximum temperature in the temperature change curve respectively; Record the real-time humidity change data of dyed polyester fabric samples under current production conditions; and record the humidity change curve; Analyze the humidity change curve, determine its maximum and minimum values, and set the predetermined humidity range as [SD init , S.D. final ], where SD init , S.D. final They are the minimum and maximum humidity in the humidity change curve respectively.

3. The method for color measurement of dyed polyester fabric based on computer vision according to claim 2, characterized in that: Extract key features from dynamic thermal imaging characteristics, including: 1.1) For dynamic temperature change processing: Based on the predetermined temperature interval [T init , T final ], the heating rate of the heating process is defined as CSv; The cooling rate of the cooling process is defined as CJv; Among them, ΔT1 and ΔT2 are the temperature changes of the heating process and the cooling process in unit time t respectively; 1.11) Mark the number of each heating process to form a heating process number sequence set {1, 2, ..., i, ..., n}; 1.12) Each cooling process is marked to form a cooling process number sequence set {1, 2, ..., j, ..., m}, where j represents the index of the jth cooling process and m represents the total number of cooling processes; 1.2) For dynamic thermal imaging feature capture: The acquisition frequency of thermal imaging is defined as f. During the heating or cooling process, the thermal imaging of the dyed polyester fabric sample is captured at the acquisition frequency f. 1.3) Extraction of key features corresponding to each heating or cooling process: Perform median filtering to denoise the captured thermal images; During the heating or cooling process, the average temperature gradient is defined as The average thermal diffusion rate is defined as The measurement error of the RGB value of the dyed polyester fabric sample corresponding to the i-th heating process is defined as ΔRGB i =(ΔR i ,ΔG i ,ΔB i );and ΔRGB i The corresponding key features are expressed in a mapping manner as follows: where f i,R ,f i,G ,f i,B Map the corresponding features to ΔRGB respectively i The function of each component in (ΔR i ,ΔG i ,ΔB i ) represents the measurement error of the intensity of each color channel in the RGB value during the i-th heating process; The measurement error of the RGB value of the dyed polyester fabric sample corresponding to the jth cooling process is defined as ΔRGB j =(ΔR j ,ΔG j ,ΔB j ), and ΔRGB j The corresponding key features are expressed in a mapping manner as follows: where f j,R ,f j,G ,f j,B Map the corresponding features to ΔRGB respectively j The function of each component in (ΔR j ,ΔG j ,ΔB j ) represents the measurement error of the intensity of each color channel in the RGB value during the j-th cooling process.

4. The method for color measurement of dyed polyester fabric based on computer vision according to claim 3, characterized in that: The first measurement error rate and the second measurement error rate specifically include: The first measurement error rate corresponding to the i-th heating process is defined as The calculation formula is as follows: Among them, ΔRGB i-1 and ΔRGB i+1 are the RGB value measurement errors corresponding to the previous heating process and the next heating process respectively; Respectively represent the measurement error rates corresponding to the changes in the R, G, and B values ​​during the i-th heating process; Using the Sigmoid function, The output value is limited to the range (0,1), and the specific form is: in is the adjusted first measurement error rate; The closer the value is to 1, the higher the rate of change of the RGB measurement error in the i-th heating process; The closer the value is to 0, the lower the change rate of the RGB measurement error in the i-th heating process; The second measurement error rate corresponding to the jth cooling process is defined as The calculation formula is as follows: Among them, ΔRGB j-1 and ΔRGB j+1 They are the RGB value measurement errors corresponding to the previous cooling process and the next cooling process respectively; Respectively represent the measurement error rates corresponding to the changes in the R, G, and B values ​​during the jth cooling process; Using the Sigmoid function, The output value is limited to the range (0,1), and the specific form is: in is the adjusted second measurement error rate; The closer the value is to 1, the higher the rate of change of the RGB measurement error in the jth cooling process; The closer the value is to 0, the lower the change rate of the RGB measurement error in the jth cooling process.

5. The method for color measurement of dyed polyester fabric based on computer vision according to claim 4, characterized in that: The predetermined humidity interval is divided into several continuous humidity zones, and the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone are collected. These RGB value measurement error data are analyzed to generate a humidity correction coefficient. The humidity correction coefficient is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample, specifically including: The preset humidity interval [SD init , S.D. final ] is divided into H consecutive humidity partitions of the same length, and these humidity partitions are recorded as {1,2,…,h,…,H}, where h represents the index of the hth humidity partition and H is the total number of humidity partitions; The number of times the RGB value measurement error data is collected in each humidity partition is recorded as {1, 2, ..., u, ..., U}, where u represents the index of the u-th number of times the RGB value measurement error is collected, and U is the total number of times collected; Define the humidity correction coefficient of the hth humidity zone as HC h , the calculation formula is as follows: Setting HC h The valid value range is (0,1); ΔRGB h,u It represents the RGB value measurement error of the h-th humidity partition under the u-th RGB value measurement error data; When HC h The closer it is to 1, the greater the change rate of the RGB measurement error of the hth humidity partition; When HC h The closer it is to 0, the smaller the change rate of the RGB measurement error of the hth humidity partition is.

6. The method for color measurement of dyed polyester fabric based on computer vision according to claim 5, characterized in that: The first measurement error calibration index and the second measurement error calibration index are obtained, and the error calibration of the RGB value of the current dyed polyester fabric sample is performed, specifically including: The temperature of the dyed polyester fabric sample at the current moment is compared with the temperature at the previous moment. If the temperature value increases, it is a heating process. The temperature at the current moment is matched with each corresponding temperature change interval in the heating process number sequence set {1, 2, ..., i, ..., n} to determine the index i of the heating process number. If the temperature value decreases, it is a cooling process. The temperature value at the current moment is matched with each corresponding temperature change interval in the cooling process number sequence set {1, 2, ..., j, ..., m} to determine the index j of the cooling process number; In the i-th heating process, the first measurement error calibration index is defined as ΔEc i , the calculation formula is as follows: Where a1 and a2 are weight coefficients of the corresponding parameters, and a1+a2=1, and the value ranges of a1 and a2 are both in the interval (0,1); μ1 is the first adjustment factor, 0.11≤μ1≤0.65, μ1 is used to ensure ΔEc i The valid value range of is the interval (0,1); set up and HC h The benchmark judgment thresholds are q1 i and q2 h ; and q1 i and q2 h The value range is set to (0.1, 0.9); set ΔEc i The valid value range of is (0,1); When ΔEc i ≥0.5, and HC h ≥q2 h When a1 and a2 are both set to 0.5, a1=a2, indicating and HC h are equally important, then and HC h Both have a great influence on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity partition h corresponding to the current i-th heating process, the RGB value of the dyed polyester fabric sample is adjusted in a first-level strategy; When ΔEci≥0.5, and and HC h When any one of them is less than the corresponding reference judgment threshold, or HC h The color measurement error of the dyed polyester fabric sample has a great influence. Under the key features and humidity partition h corresponding to the current i-th heating process, the RGB value of the dyed polyester fabric sample is adjusted by a secondary strategy; When ΔEc i <0.5, and and HC h When both are less than the corresponding benchmark judgment threshold, a1 and a2 are both set to 0.5, a1=a2, indicating and HC h are equally important, then and HC h The color measurement error of the dyed polyester fabric sample is slightly affected. Under the key features and humidity partition h corresponding to the current i-th heating process, the RGB value of the dyed polyester fabric sample is adjusted in three levels; In the jth cooling process, the second measurement error calibration index is defined as ΔEc j , the calculation formula is as follows: Where b1 and b2 are weight coefficients of corresponding parameters, and b1+b2=1, and the value range of b1 and b2 is the interval (0,1); μ2 is the second adjustment factor, 0.12≤μ2≤0.71, μ2 is used to ensure ΔEc j The valid value range of is the interval (0,1); set up The benchmark judgment thresholds are q1 j ; and q1 j The value range is set to (0.1, 0.9); set ΔEc j The valid value range of is (0,1); When ΔEc j ≥0.5, and HC h ≥q2 h When b1 and b2 are both set to 0.5, b1=b2, indicating and HC h are equally important, then and HC h All of them have a great influence on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity partition h corresponding to the current j-th cooling process, the RGB value of the dyed polyester fabric sample is adjusted in four levels; When ΔEc j ≥0.5, and and HC h When any one of them is less than the corresponding reference judgment threshold, or HC h The color measurement error of the dyed polyester fabric sample has a great influence. Under the key features and humidity partition h corresponding to the current j-th cooling process, the RGB value of the dyed polyester fabric sample is adjusted in five levels; When ΔEc j <0.5, and and HC h When both are less than the corresponding benchmark judgment threshold, b1 and b2 are both set to 0.5, b1 = b2, indicating that and HC h are equally important, then and HC h All of them have little effect on the color measurement error of the dyed polyester fabric sample. Under the key features and humidity partition h corresponding to the current j-th cooling process, the RGB value of the dyed polyester fabric sample is adjusted in six levels; The adjustment ranges of the first-level adjustment strategy, the second-level adjustment strategy, and the third-level adjustment strategy decrease in turn; The adjustment ranges of the fourth-level adjustment strategy, the fifth-level adjustment strategy and the sixth-level adjustment strategy decrease successively.

7. A color measurement system for dyed polyester fabric based on computer vision, characterized in that: The system is used to perform the computer vision-based color measurement method for dyed polyester fabric according to any one of claims 1 to 6, comprising: Predetermined interval setting module: used to collect temperature change data and humidity change data of dyed polyester fabric samples under current production conditions, and set the temperature range determined by the temperature change data as the predetermined temperature interval; set the humidity range determined by the humidity change data as the predetermined humidity interval; Key feature extraction module: used to apply uniform dynamic temperature change processing to the dyed polyester fabric sample within a predetermined temperature range, and to capture the dynamic thermal imaging characteristics and RGB values ​​of the dyed polyester fabric sample in the heating or cooling process in real time, and to extract the key features corresponding to each heating or cooling process from these dynamic thermal imaging characteristics through computer vision technology. The key features include the average temperature gradient and the average thermal diffusion rate; A measurement error rate calculation module: used for calculating a first measurement error rate representing the degree of visual measurement color difference error of the current heating process according to the visual measurement color difference error of the dyed polyester fabric sample between the current, previous and next heating processes for each key feature during the heating process; During the cooling process, for each key feature, according to the visual measurement color difference errors of the dyed polyester fabric sample between the current, previous and next cooling processes, a second measurement error rate indicating the degree of visual measurement color difference error of the current cooling process is calculated; Humidity correction coefficient generation module: used to divide the predetermined humidity interval into several continuous humidity zones, collect the RGB value measurement error data of the dyed polyester fabric sample in each humidity zone, and analyze these RGB value measurement error data to generate a humidity correction coefficient. The humidity correction coefficient is used to evaluate the influence of each humidity zone on the visual measurement color difference error of the dyed polyester fabric sample; The measurement error calibration index generation module is used to determine whether the dyed polyester fabric sample is in a heating process at the current moment. If it is a heating process, the first measurement error rate and the humidity correction coefficient are combined and analyzed to generate a first measurement error calibration index; If it is a cooling process, the second measurement error rate and the humidity correction coefficient are combined and analyzed to generate a second measurement error calibration index; The first measurement error calibration index and the second measurement error calibration index can both be used to generate different levels of adjustment strategies for error calibration of the RGB value of the current dyed polyester fabric sample; Error calibration module: used to extract the RGB value of the current dyed polyester fabric sample using computer vision technology, and perform error calibration on the RGB value of the current dyed polyester fabric sample according to different levels of adjustment strategies.

Images