A color mark rapid detection method for multi-color gravure printing machine

Abstract

The present invention discloses a color mark rapid detection method for a multi-color group gravure printing machine, the color mark rapid detection method comprising: S1: inputting a printed finished product sample; S2: collecting the sample color mark signal in the printed finished product sample; S3: according to the sample color mark signal, using a sliding window and clustering algorithm to obtain the printing color mark signal of each color group in the multi-color group gravure printing machine. The color mark rapid detection method for a multi-color group gravure printing machine provided by the present invention can monitor the photoelectric eye signal of the color registration system in real time, can realize the color mark signal feature extraction and color mark position marking, and assist the color registration system to quickly realize automatic registration.

Description

Technical Field

[0001] The invention relates to the technical field of gravure printing, and in particular to a color mark rapid detection method for a multi-color group gravure printing machine. Background Art

[0002] With the improvement of people's material and cultural living standards and the change of consumption concepts, the domestic packaging market has grown at a rate of more than 20% per year in the past decade. Among them, the packaging printing market has grown even faster, and the quality and taste of printing are also constantly improving. Packaging printing is no longer a simple description and text, but the most effective carrier of corporate culture, corporate image, and product brand. It is also the perfect unity of the product's internal quality and product appearance. At the same time, the printing packaging with novel design and exquisite printing gives customers a deeper impression and great appeal.

[0003] Gravure printing is named for its layout characteristics. Due to its unique printing method, gravure printing has the following advantages: thick ink layer, bright colors, rich layers, high printing durability, relatively low cost, suitable for printing continuous patterns, wide application range, gravure printing can be printed on traditional paper, and can also be printed on other materials such as film, aluminum foil, transfer paper, etc., suitable for long-term investment.

[0004] The color registration control system is an important part of the gravure printing machine. It is responsible for controlling the accurate positioning of each printing color and plays a decisive role in the printing quality. With the development of motor control technology, computer technology and information technology, the photoelectric color registration system of the gravure printing machine has become a key technology for the advancement of equipment technology. The color mark detection photoelectric sensor system is widely used in the registration control system of the packaging, printing and other industries to realize the discrimination of colors and the detection of color marks. Through the use of other equipment and devices, the detection of physical quantities such as position, length and quantity can be realized.

[0005] In color printing, a color pattern is composed of multiple colors. When printing, each color pattern is printed separately in a certain order. When the relative positions of different colors are not accurate, the printed pattern will appear double images, which is a problem of printing accuracy. In order to obtain clear color patterns, the monochrome patterns of different colors must be accurately overprinted in the same position, so a servo motor position closed-loop control system is required during the printing process to correct the color registration error in real time. In order to detect and correct the color registration error, a sensor system that can detect parameters such as the overprint position and error distance is required. The color mark detection photoelectric sensor system can accurately and timely detect the color mark signal representing the printing deviation in the color registration system, providing a reliable guarantee for the stable operation of the color printing automatic registration system.

[0006] In the actual overprinting process, due to the high speed of the overprinting assembly line, unstable tension control, relative position sliding of the guide roller and the pressure roller, and the temperature sensitivity of the substrate, the deviation of the pattern overprinting position is caused. The superposition of various deviations seriously affects the quality of printed products. Due to the requirements for printing accuracy and speed, it is difficult to correct such deviations manually. The error detection mechanism uses a color mark detection photoelectric sensor system to accurately detect overprinting errors. The rapidity, real-time performance, and accuracy of the color mark detection photoelectric sensor system directly affect the error correction of the error detection mechanism and the control system, thereby affecting the quality of packaging and printed products. After the printing process of the color mark detection photoelectric sensor system begins, how can the color mark signal detected by the photoelectric system be quickly and accurately processed to obtain the accurate color mark signal position? The rapidity and accuracy of this operation seriously affect the printing quality and the amount of waste, and have a great impact on the benefits of the enterprise. Therefore, it is very necessary to quickly and accurately process the signal of the photoelectric detection system. Summary of the invention

[0007] The purpose of the present invention is to provide a color mark rapid detection method for a multi-color group gravure printing machine, which can monitor the photoelectric eye signal of the color registration system in real time, realize color mark signal feature extraction and color mark position marking, and assist the color registration system to quickly realize automatic registration.

[0008] The technical solution of the present invention to solve the above technical problems is as follows:

[0009] The present invention provides a color mark rapid detection method for a multi-color group gravure printing machine, the color mark rapid detection method comprising:

[0010] S1: Input the sample of the printed product;

[0011] S2: Collecting the sample color signal in the printed product sample;

[0012] S3: According to the sample color mark signal, a sliding window and clustering algorithm are used to obtain the printing color mark signal of each color group in the multi-color group gravure printing machine.

[0013] Optionally, in step S2, photoelectric eyes of multiple color groups in the multi-color group gravure printing machine are used to collect sample color mark signals in the sample image of the printed product.

[0014] Optionally, in step S3, the sliding window and clustering algorithm includes:

[0015] A1: dividing the printing color mark signal into a plurality of subsequences;

[0016] A2: extracting and identifying color-coded subsequences from a plurality of subsequences;

[0017] A3: using a clustering algorithm to classify and divide the color mark values ​​in the color mark subsequence;

[0018] A4: Output the color mark value as the printing color mark signal.

[0019] Optionally, the step A1 includes:

[0020] Use a sliding window of length l to divide the time series into equal lengths.

[0021] Optionally, the time series expression is:

[0022] Y=(y(t1),y(t2),y(t3)…y(t n ))

[0023] Among them, Y represents a time series with a length of n, y(t i ) is t i Data recorded at time, t i is the collection time, i=1,2,…,n.

[0024] Optionally, the step A2 comprises:

[0025] A21: Obtain a target subsequence from the plurality of subsequences;

[0026] A22: Calculate the slope between any two adjacent data in the target subsequence;

[0027] A23: repeating the steps A21 and A22 until all slopes in the target subsequence are obtained;

[0028] A24: according to all the slopes, obtaining the confidence interval distance radius of all the slopes;

[0029] A25: Determine whether the confidence interval distance radius is greater than a preset threshold. If so, output the target subsequence as a color mark subsequence; otherwise, return to step A21.

[0030] Optionally, in step A22, the slope is:

[0031]

[0032] Among them, k i represents the slope of the i-th data; Y j (i) the i-th data of the j-th subsequence; Y j (i-1) represents the i-1th data of the jth subsequence; t i is the collection time of the i-th data; t i-1 Indicates the collection time of the i-1th data.

[0033] Optionally, in step A24, the confidence interval distance radius is:

[0034]

[0035] Among them, d j is the confidence interval distance radius of the jth subsequence; is the upper confidence limit, and θ j is the lower confidence limit, and Z is a random variable that satisfies the normal distribution of N(0,1), a is the confidence level, and a=0.05 is taken. represents the mean of the slope of the jth subsequence; σ j Expressed as the mean square error of the slope of the jth subsequence.

[0036] Optionally, the step A3 comprises:

[0037] A31: cluster all signals of each color group in a multi-color gravure printing machine to obtain a cluster sample set;

[0038] A32: performing a classification operation on the cluster sample set to obtain a plurality of classified cluster samples;

[0039] A33: Calculate the cluster center of the target cluster sample;

[0040] A34: obtaining a membership matrix of the target cluster samples according to the cluster centers of the target cluster samples;

[0041] A35: Determine whether the absolute value of the difference between the membership matrix of the target cluster sample and the membership matrix of the previous cluster sample is less than the allowable error. If so, output the target cluster sample as the color scale value; otherwise, add 1 to the number of iterations and return to step A33.

[0042] Optionally, in step A33, the cluster center of the target cluster sample subset is calculated as:

[0043]

[0044] in, is the cluster center of the i-th cluster; represents the probability that the kth sample data in the membership matrix U of the previous cluster sample belongs to the i-th cluster, and is the cluster sample x k To cluster center v i distance; m represents the weighted index; x k represents the cluster sample of the kth sample data; C is a constant and represents the total number of sample data; h is the number of iterations.

[0045] The present invention has the following beneficial effects:

[0046] The present invention changes the shortcomings of the existing color registration system that it takes a long time and is inaccurate to manually find the color mark position. The color mark rapid detection method for multi-color group gravure printing machines based on sample images can monitor the photoelectric eye signal status in real time, record and process it in real time, and can quickly detect the color mark position and complete the marking during the initial printing stage of the unit or when changing the printing formula, assisting the color registration system to quickly complete the color registration, saving materials and improving efficiency. The method can also monitor and predict the color registration effect based on the recorded real-time data of the photoelectric eye. If the prediction results show that a certain color group of the printing system deviates, a real-time alarm will be issued to remind the operator to pay attention to changes in tension or other factors in the printing process, and timely adjustments will be made to quickly restore the system to stability. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A flow chart of a color mark rapid detection method for a multi-color group gravure printing machine provided by the present invention;

[0048] Figure 2 It is the structural diagram of the control system of the gravure printing group;

[0049] Figure 3 This is a schematic diagram of photoelectric eye color mark detection;

[0050] Figure 4 It is the flow chart of sliding window and clustering algorithm;

[0051] Figure 5 Schematic diagram of sliding window structure;

[0052] Figure 6 is a step-by-step flow chart of step A2;

[0053] Figure 7 This is a step-by-step flow chart of step A3. DETAILED DESCRIPTION

[0054] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0055] Example

[0056] The present invention provides a color mark rapid detection method for a multi-color group gravure printing machine, referring to Figure 1 As shown, the color mark rapid detection method includes:

[0057] S1: Input the sample of the printed product;

[0058] Here, the printed product sample is mainly input into the gravure group control system, referring to Figure 2In addition, in the color mark rapid detection method for a multi-color gravure printing machine provided by the present invention, the printed product sample image can be a printed sample image used in the previous printing, or a printed sample image used in a new printing scheme, and the present invention does not make any specific restrictions.

[0059] S2: Collecting the sample color signal in the printed product sample;

[0060] Here, reference Figure 3 As shown, the sample image color mark signal in the printed product sample image is mainly collected by the photoelectric eyes of each color group on the multi-color group gravure printing machine. Of course, those skilled in the art can also select other methods in combination with the present invention and actual conditions to collect the sample image color mark signal, and the present invention does not make specific limitations.

[0061] In addition, the sample color code signal collected here will be used as the color code reference signal for subsequent work.

[0062] S3: According to the sample color mark signal, a sliding window and clustering algorithm are used to obtain the printing color mark signal of each color group in the multi-color group gravure printing machine.

[0063] Optionally, in step S3, reference Figure 4 As shown, the sliding window and clustering algorithm includes:

[0064] A1: dividing the printing color mark signal into a plurality of subsequences;

[0065] Here, a sliding window of length l is used to divide the time series into equal lengths.

[0066] The time series expression is:

[0067] Y=(y(t1),y(t2),y(t3)…y(t n ))

[0068] Among them, Y represents a time series with a length of n, y(t i ) is t i Data recorded at time, t i is the collection time, i=1,2,…,n.

[0069] In the subsequence segmentation of the time series, a sliding window of length l (l ≤ n) is used to segment the time series into equal lengths, and then moves backward by a step length r (usually r = 1), sliding continuously (n-1) / r times to form subsequence fragments of equal length, see Figure 5 .

[0070] A2: extracting and identifying color-coded subsequences from a plurality of subsequences;

[0071] Optionally, refer to Figure 6 As shown, the step A2 includes:

[0072] A21: Obtain a target subsequence from the plurality of subsequences;

[0073] A22: Calculate the slope between any two adjacent data in the target subsequence;

[0074] A23: repeating the steps A21 and A22 until all slopes in the target subsequence are obtained;

[0075] The slope can describe the structural characteristics of the subsequence, but the slope information is randomly distributed. In order to accurately describe the distribution of the subsequence slope information, the confidence interval distance radius of the slope is used as the feature representation of the subsequence. The feature extraction steps are as follows:

[0076] Assume that one of the subsequences is The slope between any two adjacent data points in the subsequence is calculated according to the following formula:

[0077]

[0078] Among them, k i represents the slope of the i-th data; Y j (i) the i-th data of the j-th subsequence; Y j (i-1) represents the i-1th data of the jth subsequence; t i is the collection time of the ith data; t i-1 Indicates the collection time of the i-1th data.

[0079] Then in this subsequence Y j In the above equation, we get l-1 slope values, and the confidence interval distance radius of the slope is:

[0080] The confidence interval distance radius is:

[0081]

[0082] Among them, d j is the confidence interval distance radius of the jth subsequence; is the upper confidence limit, and θ j is the lower confidence limit, and Z is a random variable that satisfies the normal distribution of N(0,1), a is the confidence level, and a=0.05 is taken. represents the mean of the slope of the jth subsequence; σ j Expressed as the mean square error of the slope of the jth subsequence.

[0083] Based on the extracted subsequence features, a judgment model for suspected color-coded subsequences is given: Assume that the time series Y containing n variables is Y=(y(t1), y(t2), y(t3)…y(t n )), there exists a j-th subsequence The confidence interval radius d of the slope j >γ, where γ is the threshold, then the jth subsequence y in the time series Y is considered j is a color code subsequence, which contains color code feature values, and the classification of color code feature values ​​and sample feature values ​​is to be further realized.

[0084] A24: according to all the slopes, obtaining the confidence interval distance radius of all the slopes;

[0085] A25: Determine whether the confidence interval distance radius is greater than a preset threshold. If so, output the target subsequence as a color mark subsequence; otherwise, return to step A21.

[0086] A3: using a clustering algorithm to classify and divide the color mark values ​​in the color mark subsequence;

[0087] Optionally, refer to Figure 7 As shown, the step A3 includes:

[0088] A31: cluster all signals of each color group in a multi-color gravure printing machine to obtain a cluster sample set;

[0089] A32: performing a classification operation on the cluster sample set to obtain a plurality of classified cluster samples;

[0090] A33: Calculate the cluster center of the target cluster sample;

[0091] Optionally, in step A33, the cluster center of the target cluster sample subset is calculated as:

[0092]

[0093] in, is the cluster center of the i-th cluster; represents the probability that the kth sample data in the membership matrix U of the previous cluster sample belongs to the i-th cluster, and is the cluster sample x k To cluster center v i distance; m represents the weighted index; x k represents the cluster sample of the kth sample data; C is a constant and represents the total number of sample data; h is the number of iterations.

[0094] A34: obtaining a membership matrix of the target cluster samples according to the cluster centers of the target cluster samples;

[0095] A35: Determine whether the absolute value of the difference between the membership matrix of the target cluster sample and the membership matrix of the previous cluster sample is less than the allowable error. If so, output the target cluster sample as the color scale value; otherwise, add 1 to the number of iterations and return to step A33.

[0096] Specifically, let the cluster sample set X = {x1, x2, x3, ... x n}, now divide the data X into C categories (2≤C≤N), and set its cluster center to be V={v1,v2,v3…v C}; Let the membership matrix be U = [u ik ] C×N , where element u ik ∈[0,1] represents the probability that the k-th sample data belongs to the i-th cluster (1≤k≤N,1≤i≤C). The objective function is minimized by iteratively adjusting (U, V):

[0097]

[0098] Where: m represents the weighted index and usually m = 2; is the cluster sample x k To cluster center v i distance.

[0099] The specific steps are as follows: First, calculate the cluster center V = {v1, v2, v3…v C}

[0100]

[0101] Where h is the number of iterations.

[0102] Then update the membership matrix:

[0103]

[0104] Until ||U h -u h-1 ||<ε, where ε is the allowable error; otherwise, let h=h+1 and repeat the above steps until the conditions are met.

[0105] A4: Output the color mark value as the printing color mark signal.

[0106] The present invention changes the shortcomings of the existing color registration system that it takes a long time and is inaccurate to manually find the color mark position. The color mark rapid detection method of the multi-color group gravure printing machine based on the sample image can monitor the photoelectric eye signal status in real time, record and process in real time, and can quickly detect the color mark position and complete the marking when the unit is in the initial printing stage or when the printing formula is changed, and assist the color registration system to quickly complete the color registration, saving materials and improving efficiency. The method can also monitor and predict the color registration effect according to the recorded real-time data of the photoelectric eye. If the prediction result finds that a color group of the printing system deviates, it will alarm in real time to remind the operator to pay attention to the changes in the tension or other factors in the printing process, and make timely adjustments to make the system quickly restore stability.

[0107] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A color mark rapid detection method for a multi-color gravure printing machine, characterized in that: The color mark rapid detection method comprises the following steps: S1: Input the sample of the printed product; S2: Collecting the sample color signal in the printed product sample; S3: According to the sample color code signal, a sliding window and clustering algorithm are used to obtain the printing color code signal of each color group in the multi-color group gravure printing machine; The method of using a sliding window and clustering algorithm to obtain the printing color mark signal of each color group in the multi-color group gravure printing machine includes the following steps: A1: dividing the printing color mark signal into a plurality of subsequences; A2: extracting and identifying color-coded subsequences from a plurality of subsequences; A3: using a clustering algorithm to classify and divide the color mark values ​​in the color mark subsequence; A4: Outputting the color mark value as the printing color mark signal; The step A2 comprises: A21: Obtain a target subsequence from the plurality of subsequences; A22: Calculate the slope between any two adjacent data in the target subsequence; A23: repeating the steps A21 and A22 until all slopes in the target subsequence are obtained; A24: according to all the slopes, obtaining the confidence interval distance radius of all the slopes; A25: Determine whether the confidence interval distance radius is greater than a preset threshold. If so, output the target subsequence as a color mark subsequence; otherwise, return to step A21.

2. The color mark rapid detection method of a multi-color group gravure printing machine according to claim 1, characterized in that: In the step S2, the photoelectric eye of each color group in the multi-color group gravure printing machine is used to collect the sample image color code signal in the sample image of the printed product.

3. The color mark rapid detection method of a multi-color group gravure printing machine according to claim 1, characterized in that: The step A1 comprises: Use a sliding window of length l to divide the time series into equal lengths.

4. The color mark rapid detection method of a multi-color group gravure printing machine according to claim 3, characterized in that: The time series expression is: Y=(y(t1),y(t2),y(t3)…y(t n )) Among them, Y represents a time series with a length of n, y(t i ) is t i Data recorded at time, t i is the collection time, i=1,2,…,n.

5. The color mark rapid detection method of a multi-color group gravure printing machine according to claim 1, characterized in that: In step A22, the slope is: Among them, k i represents the slope of the i-th data; Y j (i) the i-th data of the j-th subsequence; Y j (i-1) represents the i-1th data of the jth subsequence; t i is the collection time of the ith data; t i-1 Indicates the collection time of the i-1th data.

6. The color mark rapid detection method of a multi-color group gravure printing machine according to claim 1, characterized in that: In step A24, the confidence interval distance radius is: Among them, d j is the confidence interval distance radius of the jth subsequence; is the upper confidence limit, and θ j is the lower confidence limit, and Z is a random variable that satisfies the normal distribution of N(0,1), a is the confidence level, and a=0.05 is taken. represents the mean of the slope of the jth subsequence; σ j Expressed as the mean square error of the slope of the jth subsequence.

7. The color mark rapid detection method of a multi-color group gravure printing machine according to claim 1, characterized in that: The step A3 comprises: A31: cluster all signals of each color group in a multi-color gravure printing machine to obtain a cluster sample set; A32: performing a classification operation on the cluster sample set to obtain a plurality of classified cluster samples; A33: Calculate the cluster center of the target cluster sample; A34: obtaining a membership matrix of the target cluster samples according to the cluster centers of the target cluster samples; A35: Determine whether the absolute value of the difference between the membership matrix of the target cluster sample and the membership matrix of the previous cluster sample is less than the allowable error. If so, output the target cluster sample as the color scale value; otherwise, add 1 to the number of iterations and return to step A33.

8. The color mark rapid detection method of a multi-color group gravure printing machine according to claim 7, characterized in that: In step A33, the cluster center of the target cluster sample is calculated as: in, is the cluster center of the i-th cluster; represents the probability that the kth sample data in the membership matrix U of the previous cluster sample belongs to the i-th cluster, and is the cluster sample x k To cluster center v i distance; m represents the weighted index; x k represents the clustering sample of the kth sample data; C is a constant and represents the total number of sample data; h is the number of iterations.

Images