CN106897996A - chromatography error detection method based on machine vision - Google Patents

Abstract

The invention provides a method for detecting overprint errors based on machine vision. The target image is collected through the camera, and the target image contains multiple positioning marks; the area of the positioning mark is calculated. If the difference between the marked area and the standard area is not greater than the specified threshold, template matching is used to extract the center of the positioning mark. According to the offset between the center position and the standard position To detect overprint error; if the difference between the marked area and the standard area is greater than the specified threshold, edge segmentation is used to extract the edge of the positioning mark, and the overprint error is detected according to the offset between the edge position and the standard position. After the error detection is completed, the printer is automatically adjusted. The present invention uses image segmentation and template matching to solve the technical problems of the prior art, which are large in calculation amount, large in error, weak in robustness and unable to automatically adjust according to the error, and is suitable for popularization in the printing machine industry.

Description

Overprint error detection method based on machine vision

technical field

The invention relates to intelligent manufacturing technology in the printing machine industry, in particular to a method for detecting overprint errors based on machine vision.

Background technique

Multi-color printing requires overlapping registration of different color plates and different graphic printing, and printing in a certain order. Therefore, the key to checking the quality of printed products is the overprint accuracy of the printing machine.

The traditional overprint marks mainly include three kinds of cat's eye, single circle and cross mark. This kind of overprint mark is invented for manual detection of overprint error, so it is seriously affected by human subjective information and the efficiency of manual judgment is low and the accuracy is poor. It is difficult to apply to the current printing situation. . With the development of machine vision and digital image processing technology, it has become a trend to use digital image processing technology to detect overprint errors.

Some existing overprint error detection methods based on machine vision, for example, for the reticle image, there is a method of determining the center coordinates of each primary color through matching technology, a method of determining the registration error from the deviation of the coordinates, and calculating the template image based on the difference shadow method However, once the primary colors overlap, the corresponding primary color centers cannot be found, indicating that the four crosshair images are not suitable for machine vision overprint error detection. The overprint error detection of the "米" font icon relies on the head and tail positions and inclination angles of different color lines to calculate the overprint error. However, pixels may be lost during the printing process, and the calculation of the ending position will be seriously affected. It is difficult to guarantee robustness and calculation. precision. The primary color dot icon measures the distance between the centers of the color dots and compares it with the set standard distance to represent the registration accuracy. However, it is only suitable for printing machines with fewer colors and cannot solve the problem of mark overlap.

Contents of the invention

The invention provides an overprint error detection method based on machine vision to solve the technical problem that the existing technology has a large amount of calculation and cannot solve the technical problem of large detection errors caused by overlapping and missing positioning marks, so that the printing machine can automatically calibrate the overprint error.

The invention provides a method for detecting overprint errors based on machine vision, comprising the following steps:

s1 collects target detection images. After a printing unit has passed through all printing steps, the camera automatically captures a fixed area on the printing unit that contains all positioning marks.

s2 establishes the xoy coordinate system according to the left and bottom edge directions of the collected target image and the specified origin position, and calculates its deflection angle relative to the xoy coordinate system through the edge of the black positioning mark. By translation and rotation, the center point of the mark is placed at the origin of the xoy coordinate system, so that the edge of the mark coincides completely with the coordinate system, and there is no deflection angle;

s3 After positioning the black, use the xoy coordinate system as a standard to calculate the sum of the area of the other color positioning marks except printing machine black, if the difference between the sum of the standard area of the mark and the calculation result is less than or equal to the specified threshold, then execute s4 , if the difference between the sum of the marked standard area and the calculation result is greater than the specified threshold, execute s5;

s4 performs template matching on the acquisition target image. Find out the upward deviation and downward deviation between the center coordinates of the positioning mark and the standard central position, determine the corresponding printing machine color according to the central position, and determine the adjustment direction and adjustment amount of the horizontal and vertical calibration device where the corresponding color is located according to the offset;

s5 performs edge extraction on the acquisition target image. According to the edge coordinate value and center position of the positioning mark corresponding to each color of the printing machine, the registration error is analyzed and calculated, and the adjustment direction and adjustment amount of the horizontal and vertical calibration device corresponding to the color of the printing machine are determined according to the offset.

After adopting the above technical solution, the technical solution of the present invention has the following beneficial effects:

Using machine vision technology, the overprint error detection system installed on the industrial computer analyzes the target image, obtains the error caused by overprint, returns the error to the calibration device, and the calibration device adjusts the error. The printing machine can accurately adjust the next step according to the offset of the corresponding color mark. The overprint position of the printing unit realizes automatic calibration of printing overprint errors. The method uses the cross to represent the black plate of the printing machine, and the other positioning marks of different shapes represent different printing machines. The color can be applied to various printing machines, which not only gets rid of the influence caused by the color, but also reduces the image preprocessing process to save time, and can use black and white Camera captures images to save cost. In the case of no overlap, using template matching can overcome the problems of irregular edges and missing edges caused by various factors in the actual printing process, and can quickly and effectively locate the center of the mark and calculate the overprint error. In the case of overlap, image edge segmentation is used to calculate the registration error, which solves the problem that the image overlap cannot locate the center, and the calculation method is simple and does not need to consume too much time.

Preferably, different printing machine colors correspond to different shapes of positioning marks, the width and height of the positioning marks are consistent, the distance between the marks is half of the mark width, and the center of the mark can also be kept at the same horizontal line when the bottom edge of the mark is on the same horizontal line. same level. Press black registration marks in the shape of a "ten".

Preferably, the threshold value of the difference between the sum of the standard areas of the positioning marks except printing machine black and the calculation result is specified as 1/20-1/30 of the sum of the areas. Before overprinting, it is necessary to design the size of the positioning mark in advance, and set the threshold according to the size. If the threshold is too large, it will interfere with further image analysis, which is not conducive to the accurate identification of the positioning mark.

Preferably, overlapping edges are not considered after edge extraction, the upper and lower edges of the color block are determined simply, and the places where the gray scale changes greatly are taken out, and the position of the overlapping edges is estimated according to the non-overlapping edges and the width and height of the shape. When calculating the lateral error of two positioning marks, only the edges of these two marks need to be separated.

Preferably, the size of the matching template is consistent with the size of the image captured by the camera, including all colors corresponding to positioning marks and marking standard positions.

Further, the upward and downward deviations are calculated using pixel values, and after calculation, the pixel values are converted into corresponding distance values according to the resolution of the CCD camera used to obtain the actual overprint error.

Description of drawings

In order to more clearly describe the technical methods in the embodiments of the present invention, the following will briefly describe the drawings used in the implementation manners.

Fig. 1 is the flow chart of Embodiment 1 of the overprint error detection method of the present invention

Fig. 2 is a schematic diagram of positioning marks suitable for the overprint error detection and positioning method of the present invention

Fig. 3 is the structural diagram of Embodiment 2 of the overprint error detection method of the present invention

Wherein, reference sign is:

w1, positioning mark width; h1, positioning mark height; w2, positioning mark spacing; a, printing unit; b, positioning mark location; c, target image acquisition range; 1, transmission line; 2, transmission line; 3, black and white linear array CCD camera; 4. LED light source; 5. Four-color printing machine; 6. Horizontal calibration device; 7. Vertical calibration device; 8. Industrial computer; 9. Board card.

detailed description

The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

Example 1

As shown in Figure 1, this embodiment provides a method for detecting overprint errors based on machine vision, which specifically includes the following steps:

s1 collects target detection images. After a printing unit has passed through all printing steps, the camera automatically captures a fixed area on the printing unit that contains all positioning marks.

s2 establishes the xoy coordinate system according to the left and bottom edge directions of the collected target image and the specified origin position, and calculates its deflection angle relative to the xoy coordinate system through the edge of the black positioning mark. By translation and rotation, the center point of the mark is placed at the origin of the xoy coordinate system, so that the edge of the mark coincides completely with the coordinate system, and there is no deflection angle;

Specifically, taking a four-color printing machine as an example, FIG. 2 is a schematic diagram of an overprint error detection positioning mark according to the present invention. The printing press's black K-plate is marked in the shape of a "ten". The cyan, magenta, and yellow colors of the printing machine are respectively squares, circles, and hexagons with the same width w1 and height h1 as the "ten", and their centers are on the same horizontal line under standard conditions, with a distance of w2. The coordinates of the center of the black mark should be placed at the origin (0,0), and the direction of the mark is the direction of the left and bottom edges of the image. The angle required for the rotation and translation operations involved in the process is the black printing error.

s3 After positioning the black, use the xoy coordinate system as a standard to calculate the sum of the area of the other color positioning marks except printing machine black, if the difference between the sum of the standard area of the mark and the calculation result is less than or equal to the specified threshold, then execute s4 , if the difference between the sum of the marked standard area and the calculation result is greater than the specified threshold, execute s5;

s4 performs template matching on the acquisition target image. Find out the upward deviation and downward deviation between the center coordinates of the positioning mark and the standard central position, determine the corresponding printing machine color according to the central position, and determine the adjustment direction and adjustment amount of the horizontal and vertical calibration device where the corresponding color is located according to the offset;

Specifically, template matching is to search for the target in the collected target image. It is known that there are four positioning marks corresponding to four colors in the image, and the target has the same size, direction and positioning mark image as the template. The feature template matching algorithm finds the target positioning mark in the graph and determines its center coordinate position. After finding the center positions of the four positioning marks, calculate the horizontal and vertical offsets between the center positions of each color positioning mark and the set standard position, and the obtained pixel value is the overprint error.

s5 performs edge extraction on the acquisition target image. According to the edge coordinate value and center position of the positioning mark corresponding to each color of the printing machine, the registration error is analyzed and calculated, and the adjustment direction and adjustment amount of the horizontal and vertical calibration device corresponding to the color of the printing machine are determined according to the offset.

Specifically, for example, in a four-color printing machine, the four colors black (K), cyan (C), magenta (M), and yellow (Y) correspond to four different positioning marks. If the C and M positioning markers overlap. When calculating the overprint error of C and M, only the edge information of C and M is needed. When calculating the C-color lateral overprint error, separate the edge to get the position of the leftmost edge of the C-color positioning mark, calculate its lateral distance from the origin, and make a difference between this distance and 1 times the standard width of the positioning mark to obtain C-color Horizontal overprint error. When calculating the M-color lateral overprint error, separate the edge to get the position of the rightmost edge of the M-color positioning mark, calculate its lateral distance from the origin, and make a difference between this distance and 3.5 times the standard width of the positioning mark to obtain the M-color Horizontal overprint error. When calculating the C-color vertical overprint error, separate the edge to get the position of the uppermost edge of the C-color positioning mark, calculate its longitudinal distance from the origin, and make a difference between this distance and 0.5 times the standard width of the positioning mark to obtain the C-color vertical direction Overprint error. When calculating the M-color vertical overprint error, separate the edge to get the position of the uppermost edge of the M-color positioning mark, calculate its longitudinal distance from the origin, and make a difference between this distance and 0.5 times the standard width of the positioning mark to obtain the M-color vertical Overprint error. When obtaining the uppermost edge of the corresponding color, judge which color the pixel on this segment and its left and right pixels belong to according to the leftmost or rightmost edge separated when calculating the lateral error.

Preferably, different printing machine colors correspond to different shapes of positioning marks, the width and height of the positioning marks are consistent, the distance between the marks is half of the mark width, and the center of the mark can also be kept at the same horizontal line when the bottom edge of the mark is on the same horizontal line. same level. Press black registration marks in the shape of a "ten".

Preferably, the threshold value of the difference between the sum of the standard areas of the positioning marks except printing machine black and the calculation result is specified as 1/20-1/30 of the sum of the areas. Before overprinting, it is necessary to design the size of the positioning mark in advance, and set the threshold according to the size. If the threshold is too large, it will interfere with further image analysis, which is not conducive to the accurate identification of the positioning mark.

Preferably, the size of the matching template is consistent with the size of the image captured by the camera, including all colors corresponding to positioning marks and marking standard positions.

Further, pixel values are used in the calculation of the upward and downward deviations, and after calculation, the pixel values are converted into distance values according to the resolution of the CCD camera to obtain the actual overprint error.

Example 2

As shown in Figure 3, this embodiment provides a machine vision-based overprint error detection device, which is suitable for automatic overprint error detection and calibration using the overprint error detection method in the above-mentioned embodiment 1, including an industrial computer 8, the industrial computer 8. An overprint error detection system is installed to realize the processing of Embodiment 1;

It also includes a black and white linear array CCD camera 3 for image acquisition, an LED light source 4 for lighting, a four-color printing machine 5, a horizontal calibration device 6 and a vertical calibration device 7;

The horizontal calibration 6 and the vertical calibration device 7 are connected to the industrial computer 8, and the calibration devices 6 and 7 are installed on the printing press 5 to calibrate printing according to the returned overprint error;

Also comprises board 9, described LED light source 4 is connected with industrial computer 8 through board 9, and described CCD camera 3 is connected with board 9, collects data through board 9, and industrial computer 8 sends out data after processing, adjusts The calibration devices 6 and 7, the boards are installed on the industrial computer.

Specifically, the detection device is mainly used for automatic adjustment of overprint error detection, and an image acquisition system and an overprint error detection system are installed in the industrial computer 8 . A series of parameters such as the shutter speed and depth of field of the camera 3 are controlled by the acquisition system to obtain the target image, and the target image is transmitted to the overprint error detection system. After the overprint error detection system obtains the image, it directly calculates the area of the positioning mark according to the pixels and judges whether it overlaps. Different error detection methods are used. After calculating the overprint error, the pixel value is converted into a distance value according to the resolution of the CCD camera 3 to obtain the actual overprint. error. The industrial computer 8 transmits the overprint error to the horizontal overprint calibration device 6 and the vertical overprint 7 calibration device. The calibration device performs calibration, and the printing machine 5 continues to print the next printing unit.

Preferably, the camera 3 is installed on the support of the printing press 5 with the preset horizontal and vertical angles maintained. Specifically, the longitudinal direction of the camera 3 is parallel to the standard paper advance direction, and the camera 3 and the light source 4 are placed on a side where the printed matter has a printed mark. On the side, the camera 3 is held to be able to acquire images at a predetermined angle.

Further, the shutter speed, focal length, aperture, depth of field and vertical distance from the printing machine 5 of the camera 3 are controlled by the industrial computer 8 .

The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them. Those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or modify some of the technical features. Equivalent replacements are carried out; and these modifications or replacements do not mean that the essence of the corresponding technical solutions deviates from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims (1)

1. a kind of chromatography error detection method based on machine vision, it is characterised in that comprise the following steps：

S1 gathers target detection image, in printing element by after all print steps, on camera automatic data collection printing element FX, the region includes all telltale marks；

S2 sets up xoy coordinates according to the collection target image left side with following two edge directions and the origin position for providing System, its deflection angle relative to xoy coordinate systems is gone out by the edge calculations of black telltale mark.By translating and rotating, will The central point of mark is placed in the origin of xoy coordinate systems so that marker edge is completely superposed with coordinate system, does not have deflection angle；

After s3 positions black, using xoy coordinate systems as standard, to other color telltale mark meters in addition to printing machine black Area summation is calculated, if mark standard area is added summation with result of calculation difference less than or equal to the threshold value for providing, is performed S4, if mark standard area is added summation and is more than the threshold value for providing with the difference of result of calculation, performs s5；

S4 carries out template matches to collection target image, and the centre coordinate for finding out telltale mark is inclined with the upper shifting of center for standard position Deviation is differed from and moved down, correspondence printing press color, the transverse and longitudinal according to where side-play amount determines corresponding color are determined according to center The adjustment direction and adjustment amount of calibrating installation；

S5 to collection target image carry out edge extracting, according to the corresponding telltale mark edge coordinate value of each color of printing machine and in Heart position, analytical calculation registration error, the adjustment of the transverse and longitudinal calibrating installation according to where side-play amount determines correspondence printing press color Direction and adjustment amount.