CN118072319A - Image-text printing quality analysis system based on machine vision - Google Patents

Abstract

The invention relates to the technical field of image analysis, in particular to a graphic printing quality analysis system based on machine vision, which comprises an acquisition unit, a segmentation unit, an image feature library and an analysis unit. The invention divides the print collection image of the graphic printing product by the dividing unit, carries out quality analysis and judgment on the character area and the picture area, is convenient for feedback judgment on the printing quality according to the character characteristics and the picture characteristics, determines whether the printing of the character part has defects according to the self characteristics of the printed characters, simultaneously divides the picture characteristics according to the mode of creating the color difference line according to the color difference value in the standard image, then compares the divided picture characteristics with the similarity of the basic element image in the image characteristic library to determine whether the printing part of the picture has defects, realizes intelligent printing quality judgment on the basis of no printing original edition input, and greatly improves the quality detection efficiency and accuracy aiming at digital graphic printing.

Description

Image-text printing quality analysis system based on machine vision

Technical Field

The invention relates to the technical field of image analysis, in particular to an image-text printing quality analysis system based on machine vision.

Background

The graphic printing is a technique of transferring ink onto the surface of a material such as paper or textile by plate making, inking, pressurizing and the like, and copying the contents of the original in batches, but the printing quality of a printed product is uneven due to different printing modes or printing conditions, so that quality screening of the printed product is required, and most of the printing quality is compared manually, thus not only consuming manpower, but also being easy to cause visual fatigue to damage human body under long-time working.

Chinese patent publication No.: CN110530895A discloses a method and apparatus for detecting print quality; the technical point is that a data point group of a point pattern is extracted from a printed product, information of each nominal position in the point pattern is further extracted, and printing quality detection is realized according to coding rules, so that in the existing printing detection technology, corresponding printing image-text finishing quality detection can be realized only by comparing the coding rules or image-text original plates set in advance, and the printing quality detection system needs to input printing templates in advance, so that the batch printing type has higher printing efficiency, but higher detection efficiency cannot be realized for digital image-text printing with higher relative freedom at present, and the accuracy of image-text printing quality detection is lower.

Disclosure of Invention

Therefore, the invention provides a graphic printing quality analysis system based on machine vision, which is used for solving the problems of poor quality detection efficiency and lower accuracy of digital graphic printing in the prior art.

In order to achieve the above object, the present invention provides a machine vision-based graphic printing quality analysis system, comprising,

The acquisition unit is used for scanning the image-text printing product to form a printing acquisition image of the image-text printing product;

The segmentation unit is connected with the acquisition unit and used for carrying out character recognition on the printing acquisition image and segmenting the printing acquisition image into a character area and a picture area;

the image feature library is internally stored with a plurality of basic element images;

The analysis unit is respectively connected with the segmentation unit and the image feature library and comprises a character analysis module and a picture analysis module;

The character analysis module is used for carrying out calculation and judgment according to the positions of the pixel points corresponding to the maximum value and the minimum value of the color values in the character image, determining that the character image is marked as a differential character or a defective character, and determining whether the character area of the printing collected image is qualified according to the real-time defect ratio of the area of each character image marked as the defective character to the whole area of the printing collected image;

The picture analysis module is used for creating color difference lines on adjacent edges of two adjacent pixel points according to color difference values in the standard graph, extracting contours and colors in a closed area formed by the color difference lines, calculating and judging contour similarity and color similarity with each basic element image in the image feature library to determine whether the closed area is marked as a defect area, and determining whether the picture area of the printing collected image is qualified according to the ratio of the defect area to the whole area of the printing collected image.

Further, the character analysis module is internally provided with a standard character color difference value, the character analysis module can acquire a character image of any character in a character area of the printing acquisition image, extract the color value of each pixel in the character image, select the maximum value and the minimum value of the color value in each pixel of the character image to calculate the difference value, obtain the real-time color difference value of the character image for judgment,

If the real-time color difference value is smaller than or equal to the standard color difference value, the character analysis module marks the character image as a uniform color character, calculates the color average value of the character image, and judges the character image of the next character in the character area of the printing acquisition image;

if the real-time color difference value is larger than the standard color difference value, the text analysis module calculates the deviation distance ratio according to the position of the pixel point corresponding to the maximum value of the color value and the pixel point corresponding to the minimum value of the color value and judges the deviation distance ratio so as to determine that the text image is marked as a poor-color text or a defective text.

Further, when the real-time color difference value of the character image of any character is larger than the standard color difference value, the character analysis module obtains the transverse distance and the longitudinal distance of the character image to calculate the character basic distance, obtains the deviation relative distance between the pixel point corresponding to the maximum value of the color value and the pixel point corresponding to the minimum value of the color value in the character image, and calculates the deviation distance ratio according to the deviation relative distance and the character basic distance;

Where ap=ld/Lu, ap is the offset distance ratio, ld is the offset relative distance, lu is the text base distance.

Further, the text analysis module is internally provided with a preset distance ratio, the text analysis module can judge the deviation distance ratio according to the preset distance ratio,

If the deviation distance ratio is smaller than the preset distance ratio, the character analysis module marks the character image as a defect character;

If the deviation distance ratio is greater than or equal to the preset distance ratio, the character analysis module creates a check line in the character image, and performs division judgment on pixel points on the check line so as to determine that the character image is marked as a differential character or a defective character;

the check line is a connection line of a pixel point corresponding to the maximum value of the color value and a pixel point corresponding to the minimum value of the color value in the text image.

Further, a preset grouping number is arranged in the text analysis module, when the deviation distance ratio is greater than or equal to the preset distance ratio, the text analysis module creates the check line in the text image, acquires the number of all pixel points on the check line, calculates the number of the pixel points in any group according to the preset grouping number, sequentially groups the pixel points to the other end by taking any one end of the check line as a starting point, calculates the average value of the colors of the pixel points in each group, and judges,

If the average value of the colors of the groups of pixel points from the starting point is increased or decreased, the character analysis module marks the character image as a differential color character;

If the average value of the colors of the pixel points in each group from the starting point is not increased or decreased, the text analysis module marks the text image as a defect text.

Further, the text analysis module is internally provided with a preset defect ratio, after finishing marking the text images of all the texts in the text area of the printing collected image, the text analysis module calculates the real-time defect ratio of the area of each text image marked as the defect text to the whole area of the printing collected image,

If the real-time defect ratio is smaller than or equal to the preset defect ratio, the character analysis module judges that the character area of the printing collected image is qualified;

and if the real-time defect ratio is larger than the preset defect ratio, the character analysis module judges that the character area of the printing collected image is unqualified.

Further, the picture analysis module is internally provided with a standard picture internal color difference value, the picture analysis module can acquire a picture area of the printing collected image, calculate the difference value of the color values of any two adjacent pixel points in the picture area and record the picture color difference value,

If the difference value of the picture color is smaller than or equal to the color difference value in the standard picture, the picture analysis module does not create a color difference line between the current two pixel points;

If the picture color difference is larger than the color difference value in the standard picture, the picture analysis module creates a color difference line on the adjacent edges of the two pixel points.

Further, the picture analysis module judges each color difference line when finishing color difference judgment and creation of color difference lines for all adjacent two pixel points of the picture area of the printing collected image,

If a plurality of color difference lines exist in the picture area of the printing acquired image and do not form a closed area, the picture analysis module adjusts the color difference value in the standard chart according to the number of the color difference lines which do not form the closed area, and increases the color difference value in the standard chart until the created color difference lines can form the closed area;

If all color difference lines in the picture area of the printing collected image can form a plurality of closed areas, the picture analysis module performs feature extraction in any area to determine whether the closed area is marked as a defect area;

wherein, E '=e× [1+ (Qx/4×qs) ], E' is the color difference value in the standard chart after adjustment, E is the color difference value in the standard chart before adjustment, qx is the number of color difference lines which do not form a closed area, and Qs is the number of all pixels of the printing acquisition image picture area;

The closed area is an area formed by connecting a plurality of color difference lines end to end, or an area formed by connecting a plurality of color difference lines end to end with the edges of the picture area of the printing acquired image;

the feature extraction includes contour extraction and color extraction.

Further, the picture analysis module is internally provided with standard shape similarity and standard color similarity, when all color difference lines in the picture area of the printing collected image can form a plurality of closed areas, the picture analysis module can extract the outline shape of any one closed area and calculate the outline shape of the closed area and the outline similarity of each basic element image in the image feature library,

If the outline shape of the closed region and the outline similarity of each basic element image in the image feature library are smaller than the standard shape similarity, the picture analysis module marks the closed region as a defect region;

If the contour shape of the closed region and the contour similarity of any basic element image in the image feature library are both greater than or equal to the standard shape similarity, the picture analysis module marks the basic element image as a similar element image, extracts the color distribution of the closed region, calculates the color similarity of the color distribution of the closed region and each similar element image, and marks the closed region as a defect region when the color similarity of the closed region and each similar element image is smaller than the standard color similarity; when the color similarity between the closed area and any similar element image is larger than or equal to the standard color similarity, the picture analysis module does not mark the closed area and judges the next closed area.

Further, the picture analysis module is also internally provided with a preset defect ratio, when finishing judging marks of all closed areas in the picture area of the printing collected image, the picture analysis module calculates the picture defect ratio of all defect areas of the marks to the whole area of the printing collected image,

If the ratio of the picture defects is smaller than or equal to the preset defect ratio, the picture analysis module judges that the picture area of the printing collected image is qualified;

If the picture defect ratio is larger than the preset defect ratio, the picture analysis module judges that the picture area of the printing collected image is unqualified.

Compared with the prior art, the invention has the advantages that the segmentation unit is used for segmenting the characters and the pictures of the printing collection image of the image-text printing product, the character analysis module and the picture analysis module are respectively arranged for respectively carrying out quality analysis and judgment on the character area and the picture area of the printing collection image, the feedback judgment on the printing quality is convenient according to the character characteristics and the picture characteristics, the printing of the character part is determined to have defects according to the self characteristics of the printing characters, the segmentation of the picture characteristics is carried out according to the mode of creating the color difference line in the standard image, and the similarity comparison with the basic element image in the image characteristic library is used for determining whether the picture printing part has defects, so that the intelligent printing quality judgment is realized on the basis of no printing original edition input, and the quality detection efficiency and the accuracy of digital image-text printing are greatly improved.

Further, by setting the color difference value in the standard character in the character analysis module, the character image of any segmented character can be judged, the color state of the same character in the preparation of graphic printing can be represented by selecting the color value maximum value and the color value minimum value in each pixel of the character image to calculate the difference value, and when the real-time color difference value is smaller than or equal to the standard color difference value, the printing color error of the character is smaller, so that the character is marked as a uniform color character, and the color average value of the character image is calculated so as to be conveniently used as the parameter output of quality detection, and meanwhile, the detection efficiency of the printing quality is improved.

Further, when the real-time color difference value of the character image is larger than the standard color difference value, the deviation distance between the pixel point corresponding to the maximum value of the color value in the character image and the pixel point corresponding to the minimum value of the color value and the deviation distance ratio are obtained, so that the fact that the position of the maximum value of the color value is far away from or far back relative to the position of the minimum value of the color value and the position of the minimum value of the color value are far away from or far back relative to the character is indicated, and the fact that the deviation distance ratio is smaller than the preset distance ratio indicates that the two color maximum values in the character image are relatively close to each other is indicated, therefore, the character printing unevenness can be rapidly determined, the character printing unevenness is marked as a defect character, and the printing defect position is rapidly locked.

In particular, when the color average value of the pixel points in the grouping is determined, if the color average value of each group gradually decreases or increases from one end of the checking line to the other end, and the changing distance, namely the deviation relative distance, is not lower than the requirement of the preset distance ratio, the character is determined to be rendered, so that the character is marked as a poor color character, the defect is defaulted, otherwise, the color change on the checking line is not correspondingly and gradually changed, the part is determined to have a printing defect, so that the character is marked as a defective character, whether the character area of the printing collected image is qualified is determined according to the area ratio of the defective character, and the accuracy of the printing quality detection analysis is further ensured.

Further, by setting the color difference value in the standard chart in the picture analysis module, the area with larger color difference value of two adjacent pixel points can be rapidly determined, and the color difference line is created between the two current pixel points for the larger color difference value of the two adjacent pixel points, so that the integral picture area can be rapidly segmented, the segmented closed areas can be conveniently extracted in characteristics, and the characteristic extraction and comparison result is more accurate.

In particular, a large number of basic element images are stored by setting an image feature library, and basic element images similar to a closed area are searched by adopting two features of outlines and colors, if the basic element images similar to the closed area exist, a certain picture element in the part of picture areas is judged, if the image features of the closed area are not similar to those of the basic element images, the part of picture areas is judged to have printing defects, whether the picture areas of the printing acquisition images are qualified or not is determined by setting a preset defect ratio mode, and the printing quality judgment without an original edition is realized on the basis of ensuring the printing quality detection efficiency of the printing acquisition images, and meanwhile, the accuracy of printing quality analysis detection is improved.

Drawings

FIG. 1 is a schematic diagram of the connection of the machine vision-based graphic printing quality analysis system according to the present embodiment;

FIG. 2 is a schematic diagram showing the connection of the analysis unit according to the present embodiment;

FIG. 3 is a flow chart of the embodiment in which the image analysis module determines that a closed region is formed;

FIG. 4 is a flowchart illustrating the determination of the defect area mark by the picture analysis module according to the present embodiment.

Detailed Description

In order that the objects and advantages of the invention will become more apparent, the invention will be further described with reference to the following examples; it should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.

Preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are merely for explaining the technical principles of the present invention, and are not intended to limit the scope of the present invention.

It should be noted that, in the description of the present invention, terms such as "upper," "lower," "left," "right," "inner," "outer," and the like indicate directions or positional relationships based on the directions or positional relationships shown in the drawings, which are merely for convenience of description, and do not indicate or imply that the apparatus or elements must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present invention.

Furthermore, it should be noted that, in the description of the present invention, unless explicitly specified and limited otherwise, the terms "mounted," "connected," and "connected" are to be construed broadly, and may be either fixedly connected, detachably connected, or integrally connected, for example; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention can be understood by those skilled in the art according to the specific circumstances.

Referring to fig. 1, which is a schematic connection diagram of a machine vision-based image-text print quality analysis system according to the present embodiment, the present embodiment provides a machine vision-based image-text print quality analysis system, which includes,

The acquisition unit is used for scanning the image-text printing product to form a printing acquisition image of the image-text printing product;

The segmentation unit is connected with the acquisition unit and used for carrying out character recognition on the printing acquisition image and segmenting the printing acquisition image into a character area and a picture area;

the image feature library is internally stored with a plurality of basic element images;

The analysis unit is respectively connected with the segmentation unit and the image feature library and comprises a character analysis module and a picture analysis module;

The character analysis module is used for carrying out calculation and judgment according to the positions of the pixel points corresponding to the maximum value and the minimum value of the color values in the character image, determining that the character image is marked as a differential character or a defective character, and determining whether the character area of the printing collected image is qualified according to the real-time defect ratio of the area of each character image marked as the defective character to the whole area of the printing collected image;

The picture analysis module is used for creating color difference lines on adjacent edges of two adjacent pixel points according to color difference values in the standard graph, extracting contours and colors in a closed area formed by the color difference lines, calculating and judging contour similarity and color similarity with each basic element image in the image feature library to determine whether the closed area is marked as a defect area, and determining whether the picture area of the printing collected image is qualified according to the ratio of the defect area to the whole area of the printing collected image.

Referring to fig. 2, 1 is a schematic connection diagram of an analysis unit according to the present embodiment;

According to the embodiment, the segmentation unit is used for segmenting characters and pictures of a printing collection image of a graphic printing product, the character analysis module and the picture analysis module are respectively arranged in the analysis unit and respectively conduct quality analysis and judgment on a character area and a picture area of the printing collection image, feedback judgment on printing quality according to character features and picture features is facilitated, whether defects exist in printing of character parts or not is determined according to self characteristics of the printing characters, segmentation of picture features is conducted according to a mode of creating color difference lines according to color difference values in a standard chart, whether defects exist in a picture printing part or not is determined according to similarity comparison of basic element images in an image feature library, intelligent printing quality judgment is achieved on the basis of no printing original edition input, and quality detection efficiency and accuracy of digital graphic printing are improved greatly.

Specifically, the character analysis module is internally provided with a standard character color difference value, the character analysis module can acquire a character image of any character in a character area of the printing acquisition image, extract the color value of each pixel in the character image, select the maximum value and the minimum value of the color value in each pixel of the character image to calculate the difference value, obtain the real-time color difference value of the character image for judgment,

If the real-time color difference value is smaller than or equal to the standard color difference value, the character analysis module marks the character image as a uniform color character, calculates the color average value of the character image, and judges the character image of the next character in the character area of the printing acquisition image;

if the real-time color difference value is larger than the standard color difference value, the text analysis module calculates the deviation distance ratio according to the position of the pixel point corresponding to the maximum value of the color value and the pixel point corresponding to the minimum value of the color value and judges the deviation distance ratio so as to determine that the text image is marked as a poor-color text or a defective text.

The color difference value in the standard character is set in the character analysis module, the character image of any segmented character can be judged, the color value maximum value and the color value minimum value in each pixel of the character image are selected to calculate the difference value, the color state of the same character in the preparation of the graphic printing can be represented, when the real-time color difference value is smaller than or equal to the standard color difference value, the printing color error of the character is represented to be smaller, the character is marked as a uniform color character, and the color average value of the character image is calculated so as to be conveniently used as the parameter output of quality detection, and meanwhile, the detection efficiency of the printing quality is improved; in this embodiment, the color value set is the total RGB value, that is, the total color value obtained by r+g+b, the color difference value in the standard character set is generally set according to the printing requirement, if the printing font is a special color or an artistic font, the color difference value in the standard character needs to be correspondingly adjusted, and in order to ensure higher quality detection accuracy, the color difference value in the standard character is generally set to not more than 12.

Specifically, when the real-time color difference value of the character image of any character is larger than the standard color difference value, the character analysis module obtains the transverse distance and the longitudinal distance of the character image to calculate the character basic distance, obtains the deviation relative distance between the pixel point corresponding to the maximum value of the color value and the pixel point corresponding to the minimum value of the color value in the character image, and calculates the deviation distance ratio according to the deviation relative distance and the character basic distance;

Where ap=ld/Lu, ap is the offset distance ratio, ld is the offset relative distance, lu is the text base distance.

Specifically, the text analysis module is internally provided with a preset distance ratio, the text analysis module can judge the deviation distance ratio according to the preset distance ratio,

If the deviation distance ratio is smaller than the preset distance ratio, the character analysis module marks the character image as a defect character;

If the deviation distance ratio is greater than or equal to the preset distance ratio, the character analysis module creates a check line in the character image, and performs division judgment on pixel points on the check line so as to determine that the character image is marked as a differential character or a defective character;

the check line is a connection line of a pixel point corresponding to the maximum value of the color value and a pixel point corresponding to the minimum value of the color value in the text image.

When the real-time color difference value of the character image is larger than the standard color difference value, the deviation distance between the pixel point corresponding to the maximum value of the color value in the character image and the pixel point corresponding to the minimum value of the color value and the deviation distance ratio of the character basic distance are obtained to indicate whether the position of the maximum value of the color value and the position of the minimum value of the color value are far away or far back relative to the character, and when the deviation distance ratio is smaller than the preset distance ratio, the two color maximum values in the character image are relatively close, so that the character printing unevenness can be rapidly determined, the character is marked as a defective character, and the printing defect position is rapidly locked.

In this embodiment, the horizontal distance and the vertical distance of the text image are obtained to calculate the text base distance, where a smaller value of the horizontal span distance and the vertical span distance of the text image is used as the text base distance, which can effectively avoid the decision error caused by the gradual change or color change effect caused by font rendering, and the text cannot be rendered independently at a distance less than fifty percent of the text base distance, so the preset distance ratio is generally about 0.5.

Specifically, the text analysis module is internally provided with a preset grouping number, when the deviation distance ratio is greater than or equal to the preset distance ratio, the text analysis module creates the check line in the text image, acquires the number of all pixel points on the check line, calculates the number of the pixel points in any group according to the preset grouping number, sequentially groups the pixel points to the other end by taking any one end of the check line as a starting point, calculates the average value of the colors of the pixel points in each group for judgment,

If the average value of the colors of the groups of pixel points from the starting point is increased or decreased, the character analysis module marks the character image as a differential color character;

If the average value of the colors of the pixel points in each group from the starting point is not increased or decreased, the text analysis module marks the text image as a defect text.

Specifically, the text analysis module is internally provided with a preset defect ratio, after finishing marking the text images of all the texts in the text area of the printing collected image, the text analysis module calculates the real-time defect ratio of the area of each text image marked as the defective text to the whole area of the printing collected image,

If the real-time defect ratio is smaller than or equal to the preset defect ratio, the character analysis module judges that the character area of the printing collected image is qualified;

and if the real-time defect ratio is larger than the preset defect ratio, the character analysis module judges that the character area of the printing collected image is unqualified.

When the color average value of the grouped pixel points is determined, if the color average value of each group gradually decreases or increases from one end of the check line to the other end, and the changing distance, namely the deviation relative distance, is not lower than the requirement of the preset distance ratio, the character is determined to be rendered, so that the character is marked as a poor color character, and is defaulted to be free of defects, otherwise, the color change on the check line is determined to be free of corresponding gradual change rules, the part is determined to be defective in printing, so that the character is marked as a defective character, whether the character area of the printing collected image is qualified is determined according to the area ratio of the defective character, and the accuracy of printing quality detection analysis is further ensured.

The preset defect ratio is set to be the ratio of the determined defect area, and is generally set according to the printing requirement, in this embodiment, the preset defect ratio is set to be 0.015, that is, when the defect area of the text image is lower than 1.5% of the total area, the text image is qualified.

FIG. 3 is a flowchart showing the picture analysis module determining that a closed region is formed;

Specifically, the picture analysis module is internally provided with a standard picture internal color difference value, the picture analysis module can acquire a picture area of the printing collected image, calculate the difference value of the color values of any two adjacent pixel points in the picture area and record the picture color difference value,

If the difference value of the picture color is smaller than or equal to the color difference value in the standard picture, the picture analysis module does not create a color difference line between the current two pixel points;

If the picture color difference is larger than the color difference value in the standard picture, the picture analysis module creates a color difference line on the adjacent edges of the two pixel points.

Specifically, when the picture analysis module completes the color difference value judgment and the creation of color difference lines for all two adjacent pixel points of the picture area of the printing and collecting image, the picture analysis module judges the created color difference lines,

If a plurality of color difference lines exist in the picture area of the printing acquired image and do not form a closed area, the picture analysis module adjusts the color difference value in the standard chart according to the number of the color difference lines which do not form the closed area, and increases the color difference value in the standard chart until the created color difference lines can form the closed area;

If all color difference lines in the picture area of the printing collected image can form a plurality of closed areas, the picture analysis module performs feature extraction in any area to determine whether the closed area is marked as a defect area;

wherein, E '=e× [1+ (Qx/4×qs) ], E' is the color difference value in the standard chart after adjustment, E is the color difference value in the standard chart before adjustment, qx is the number of color difference lines which do not form a closed area, and Qs is the number of all pixels of the printing acquisition image picture area;

The closed area is an area formed by connecting a plurality of color difference lines end to end, or an area formed by connecting a plurality of color difference lines end to end with the edges of the picture area of the printing acquired image;

the feature extraction includes contour extraction and color extraction.

The color difference value in the standard graph can be quickly determined by setting the color difference value in the picture analysis module, the color difference value in the standard graph can be quickly determined for the areas with larger color difference values of the two adjacent pixel points, the color difference value in the standard graph can be quickly segmented by establishing a color difference line between the two current pixel points, the segmented closed areas can be conveniently extracted, the characteristic extraction and comparison result is more accurate, meanwhile, the color difference value in the standard graph is adjusted until the established color difference lines can form the closed areas when a plurality of color difference lines exist in the picture area of the printing acquisition image, the color difference value in the initial standard graph is not required to be excessively set, the self-adaptive mechanism can be utilized for repeatedly adjusting and judging for a plurality of times, the finally adjusted color difference value in the standard graph reaches a state suitable for printing acquisition image, and the color difference value in the initial standard graph is generally set to be the same as the color difference value in the standard graph, and the setting of the color difference value in the initial standard graph is not more than 12 in the embodiment.

FIG. 4 is a flowchart showing the picture analysis module determining the defect area mark according to the present embodiment;

Specifically, the picture analysis module is internally provided with standard shape similarity and standard color similarity, when all color difference lines in the picture area of the printing collected image can form a plurality of closed areas, the picture analysis module can extract the outline shape of any closed area and calculate the outline shape of the closed area and the outline similarity of each basic element image in the image feature library,

If the outline shape of the closed region and the outline similarity of each basic element image in the image feature library are smaller than the standard shape similarity, the picture analysis module marks the closed region as a defect region;

If the contour shape of the closed region and the contour similarity of any basic element image in the image feature library are both greater than or equal to the standard shape similarity, the picture analysis module marks the basic element image as a similar element image, extracts the color distribution of the closed region, calculates the color similarity of the color distribution of the closed region and each similar element image, and marks the closed region as a defect region when the color similarity of the closed region and each similar element image is smaller than the standard color similarity; when the color similarity between the closed area and any similar element image is larger than or equal to the standard color similarity, the picture analysis module does not mark the closed area and judges the next closed area.

Specifically, the picture analysis module is also internally provided with a preset defect ratio, when finishing judging marks of all closed areas in the picture area of the printing collected image, the picture analysis module calculates the picture defect ratio of all defect areas of the marks to the whole area of the printing collected image,

If the ratio of the picture defects is smaller than or equal to the preset defect ratio, the picture analysis module judges that the picture area of the printing collected image is qualified;

If the picture defect ratio is larger than the preset defect ratio, the picture analysis module judges that the picture area of the printing collected image is unqualified.

The method comprises the steps of storing a large number of basic element images through an image feature library, searching for basic element images similar to a closed area by adopting two features of outlines and colors, judging a certain picture element in the picture area of the part if the basic element images similar to the closed area exist, judging that a printing defect exists in the part if the image features of the closed area are not similar to the basic element images, and determining whether the picture area of a printing acquisition image is qualified or not through a mode of setting a preset defect ratio, wherein the preset defect ratio is the area ratio condition, so that the preset defect ratio is the same as that used for judging the character image defect and is 0.015; on the basis of guaranteeing the printing quality detection efficiency of the printing collected images, the original-edition-free printing quality judgment is realized, and meanwhile, the accuracy of printing quality analysis and detection is improved.

Thus far, the technical solution of the present invention has been described in connection with the preferred embodiments shown in the drawings, but it is easily understood by those skilled in the art that the scope of protection of the present invention is not limited to these specific embodiments. Equivalent modifications and substitutions for related technical features may be made by those skilled in the art without departing from the principles of the present invention, and such modifications and substitutions will be within the scope of the present invention.

The foregoing description is only of the preferred embodiments of the invention and is not intended to limit the invention; various modifications and variations of the present invention will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (10)

1. A graphic printing quality analysis system based on machine vision is characterized by comprising,

The acquisition unit is used for scanning the image-text printing product to form a printing acquisition image of the image-text printing product;

The segmentation unit is connected with the acquisition unit and used for carrying out character recognition on the printing acquisition image and segmenting the printing acquisition image into a character area and a picture area;

the image feature library is internally stored with a plurality of basic element images;

The analysis unit is respectively connected with the segmentation unit and the image feature library and comprises a character analysis module and a picture analysis module;

The character analysis module is used for carrying out calculation and judgment according to the positions of the pixel points corresponding to the maximum value and the minimum value of the color values in the character image, determining that the character image is marked as a differential character or a defective character, and determining whether the character area of the printing collected image is qualified according to the real-time defect ratio of the area of each character image marked as the defective character to the whole area of the printing collected image;

The picture analysis module is used for creating color difference lines on adjacent edges of two adjacent pixel points according to color difference values in the standard graph, extracting contours and colors in a closed area formed by the color difference lines, calculating and judging contour similarity and color similarity with each basic element image in the image feature library to determine whether the closed area is marked as a defect area, and determining whether the picture area of the printing collected image is qualified according to the ratio of the defect area to the whole area of the printing collected image.

2. The machine vision-based graphic printing quality analysis system according to claim 1, wherein the character analysis module is provided with a standard character color difference value, and the character analysis module can obtain a character image of any character in a character area of the printing collection image, extract color values of pixels in the character image, select a calculated difference value between a maximum value and a minimum value of the color values in the pixels of the character image, obtain a real-time color difference value of the character image, and determine,

If the real-time color difference value is smaller than or equal to the standard color difference value, the character analysis module marks the character image as a uniform color character, calculates the color average value of the character image, and judges the character image of the next character in the character area of the printing acquisition image;

if the real-time color difference value is larger than the standard color difference value, the text analysis module calculates the deviation distance ratio according to the position of the pixel point corresponding to the maximum value of the color value and the pixel point corresponding to the minimum value of the color value and judges the deviation distance ratio so as to determine that the text image is marked as a poor-color text or a defective text.

3. The machine vision-based graphic printing quality analysis system according to claim 2, wherein the text analysis module obtains a horizontal distance and a vertical distance of a text image to calculate a text base distance when a real-time color difference value of the text image of any text is larger than a standard color difference value, obtains a deviation relative distance between a pixel point corresponding to a maximum value of color values in the text image and a pixel point corresponding to a minimum value of the color values, and calculates a deviation distance ratio according to the deviation relative distance and the text base distance;

Where ap=ld/Lu, ap is the offset distance ratio, ld is the offset relative distance, lu is the text base distance.

4. The machine vision-based graphic printing quality analysis system according to claim 3, wherein the text analysis module is provided with a preset distance ratio, and the text analysis module can determine the deviation distance ratio according to the preset distance ratio,

If the deviation distance ratio is smaller than the preset distance ratio, the character analysis module marks the character image as a defect character;

If the deviation distance ratio is greater than or equal to the preset distance ratio, the character analysis module creates a check line in the character image, and performs division judgment on pixel points on the check line so as to determine that the character image is marked as a differential character or a defective character;

the check line is a connection line of a pixel point corresponding to the maximum value of the color value and a pixel point corresponding to the minimum value of the color value in the text image.

5. The machine vision based graphic printing quality analysis system according to claim 4, wherein a preset grouping number is provided in the text analysis module, the text analysis module creates the check line in the text image and obtains the number of all the pixels on the check line when the deviation distance ratio is greater than or equal to the preset distance ratio, calculates the number of pixels in any group according to the preset grouping number, the text analysis module groups each pixel sequentially with any one end of the check line as a starting point to the other end, calculates the average value of the colors of each group of pixels, and determines,

If the average value of the colors of the groups of pixel points from the starting point is increased or decreased, the character analysis module marks the character image as a differential color character;

If the average value of the colors of the pixel points in each group from the starting point is not increased or decreased, the text analysis module marks the text image as a defect text.

6. The machine vision based graphics context print quality analysis system of claim 2, wherein the text analysis module has a preset defect ratio, and after the text analysis module completes marking the text image of all the text in the text area of the print acquisition image, calculates a real-time defect ratio of each text image area marked as a defective text to the whole area of the print acquisition image,

If the real-time defect ratio is smaller than or equal to the preset defect ratio, the character analysis module judges that the character area of the printing collected image is qualified;

and if the real-time defect ratio is larger than the preset defect ratio, the character analysis module judges that the character area of the printing collected image is unqualified.

7. The machine vision-based graphic printing quality analysis system according to claim 1, wherein the picture analysis module is provided with standard intra-picture color difference values, and the picture analysis module can obtain a picture area of the printing collected image and calculate the difference value of the color values of any two adjacent pixel points in the picture area to record the picture color difference value,

If the difference value of the picture color is smaller than or equal to the color difference value in the standard picture, the picture analysis module does not create a color difference line between the current two pixel points;

If the picture color difference is larger than the color difference value in the standard picture, the picture analysis module creates a color difference line on the adjacent edges of the two pixel points.

8. The machine vision based graphics and text print quality analysis system according to claim 7, wherein the picture analysis module determines each color difference line to be created when the color difference determination and the creation of color difference lines are completed for all adjacent two pixels of the picture area of the print acquisition image,

If a plurality of color difference lines exist in the picture area of the printing acquired image and do not form a closed area, the picture analysis module adjusts the color difference value in the standard chart according to the number of the color difference lines which do not form the closed area, and increases the color difference value in the standard chart until the created color difference lines can form the closed area;

If all color difference lines in the picture area of the printing collected image can form a plurality of closed areas, the picture analysis module performs feature extraction in any area to determine whether the closed area is marked as a defect area;

wherein, E '=e× [1+ (Qx/4×qs) ], E' is the color difference value in the standard chart after adjustment, E is the color difference value in the standard chart before adjustment, qx is the number of color difference lines which do not form a closed area, and Qs is the number of all pixels of the printing acquisition image picture area;

The closed area is an area formed by connecting a plurality of color difference lines end to end, or an area formed by connecting a plurality of color difference lines end to end with the edges of the picture area of the printing acquired image;

the feature extraction includes contour extraction and color extraction.

9. The machine vision based graphic printing quality analysis system according to claim 8, wherein the picture analysis module is provided with standard shape similarity and standard color similarity, and when all color difference lines in the picture area of the printing collected image can form a plurality of closed areas, the picture analysis module can extract the outline shape of any one of the closed areas and calculate the outline shape of the closed area and the outline similarity of each basic element image in the image feature library,

If the outline shape of the closed region and the outline similarity of each basic element image in the image feature library are smaller than the standard shape similarity, the picture analysis module marks the closed region as a defect region;

If the contour shape of the closed region and the contour similarity of any basic element image in the image feature library are both greater than or equal to the standard shape similarity, the picture analysis module marks the basic element image as a similar element image, extracts the color distribution of the closed region, calculates the color similarity of the color distribution of the closed region and each similar element image, and marks the closed region as a defect region when the color similarity of the closed region and each similar element image is smaller than the standard color similarity; when the color similarity between the closed area and any similar element image is larger than or equal to the standard color similarity, the picture analysis module does not mark the closed area and judges the next closed area.

10. The machine vision based graphics context print quality analysis system according to claim 9, wherein the graphics context analysis module is also provided with a preset defect ratio, and when finishing the judgment marking of all the closed areas in the graphics area of the print acquisition image, the graphics context analysis module calculates the ratio of all the defect areas of the mark to the graphics context defect of all the areas of the print acquisition image,

If the ratio of the picture defects is smaller than or equal to the preset defect ratio, the picture analysis module judges that the picture area of the printing collected image is qualified;

If the picture defect ratio is larger than the preset defect ratio, the picture analysis module judges that the picture area of the printing collected image is unqualified.