Search

CN-122023390-A - Digital ink-jet printing quality analysis method and system based on image processing

CN122023390ACN 122023390 ACN122023390 ACN 122023390ACN-122023390-A

Abstract

The invention discloses a digital ink-jet printing quality analysis method and system based on image processing, and belongs to the technical field of image processing. The method comprises the steps of obtaining an image sequence, recording pixel coordinates, forming a motion track of feature points, calculating motion parameters, marking a region where the feature points with the deviation degree exceeding a threshold value are located as a local deformation region, generating a local coordinate mapping from global pixel coordinates to a local dynamic reference system, obtaining a contour pixel point set of a target object to be measured, and calculating the geometric parameters of the target object to be measured to serve as a quality evaluation measurement result to be output. The invention solves the problems of image space distortion caused by production line movement and mechanical vibration in the prior art, further distortion of measured data and error process adjustment caused by the fact that a local dynamic reference system which synchronously changes with the physical structure of the fabric is constructed, and contour pixel points of a target to be measured are converted into calculation geometric parameters under the reference system.

Inventors

  • Xie Wuyu

Assignees

  • 绍兴柯桥赛邦纺织有限公司

Dates

Publication Date
20260512
Application Date
20260403

Claims (9)

  1. 1. The digital ink-jet printing quality analysis method based on image processing is characterized by comprising the following steps of: acquiring an image sequence obtained by continuously shooting a printed fabric to be detected, wherein two adjacent frames of images in the image sequence have overlapping areas; Extracting characteristic points of each frame of image in the image sequence, and recording pixel coordinates of all the characteristic points in the current frame of image; Performing cross-frame matching on the characteristic points in the continuous multi-frame images to form a motion track of the characteristic points; Carrying out statistical analysis on the motion trail of all the characteristic points, and calculating the main trend direction and the main trend speed of the motion trail to form an integral motion parameter; Calculating the deviation degree of the motion trail of each feature point and the overall motion parameter, and marking the region where the feature point with the deviation degree exceeding the threshold value is located as a local deformation region; Selecting a datum point in the local deformation area to establish a local dynamic reference system, and generating a local coordinate mapping from the global pixel coordinate to the local dynamic reference system; Determining an image area in which a target object to be detected is positioned, and acquiring a contour pixel point set of the target object to be detected when the image area belongs to a local deformation area; And converting the contour pixel point set into relative coordinates under a local dynamic reference system by utilizing local coordinate mapping, and calculating the geometric parameters of the target object to be measured under the local dynamic reference system to be used as a quality evaluation measurement result to be output.
  2. 2. The image processing-based digital inkjet printing quality analysis method according to claim 1, wherein the step of extracting feature points of each frame of image in the image sequence and recording pixel coordinates of all feature points in the current frame of image comprises: Performing traversal processing on each frame image in the image sequence, performing pixel level gray gradient calculation operation on the currently traversed frame image, obtaining pixel points with severe local gray change in the current frame image, and marking the pixel points as first type feature points; Performing edge intersection detection operation on the currently traversed frame image, positioning the boundary position of the edge of the printing pattern and marking the boundary position as a second type of characteristic point; Summarizing all the first type characteristic points and the second type characteristic points extracted from the currently traversed frame image to be used as a characteristic point set of the current frame image; for each feature point in the feature point set, establishing a two-dimensional coordinate tuple containing horizontal pixel coordinate values and longitudinal pixel coordinate values; Taking the two-dimensional coordinate tuple as pixel coordinates of the characteristic points in the current frame image and recording the pixel coordinates into a preset coordinate storage matrix; And maintaining the mapping relation between the characteristic points in each frame of image and the corresponding pixel coordinates through the coordinate storage matrix until the characteristic point extraction and pixel coordinate recording operation of all frames of images in the image sequence is completed.
  3. 3. The method for analyzing the quality of digital inkjet printing based on image processing according to claim 1, wherein the step of performing cross-frame matching on feature points in continuous multi-frame images to form a motion track of the feature points comprises the steps of: Extracting respective characteristic point sets of two adjacent frames of images on the time sequence by taking the time sequence of the image sequence as a reference, and intercepting a local pixel region with a preset size around each characteristic point in the previous frame of image as a reference template; Carrying out gray matrix similarity calculation on the reference template and candidate pixel areas with corresponding sizes around the feature points in the image of the next frame, and selecting the feature point with the highest gray matrix similarity, where the center of the candidate pixel area is located, as a homonymous feature point matched across frames; Distributing a time stamp to each frame of image according to the frame rate period of the image sequence, and serially associating pixel coordinate positions in frames with the same name feature points continuously appearing on the time sequence according to the time sequence; Extracting the pixel coordinates with the time stamp corresponding to the same characteristic point in the continuous multi-frame images, and constructing a motion state sequence taking a time variable as an independent variable and the pixel coordinates as the dependent variable; And saving the motion state sequence as the motion trail of the feature point on the time axis to a trail set, and completing the motion trail construction operation of all the cross-frame matching feature points.
  4. 4. The image processing-based digital inkjet printing quality analysis method according to claim 1, wherein the step of performing statistical analysis on the motion trajectories of all the feature points and calculating the overall motion parameters composed of the main trend direction and the main trend speed of the motion trajectories includes: Extracting motion trajectories of all feature points from a trajectory set, and calculating a space displacement vector between adjacent time stamp nodes according to the contained serialized pixel coordinates with the time stamps for each motion trajectory; carrying out direction angle solving on each space displacement vector to obtain an instantaneous direction sequence of the track, and dividing the modular length of the space displacement vector by the time difference value of the adjacent time stamps to obtain an instantaneous speed sequence of the track; carrying out direction distribution histogram statistics on the instantaneous direction sequence of the motion trail of all the characteristic points in the same time span in the image sequence, and selecting a direction central value corresponding to a distribution interval with the highest frequency as a main trend direction; Carrying out speed distribution histogram statistics on an instantaneous speed sequence of motion tracks of all feature points in the same time span in an image sequence, and selecting a speed central value corresponding to a distribution interval with the highest frequency as a main trend speed; And carrying out structured packaging combination on the obtained main trend direction and main trend speed to form an integral motion parameter representing the macroscopic motion state of the printed fabric in the image acquisition process.
  5. 5. The method for analyzing the quality of digital inkjet printing based on image processing according to claim 1, wherein the step of calculating the degree of deviation between the motion trajectory of each feature point and the overall motion parameter and marking the region where the feature point whose degree of deviation exceeds a threshold value is a local deformation region includes: extracting a main trend direction and a main trend speed in the overall motion parameters, and calling a motion track of each feature point one by one; calculating an included angle difference value between an instantaneous direction sequence average value of each characteristic point motion track and a main trend direction as an angle deviation, and calculating a scalar difference absolute value between an instantaneous speed sequence average value of each characteristic point motion track and a main trend speed as a speed deviation; judging whether the angle deviation exceeds a preset angle deviation threshold value or not and judging whether the speed deviation exceeds a preset speed deviation threshold value or not; When the angle deviation of a certain feature point exceeds a preset angle deviation threshold value or the speed deviation exceeds a preset speed deviation threshold value, marking the feature point and a pixel neighborhood with a set radius around the feature point as an abnormal candidate area; Calculating Euclidean distance between the central coordinates of each abnormal candidate region, and if the Euclidean distance between the central coordinates of a plurality of abnormal candidate regions is smaller than a preset communication distance, performing convex hull calculation on boundary points of the abnormal candidate regions; and merging and determining continuous ranges covered by the convex hull polygons generated by the convex hull calculation as local deformation areas.
  6. 6. The image processing-based digital inkjet printing quality analysis method according to claim 1, wherein selecting fiducial points within the local deformation region establishes a local dynamic reference frame, and generating a local coordinate map from global pixel coordinates to the local dynamic reference frame includes: Extracting pixel coordinates of all feature points positioned in each determined local deformation area; three non-collinear feature points are selected from the pixel coordinates to serve as reference points, and global pixel coordinates of the three non-collinear reference points in the current frame image are extracted; Taking a first datum point of the three selected datum points as a local origin, taking the direction of a connecting line of the first datum point pointing to a second datum point as the direction of a local first coordinate axis, and taking the direction of a projection of the connecting line of the first datum point pointing to a third datum point in the direction perpendicular to the direction of the local first coordinate axis as the direction of a local second coordinate axis to construct a mutually orthogonal local dynamic reference system; Constructing a coordinate conversion equation set based on the global coordinate value of the selected datum point under the global pixel coordinate system and the local initial coordinate value under the local dynamic reference system, and solving the coordinate conversion equation set to obtain a space transformation matrix parameter; And establishing a local coordinate mapping matrix from the global pixel coordinate system to the local dynamic reference system through the space transformation matrix parameters and carrying out storage calling.
  7. 7. The method for analyzing digital inkjet printing quality based on image processing according to claim 1, wherein the step of determining an image area in which the target object to be measured is located and acquiring the set of contour pixel points of the target object to be measured when the image area belongs to the local deformation area comprises: Receiving a measurement task instruction issued for a line or ink point to be analyzed in an image, and acquiring initial position coordinates of a target object to be measured according to the measurement task instruction; expanding the bounding box outwards based on the initial position coordinates to generate an image area containing the complete edge of the target object to be detected as a target analysis box; Extracting four vertex coordinates of a target analysis frame, and judging the spatial position relationship between the four vertex coordinates and convex hull polygon vertex coordinates of all local deformation areas in the current frame image; When the fact that four vertex coordinates of the target analysis frame are completely contained in any local deformation area is detected, judging that the image area belongs to the local deformation area; And carrying out morphological edge detection operation on pixels in the target analysis frame, removing background pixel points, reserving pixel points with gray gradient values reaching set boundary conditions, and carrying out sequential splicing operation on reserved edge pixel point coordinates to obtain a contour pixel point set describing the closed or open boundary of the target object to be detected.
  8. 8. The image processing-based digital inkjet printing quality analysis method according to claim 1, wherein the step of calculating the geometric parameters of the target object to be measured under the local dynamic reference system as the quality evaluation measurement result output includes: Invoking a local coordinate mapping matrix corresponding to the local deformation region, and converting the global pixel coordinate of each pixel point in the contour pixel point set into a homogeneous coordinate form; Performing matrix multiplication operation on pixel points in the form of alignment coordinates by using a local coordinate mapping matrix, and solving a transverse relative coordinate value and a longitudinal relative coordinate value of each pixel point under a local dynamic reference system; Reconstructing a relative contour graph of the target object to be detected by using the relative coordinate point set generated after conversion, and calculating a line width index and an ink point diameter index of the target object to be detected based on Euclidean distance and topological structure between the relative coordinate point sets in a local dynamic reference system; performing numerical extraction on the line width index and the ink point diameter index which are obtained by calculation based on the local dynamic reference system and eliminate the influence of the overall motion; And writing the extracted index value as a geometric parameter truly reflecting the relative geometric relation of the printed pattern, and outputting the geometric parameter as a final quality evaluation measurement result.
  9. 9. The digital ink-jet printing quality analysis system based on image processing is used for realizing the digital ink-jet printing quality analysis method based on image processing according to any one of claims 1-8, and is characterized by comprising a sequence acquisition module, a coordinate recording module, a characteristic point matching module, a parameter calculation module, a deformation area marking module, a coordinate mapping generation module, a pixel point set acquisition module and a result output module; The sequence acquisition module is used for acquiring an image sequence obtained by continuously shooting the printed fabric to be detected, wherein two adjacent frames of images in the image sequence have overlapping areas; The coordinate recording module is used for extracting the characteristic points of each frame of image in the image sequence and recording the pixel coordinates of all the characteristic points in the current frame of image; the characteristic point matching module is used for performing cross-frame matching on characteristic points in continuous multi-frame images to form a motion track of the characteristic points; the parameter calculation module is used for carrying out statistical analysis on the motion trail of all the characteristic points and calculating the overall motion parameter formed by the main trend direction and the main trend speed of the motion trail; the deformation region marking module is used for calculating the deviation degree of the motion trail of each characteristic point and the overall motion parameter, and marking the region where the characteristic point with the deviation degree exceeding the threshold value is located as a local deformation region; the coordinate map generating module is used for selecting a datum point in the local deformation area to establish a local dynamic reference system and generating a local coordinate map from the global pixel coordinate to the local dynamic reference system; The pixel point set acquisition module is used for determining an image area where the target object to be detected is located, and acquiring a contour pixel point set of the target object to be detected when the image area belongs to the local deformation area; The result output module is used for converting the contour pixel point set into relative coordinates under a local dynamic reference system by utilizing local coordinate mapping, and calculating geometrical parameters of the target object to be measured under the local dynamic reference system to be used as a quality evaluation measurement result to be output.

Description

Digital ink-jet printing quality analysis method and system based on image processing Technical Field The invention relates to the technical field of image processing, in particular to a digital ink-jet printing quality analysis method and system based on image processing. Background With the wide application of digital ink-jet printing technology in textile printing industry, higher requirements are put on-line detection and accurate evaluation of the quality of printed products. Unlike traditional analog printing, digital ink jet printing forms patterns by directly spraying ink drops onto the surface of a fabric through a spray head, and quality evaluation of the digital ink jet printing depends on accurate measurement of microscopic geometric parameters such as line width, ink dot diameter, edge roughness and the like. Existing digital inkjet printing quality analysis systems are typically implemented using machine vision-based methods. The method comprises the steps of collecting an image of printed fabric through an industrial camera, extracting outlines of lines or ink points in the image through an edge detection algorithm, calculating geometrical parameters such as line width, ink point roundness and the like based on a pixel coordinate system, and finally comparing the parameters with preset standard values to obtain a quality evaluation result. In the prior art, the printed fabric is usually in a high-speed motion state on a continuous production line, and the acquired image inevitably has motion blur and space geometric distortion due to mechanical vibration of the production line and waving of the fabric. When edge detection is performed based on an image with distortion, the originally clear edge gradient becomes gentle, so that the edge positioning is offset, and microscopic parameters such as the measured line width, the ink dot diameter and the like are systematically deviated. These deviations are independent of the process state of the inkjet printing apparatus itself, but are mixed into the measurement data, so that the measurement result does not truly reflect the head state. When the system automatically triggers process adjustment based on the distorted measurement data, the line widening caused by motion blur may be misjudged as the excessive ink jet amount, so that the ink jet voltage is erroneously reduced, and instead, new printing defects are introduced, so that process parameters are disturbed, and the stability of product quality and the operation efficiency of a production line are affected. Disclosure of Invention The embodiment of the application solves the problems of space distortion of images, further distortion of measured data and error process adjustment caused by production line movement and mechanical vibration in the prior art by providing the digital ink-jet printing quality analysis method and system based on image processing, and realizes that a measuring result truly reflecting the geometrical relationship of a printing pattern is obtained by constructing a local dynamic reference system synchronously changing with the physical structure of a fabric, thereby ensuring the process adjustment accuracy and the product quality stability. The embodiment of the application provides a digital ink-jet printing quality analysis method based on image processing, which comprises the steps of obtaining an image sequence obtained by continuously shooting a printed fabric to be detected, wherein two adjacent frames of images in the image sequence have overlapping areas; Extracting characteristic points of each frame of image in the image sequence, and recording pixel coordinates of all the characteristic points in the current frame of image; Performing cross-frame matching on the characteristic points in the continuous multi-frame images to form a motion track of the characteristic points; Carrying out statistical analysis on the motion trail of all the characteristic points, and calculating the main trend direction and the main trend speed of the motion trail to form an integral motion parameter; Calculating the deviation degree of the motion trail of each feature point and the overall motion parameter, and marking the region where the feature point with the deviation degree exceeding the threshold value is located as a local deformation region; Selecting a datum point in the local deformation area to establish a local dynamic reference system, and generating a local coordinate mapping from the global pixel coordinate to the local dynamic reference system; Determining an image area in which a target object to be detected is positioned, and acquiring a contour pixel point set of the target object to be detected when the image area belongs to a local deformation area; And converting the contour pixel point set into relative coordinates under a local dynamic reference system by utilizing local coordinate mapping, and calculating the geometric parameters of the target object to be