CN107345916B - Plane appearance detection method based on fixed contour - Google Patents

Abstract

The invention discloses a plane appearance detection method based on a fixed contour, which is mainly used for detecting the defects of color difference on the plane appearance or concave-convex defects with 3D shapes on the surface, effectively distinguishing the plane contour aiming at a digital image and then segmenting and analyzing the abnormity in the contour.

Description

Plane appearance detection method based on fixed contour

Technical Field

The invention relates to the field of machine vision detection, in particular to a plane appearance detection method based on a fixed contour.

Background

In industrial automation production, product appearance detection is widely applied, such as mobile phones, flat panels, various electronic products, refrigerators, washing machines and other household appliances. In the past, manual detection is usually adopted, the defects of high labor intensity of detection workers, poor detection stability and the like exist in the manual detection, the consistency of detection results is poor, and the risks of missed detection and false detection also exist.

With the rapid development of microelectronic technology, computer technology and image processing technology, machine vision inspection technology gradually moves from experiment to practical application stage. The automatic detection device has the advantages of high detection efficiency, good detection consistency and the like, and gradually reduces or replaces manpower, so that high automation is realized.

For example, the panel defect detection of a mobile phone or a flat panel is adopted, the sizes of panels of different models are different, the arrangement of holes on the panels is different, and after the panel parameters are generally required to be configured and input into a detection system, the fixed template is used as a template for learning and contrast detection. The input of template parameters in the process is complex, the template needs to be spatially calibrated, and the appearance defect detection can be performed after the position matching.

Disclosure of Invention

The invention aims to solve the technical problem of providing a plane appearance detection method based on a fixed contour, which is based on the contour of the plane, aims to remove the influence of the contour of the plane in the processing process, directly process and analyze the appearance defect inside the surface, simplify the image processing flow, improve the operation efficiency and optimize the processing effect.

The invention is realized by the following technical scheme: a plane appearance detection method based on fixed contours is mainly used for detecting the defects of color difference existing on the appearance of a plane or concave-convex defects with 3D shapes on the surface, aiming at a digital image, a plane contour is effectively distinguished, and then the abnormity in the contour is divided and analyzed, and the specific steps are as follows:

the method comprises the steps of firstly, acquiring an image of a plane appearance, acquiring a plane outline of an object to be detected, carrying out boundary detection aiming at the acquired image of the appearance, identifying the plane outline in the boundary image, including a peripheral outline and an internal hole pattern, carrying out quantitative extraction and storage on the position and the size of the outline, wherein the process belongs to a parameter learning and configuration process;

secondly, collecting an appearance image aiming at an object to be detected, carrying out boundary detection, filtering and gradient operation on the plane appearance image, carrying out binarization processing to obtain a binarized image, and identifying and extracting the contour of the image obtained at this time according to the contour identification obtained in the previous step;

removing the outline from the binary image of the appearance image, performing expansion processing on the identification in the second step and the extracted outline, and then subtracting the outline from the binary image, wherein the specific operation is to perform zeroing operation on the image gray value of an outline expansion area in the binary image;

fourthly, performing cluster analysis of connected regions on the rest abnormal targets in the image, and quantifying position and area information of the abnormal regions;

and fifthly, according to the result of the fourth step, fusing the information of the binary image and the gray information in the corresponding original image, performing area analysis on the appearance abnormity, and determining the type of the abnormal defect and the severity of the divided defect.

The invention has the beneficial effects that: the invention is based on the self-profile, removes the influence of the self-profile in the processing process, directly processes and analyzes the appearance defect inside the surface, simplifies the image processing flow and improves the operation efficiency and the processing effect.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

FIG. 1 is a schematic diagram of a gray scale image of a panel according to the present invention;

FIG. 2 is a schematic diagram of the contour identification and extraction image of the present invention;

FIG. 3 is a schematic diagram of a binarized image of a boundary of a planar appearance image according to the present invention;

FIG. 4 is a diagram illustrating the image effect of the invention after the boundary of the flat appearance image is expanded;

FIG. 5 is a schematic diagram illustrating the effect of the invention after the outline of the flat appearance image is removed;

FIG. 6 is a flow chart of image processing according to the present invention.

Detailed Description

All of the features disclosed in this specification, or all of the steps in any method or process so disclosed, may be combined in any combination, except combinations of features and/or steps that are mutually exclusive.

Any feature disclosed in this specification (including any accompanying claims, abstract and drawings), may be replaced by alternative features serving equivalent or similar purposes, unless expressly stated otherwise. That is, unless expressly stated otherwise, each feature is only an example of a generic series of equivalent or similar features.

As shown in fig. 1, the panel finishing profile is a rectangle with rounded corners, and includes a circular hole and a long hole inside as shown in fig. 1.

Performing boundary extraction on the image of fig. 1 to obtain a contour boundary, as shown in fig. 2, respectively identifying and extracting contour shapes in fig. 2: the outline 1 is a rectangle with round corners; 2-bit strip hole pattern on the outline; the outline 3 is a round hole; the areas of the enveloping rectangles are respectively A1, A2 and A3, and the central positions are respectively P1, P2 and P3.

During defect detection, an image is shot as shown in figure 1, an abnormal area A exists in a panel, image preprocessing is carried out, a gradient map of the image is obtained and binaryzation is carried out, and as shown in figure 3, position information and area information of a known contour are obtained;

the outline 1, the outline 2 and the outline 3 with approximate areas and positions are found in the figure 3, the image is subjected to expansion processing, namely the inside and the outside of the outline boundary are enlarged, as shown in the figure 4, the area which is not in the background color in the figure 4 is taken as the background on the figure 3, and as a result, as shown in the figure 5, the image is equivalent to only defects and has no appearance outline.

The connected region analysis is performed on the attached drawing 5, that is, the position PA and the area AA of the abnormal region a can be obtained, the region can be mapped back to the attached drawing 1, and the type and the severity of the defect can be divided according to the richer gray information in the attached drawing 1.

The method is based on the self-profile, the influence of the self-profile is removed in the processing process, the appearance defects in the surface are directly processed and analyzed, the image processing flow is simplified, and the operation efficiency and the processing effect are improved. Along with the increase of the demand of the quantity of electronic products and the improvement of the demand of the appearance detection precision, the method has better flexibility in the aspects of image processing flow and effect, and therefore, the method has wide application prospect.

As shown in fig. 6, (1) acquiring an image of a planar appearance, acquiring a planar contour of a detected object, performing boundary detection on the acquired image of the planar appearance, identifying the planar contour in the boundary image, including a peripheral contour, an internal hole pattern and the like, and quantitatively extracting and storing the position and size of the contour, wherein the process belongs to a parameter learning and configuration process;

(2) acquiring an appearance image aiming at an object to be detected, carrying out operations such as boundary detection, filtering, gradient and the like on the plane appearance image, carrying out binarization processing to obtain a binarized image, and identifying and extracting the contour of the image obtained at this time according to the contour identification obtained in the last step;

(3) removing the outline from the binary image of the appearance image, performing expansion processing on the second step identification and the extracted outline, and then subtracting the outline from the binary image, wherein the specific operation is to perform zeroing operation on the image gray value of the outline expansion area in the binary image;

(4) performing cluster analysis of connected regions on the rest abnormal targets in the image, and quantifying information such as positions, areas and the like of the abnormal regions;

(5) and according to the result of the fourth step, fusing the information of the binary image and the gray information in the corresponding original image, performing area analysis on the appearance abnormality, and determining the type of the abnormal defect and the severity of the divided defect.

The invention has the beneficial effects that: the invention is based on the self-profile, removes the influence of the self-profile in the processing process, directly processes and analyzes the appearance defect inside the surface, simplifies the image processing flow and improves the operation efficiency and the processing effect.

The method is based on the self-contour, the influence of the self-contour is removed in the processing process, the appearance defects in the surface are directly processed and analyzed, the image processing flow is simplified, and the operation efficiency and the processing effect are improved. Along with the increase of the demand of the quantity of electronic products and the improvement of the demand of the appearance detection precision, the method has better flexibility in the aspects of image processing flow and effect, and therefore, the method has wide application prospect.

The above description is only an embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that are not thought of through the inventive work should be included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope defined by the claims.

Claims (1)

1. A plane appearance detection method based on a fixed contour is characterized in that: the plane appearance defect detection mainly detects the defects of color difference or 3D-shaped concave-convex defects on the surface of a plane appearance, effectively distinguishes plane outlines aiming at a digital image, and then divides and analyzes the abnormity in the outlines, and the method comprises the following specific steps:

the method comprises the steps of firstly, acquiring an image of a plane appearance, acquiring a plane outline of an object to be detected, carrying out boundary detection aiming at the acquired image of the appearance, identifying the plane outline in the boundary image, including a peripheral outline and an internal hole pattern, carrying out quantitative extraction and storage on the position and the size of the outline, wherein the process belongs to a parameter learning and configuration process;

secondly, collecting an appearance image aiming at an object to be detected, carrying out boundary detection, filtering and gradient operation on the plane appearance image, carrying out binarization processing to obtain a binarized image, and identifying and extracting the contour of the image obtained at this time according to the contour identification obtained in the previous step;

removing the outline from the binary image of the appearance image, performing expansion processing on the identification in the second step and the extracted outline, and then subtracting the outline from the binary image, wherein the specific operation is to perform zeroing operation on the image gray value of an outline expansion area in the binary image;

fourthly, performing cluster analysis of connected regions on the rest abnormal targets in the image, and quantifying position and area information of the abnormal regions;

and fifthly, according to the result of the fourth step, fusing the information of the binary image and the gray information in the corresponding original image, performing area analysis on the appearance abnormity, and determining the type of the abnormal defect and the severity of the divided defect.

Images