A method for judging image morphology quality in industrial visual inspection

Abstract

The present invention discloses a method for determining the image morphology quality of industrial visual inspection. The method divides the detection field of view of an imaging acquisition lens into multiple regions, uses standard-sized graphics as reference calibration samples, places the samples in the divided different regions in turn, and acquires corresponding true-color images. The morphological index values ​​of the graphics and the standard-sized graphics in the true-color image corresponding to each region are calculated respectively, and by analyzing the correlation between the two, it is verified whether the image acquisition effect meets the image morphology standard. After the verification is passed, the coefficient of variation of the morphological index values ​​of the standard-sized graphics in different position regions in the true-color image is calculated, and the coefficient of variation is used to determine the quality of online imaging morphology. The method has the advantages of being simple, effective, and having a fast detection speed.

Description

Technical Field

[0001] The present invention relates to the technical field of computer machine vision, and in particular to an image morphology quality determination method for industrial vision inspection. Background Art

[0002] In the field of industrial manufacturing, computer vision plays a role in object detection and analysis in intelligent manufacturing industrial inspection. However, in actual application, due to the complexity and variability of the production environment itself and the sensitivity of computer vision technology to light, the stability of the image quality cannot be effectively guaranteed, which in turn misleads the actual production. Summary of the invention

[0003] The present invention aims to solve the problem that the quality of image acquisition on intelligent manufacturing production lines cannot be effectively guaranteed. The purpose of the present invention is to propose an image morphology quality determination method for industrial environment visual inspection, which can quickly and effectively determine the online imaging morphology quality detection, ensure the stability of online acquired images, and thus improve the quality of product production.

[0004] The specific technical solutions adopted by the present invention are as follows:

[0005] The present invention provides an image morphology quality determination method for industrial visual inspection, which has the following characteristics and comprises the following steps: step 1, selecting a standard size graphic as a reference calibration sample; step 2, dividing the detection field of view of an image acquisition lens into multiple regions to obtain different regions of the detection field of view; step 3, placing the reference calibration sample in each divided region respectively, and acquiring corresponding true color images; step 4, respectively calculating the morphology index value of the calibration sample in each true color image, and performing correlation analysis with the actual morphology index value of the standard size graphic, so as to determine whether the image acquisition effect meets the graphic morphology standard; step 5, calculating the standard deviation and the coefficient of variation of the morphology index value of the calibration sample in different regions, and determining whether the image morphology quality meets the online imaging acquisition requirement according to the calculated standard deviation and the coefficient of variation results.

[0006] Furthermore, in the image morphology quality determination method for industrial visual inspection provided by the present invention, it can also have the following characteristics: wherein, in step one, the standard size graphics include rectangles, triangles, and circles.

[0007] Furthermore, in the image morphology quality determination method for industrial visual inspection provided by the present invention, it can also have the following characteristics: wherein, the detection field of view in step 2 is rectangular, and the detection field of view is divided into three equal-sized areas of left, middle and right, and numbered i=1, 2, 3.

[0008] Furthermore, the image morphology quality determination method for industrial visual inspection provided by the present invention may also have the following feature: wherein, in step three, each standard size graphic is placed on three areas respectively, and corresponding true color images are acquired.

[0009] Furthermore, in the image morphology quality determination method for industrial visual inspection provided by the present invention, there may also be such a feature: wherein the specific process of step 4 is as follows:

[0010] 1) Calculate the morphological index values ​​of the calibration samples in the true color image respectively:

[0011] Define the actual shape indicator value of the rectangular standard size figure: length is L 0R , width is W 0R , the perimeter is P 0R , area A 0R ;

[0012] Define the actual shape indicator value of the triangle standard size figure: length is L 0T , width is W 0T , the perimeter is P 0T , area A 0T ;

[0013] Define the actual shape indicator value of the standard size circle: length is L 0C , width is W 0C , the perimeter is P 0C , area A 0C ;

[0014] Define the morphological index value of the rectangular calibration sample in the collected true color image: length is L 1R , width is W 1R , the perimeter is P 1R , area A 1R ;

[0015] Define the morphological index value of the triangle calibration sample in the collected true color image: the length is L 1T , width is W 1T , the perimeter is P 1T , area A 1T ;

[0016] Define the morphological index value of the circular calibration sample in the collected true color image: the length is L 1C , width is W 1C , the perimeter is P 1C , area A 1C ;

[0017] 2) Calculate the correlation coefficient r between the actual morphological index values ​​of rectangular, triangular, and circular standard size graphics and the morphological index values ​​of rectangular, triangular, and circular calibration samples in the true color image according to the following formula: R 、r T 、r C :

[0018]

[0019]

[0020]

[0021] In the formula, Cov(*) represents covariance; Var(*) represents variance;

[0022] 3) Determine whether the image acquisition effect meets the graphic morphology standard:

[0023] For the size region image i=1, 2, 3, if r iR 、r iT 、r iC If all are ≥0.99, it indicates that the image acquisition effect meets the graphic morphology standard.

[0024] Furthermore, in the image morphology quality determination method for industrial visual inspection provided by the present invention, there may also be such a feature: wherein the specific process of step five is as follows:

[0025] 1) Calculate the length, width, perimeter and area of ​​the rectangular, triangular and circular calibration samples in three different regions respectively, and obtain the following three corresponding sets of morphological index values:

[0026] (L Ri ,W Ri ,P Ri ,A Ri )、(L Ti ,W Ti ,P Ti ,A Ti ), (L Ci ,W Ci ,P Ci ,A Ci ), i=1, 2, 3;

[0027] 2) Calculate the standard deviation and coefficient of variation of the morphological index values ​​of each calibration sample in three different areas:

[0028] For rectangular calibration samples, the coefficient of variation CV of length, width, perimeter and area LR ,CV WR ,CV PR ,CVAR The calculation formula is as follows:

[0029]

[0030] In formula (4), N = 3; S LR is the standard deviation of the rectangle length values; is the average value of the rectangle length;

[0031]

[0032] In formula (5), N = 3; S WR is the standard deviation of the rectangle width values; is the average value of the rectangle width;

[0033]

[0034] In formula (6), N = 3, S PR is the standard deviation of the perimeter of the rectangle, is the average of the perimeters of the rectangles;

[0035]

[0036] In formula (7), N = 3, S AR is the standard deviation of the perimeter of the rectangle, is the average of the perimeters of the rectangles;

[0037] For the triangular calibration samples and circular calibration samples, the coefficient of variation CV of the length, width, perimeter and area of ​​the triangular calibration samples can be calculated in the same way by replacing the morphological index values ​​of the rectangular calibration samples in the formulas with the morphological index values ​​of the triangular or circular calibration samples through formulas (4) to (7). LT , CV WT , CV PT , C.F. AT , and the coefficient of variation CV of the length, width, circumference, and area of ​​the circular calibration sample LC , CV WC , CV PC , CV AC ;

[0038] 3) Determine whether the image morphology detection meets the requirements of online imaging acquisition:

[0039] The coefficient of variation threshold of each morphological index value was set to 5%. If all calculated coefficients of variation were within 5%, it indicated that the requirements for online imaging acquisition were met.

[0040] Functions and effects of the present invention:

[0041] The image morphology quality determination method for industrial visual inspection of the present invention is to divide the detection field of view of the imaging acquisition lens into multiple regions, use the standard size graphics as the reference calibration samples, place the samples in the divided different regions in turn, and collect the corresponding true color images. In order to test the image morphology quality of the fixed area, the calibration samples of the true color image corresponding to each area and the morphology index values ​​on the standard size graphics are calculated respectively, and the image acquisition effect is verified whether it meets the image morphology standard by analyzing the correlation between the two. In order to verify the influence of different areas in the detection field of view on the image morphology quality, the morphology index values ​​of each graphic sample in the true color graphics of the standard size pattern in different areas are calculated respectively. By calculating the standard deviation and the coefficient of variation of the morphology index values ​​of the same graphic sample in different areas, the coefficient of variation is used to determine the quality of the online imaging morphology. The method of the present invention is simple, effective, and has the characteristics of fast detection speed. It uses machine vision technology to perform non-destructive detection on industrial images, thereby achieving the effect of quickly evaluating the quality of visual detection image morphology in industrial environments, and has broad application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 is a flow chart of an image morphology quality determination method for industrial visual inspection in an embodiment of the present invention;

[0043] Figure 2 Schematic diagram of multi-region division of the detection field of view in an embodiment of the present invention;

[0044] Figure 3 It is a true color image of a standard size graphic in different areas in an embodiment of the present invention. DETAILED DESCRIPTION

[0045] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the following embodiments are combined with the accompanying drawings to specifically illustrate the technical solutions of the present invention.

[0046] <Example>

[0047] See also Figure 1 In this embodiment, a method for determining image morphology quality of industrial visual inspection includes the following steps:

[0048] Step 1: Select a standard size graphic as the reference calibration sample.

[0049] The standard size graphics include but are not limited to rectangular standard size graphics, triangular standard size graphics and circular standard size graphics.

[0050] Step 2: Divide the detection field of view of the image acquisition lens into multiple areas to obtain different areas of the detection field of view. Figure 2As shown, the detection field of view is rectangular, and is divided into three equal-sized areas of left, middle and right, and numbered i=1, 2, 3.

[0051] Step 3: Place rectangular standard-size graphics, triangular standard-size graphics, and circular standard-size graphics in each divided area, and collect corresponding true-color images. Three true-color images are obtained for each standard-size graphic. Figure 3 , Figure 3 The middle group (R) contains three true color images corresponding to the rectangular calibration samples in three regions; Figure 3 The middle group (T) contains three true color images corresponding to the triangle calibration samples in three regions; Figure 3 The middle group of images (C) are three true color images corresponding to the circular calibration samples in three areas.

[0052] The image acquisition process used in this embodiment is implemented through an industrial visual inspection online imaging acquisition platform, which includes a light source, an image acquisition lens, a camera and a computer. The image acquisition lens determines the rectangular detection field of view, the camera is used to acquire true color images, and the computer runs a computer program such as steps one to five.

[0053] Step 4: For each area, calculate the morphological index value of the calibration sample in each true color image, and perform correlation analysis with the actual morphological index value of the standard size graphic to determine whether the image acquisition effect meets the graphic morphological standard.

[0054] The specific process of step 4 is as follows:

[0055] 1) Calculate the morphological index values ​​of the calibration samples in the true color image respectively:

[0056] Define the actual shape indicator value of the rectangular standard size figure: length is L 0R , width is W 0R , the perimeter is P 0R , area A 0R ;

[0057] Define the actual shape indicator value of the triangle standard size figure: length is L 0T , width is W 0T , the perimeter is P 0T , area A 0T ;

[0058] Define the actual shape indicator value of the standard size circle: length is L 0C , width is W 0C , the perimeter is P 0C , area A 0C .

[0059] The actual morphological index values ​​of the standard size graphics in this embodiment are shown in Table 1 below:

[0060] Table 1

[0061]

[0062]

[0063] Define the morphological index value of the rectangular calibration sample in the collected true color image: length is L 1R , width is W 1R , the perimeter is P 1R , area A 1R ;

[0064] Define the morphological index value of the triangle calibration sample in the collected true color image: the length is L 1T , width is W 1T , the perimeter is P 1T , area A 1T ;

[0065] Define the morphological index value of the circular calibration sample in the collected true color image: the length is L 1C , width is W 1C , the perimeter is P 1C , area A 1C .

[0066] The calculated morphological index values ​​of the calibration samples in the true color image in this embodiment are shown in Table 2 below:

[0067] Table 2

[0068]

[0069] 2) Calculate the correlation coefficient r between the actual morphological index value of the rectangular standard size figure and the morphological index value of the rectangular calibration sample in the true color image according to the following formula: R :

[0070]

[0071] The correlation coefficient r between the actual morphological index value of the triangle standard size figure and the morphological index value of the triangle calibration sample in the true color image T :

[0072]

[0073] The correlation coefficient r between the actual morphological index value of the circular standard size figure and the morphological index value of the circular calibration sample in the true color image C :

[0074]

[0075] In formulas (1) to (3), Cov(*) represents covariance; Var(*) represents variance.

[0076] The correlation coefficients of the actual morphological index values ​​of the standard size image template calculated in this embodiment and the morphological index values ​​of the calibration samples in the true color images of different regions are shown in the following Table 3:

[0077] Table 3

[0078]

[0079] 3) Determine whether the image acquisition effect meets the graphic morphology standard:

[0080] For the size region image i=1, 2, 3, if r iR 、r iT 、r iC If all are ≥0.99, it indicates that the image acquisition effect meets the graphic morphology standard.

[0081] As for the image acquisition effect of this embodiment, according to Table 3 above, all the correlation coefficients in the table are greater than 0.99 and close to 1, so it can be determined that this image acquisition effect meets the morphological imaging standard.

[0082] Step 5: Calculate the standard deviation and coefficient of variation of the morphological index values ​​of the calibration samples in different areas, and determine whether the image morphological detection meets the requirements of online imaging acquisition based on the calculated standard deviation and coefficient of variation results.

[0083] The specific process of step five is as follows:

[0084] 1) Calculate the length, width, perimeter and area of ​​the rectangular, triangular and circular calibration samples in three different regions respectively, and obtain the following three corresponding sets of morphological index values:

[0085] (L Ri ,W Ri ,P Ri ,A Ri )、(L Ti ,W Ti ,P Ti ,A Ti ), (L Ci ,W Ci ,P Ci ,A Ci ), i=1, 2, 3;

[0086] 2) Calculate the standard deviation and coefficient of variation of the morphological index values ​​of each calibration sample in three different areas:

[0087] For rectangular calibration samples, the coefficient of variation CV of length, width, perimeter and area LR ,CV WR ,CV PR ,CV AR The calculation formula is as follows:

[0088]

[0089] In formula (4), N = 3; S LR is the standard deviation of the rectangle length values; is the average value of the rectangle length;

[0090]

[0091] In formula (5), N = 3; S WR is the standard deviation of the rectangle width values; is the average value of the rectangle width;

[0092]

[0093] In formula (6), N = 3, S PR is the standard deviation of the perimeter of the rectangle, is the average of the perimeters of the rectangles;

[0094]

[0095] In formula (7), N = 3, S AR is the standard deviation of the perimeter of the rectangle, is the average of the perimeters of the rectangles;

[0096] For the triangular calibration samples and circular calibration samples, the coefficient of variation CV of the length, width, perimeter and area of ​​the triangular calibration samples can be calculated in the same way by replacing the morphological index values ​​of the rectangular calibration samples in the formulas with the morphological index values ​​of the triangular or circular calibration samples through formulas (4) to (7). LT , CV WT , CV PT , CV AT , and the coefficient of variation CV of the length, width, circumference, and area of ​​the circular calibration sample LC , CV WC , CV PC , CV AC .

[0097] In this embodiment, based on the data in Table 2, the coefficient of variation of each morphological index value in different regions of the calibration samples with different shapes can be calculated according to the above steps. The results are shown in Table 4:

[0098] Sample Graphics Length coefficient of variation Width coefficient of variation Coefficient of variation of circumference Area coefficient of variation rectangle 0.003186 0.006426 0.017876 0.011632 triangle 0.005713 0.004733 0.015566 0.012252 Round 0.002435 0.004525 0.015487 0.007841

[0099] 3) Determine whether the image morphology quality meets the requirements of online imaging acquisition:

[0100] The coefficient of variation threshold of each morphological index value was set to 5%. If all calculated coefficients of variation were within 5%, it indicated that the requirements for online imaging acquisition were met.

[0101] For this embodiment, according to Table 4 above, the coefficient of variation of each morphological index value of all standard samples is within 5%, so different regions have no effect on the image morphological acquisition quality, which means that the online imaging acquisition requirements are met.

[0102] The above embodiments are only preferred embodiments of the present invention and are not intended to limit the protection scope of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the protection scope of the present invention.

Claims

1. A method for determining image morphology quality in industrial visual inspection, characterized in that: The following steps are involved: Step 1: Select a standard size graphic as a reference calibration sample, wherein the standard size graphic includes a rectangle, a triangle, and a circle; Step 2: Divide the rectangular detection field of view of the image acquisition lens into three equal-sized areas of left, middle and right, and number them i=1, 2, 3 to obtain different areas of the detection field of view; Step 3: Place the reference calibration samples in each divided area and acquire the corresponding true color image; Step 4: Calculate the morphological index value of the calibration sample in each true color image, and perform correlation analysis with the actual morphological index value of the standard size graphic, so as to determine whether the image acquisition effect meets the graphic morphological standard; The specific process is as follows: 1) Calculate the morphological index values ​​of standard-sized graphics in true color images respectively: Define the actual shape indicator value of the rectangular standard size figure: length is L 0R , width is W 0R , the perimeter is P 0R , area A 0R ; Define the actual shape indicator value of the standard size triangle figure: length is L 0T , width is W 0T , the perimeter is P 0T , area A 0T ; Define the actual shape indicator value of the standard size circular figure: length is L 0C , width is W 0C , the perimeter is P 0C , area A 0C ; Define the morphological index value of the rectangular calibration sample in the collected true color image: length is L 1R , width is W 1R , the perimeter is P 1R , area A 1R ; Define the morphological index value of the triangle calibration sample in the collected true color image: the length is L 1T , width is W 1T , the perimeter is P 1T , area A 1T ; Define the morphological index value of the circular calibration sample in the collected true color image: the length is L 1C , width is W 1C , the perimeter is P 1C , area A 1C ; 2) Calculate the correlation coefficient r between the actual morphological index values ​​of rectangular, triangular, and circular standard size graphics and the morphological index values ​​of rectangular, triangular, and circular calibration samples in the true color image according to the following formula: R 、r T 、r C : In the formula, Cov(*) represents covariance; Var(*) represents variance; 3) Determine whether the image acquisition effect meets the graphic morphology standard: For the size region image i=1, 2, 3, if r iR 、r iT 、r iC If both are ≥ 0.99, it indicates that the image acquisition effect meets the graphic morphology standard; Step 5: Calculate the standard deviation and coefficient of variation of the morphological index values ​​of the calibration samples in different areas, and determine whether the image morphological quality meets the requirements of online imaging acquisition based on the calculated standard deviation and coefficient of variation results; The specific process is as follows: 1) Calculate the length, width, perimeter and area of ​​the rectangular, triangular and circular calibration samples in three different regions respectively, and obtain the following three corresponding sets of morphological index values: (L Ri ,W Ri ,P Ri ,A Ri )、(L Ti ,W Ti ,P Ti ,A Ti ),(L Ci ,W Ci ,P Ci ,A Ci ),i=1,2,3; 2) Calculate the standard deviation and coefficient of variation of the morphological index values ​​of each calibration sample in three different areas: For rectangular calibration samples, the coefficient of variation CV of length, width, perimeter and area LR , CV WR , CV PR , CV AR The calculation formula is as follows: In formula (4), N=3; S LR is the standard deviation of the rectangle length values; is the average value of the rectangle length; In formula (5), N=3; S WR is the standard deviation of the rectangle width values; is the average value of the rectangle width; In formula (6), N = 3, S PR is the standard deviation of the perimeter of the rectangle, is the average of the perimeters of the rectangles; In formula (7), N = 3, S AR is the standard deviation of the perimeter of the rectangle, is the average of the perimeters of the rectangles; For the triangular calibration samples and circular calibration samples, the coefficient of variation CV of the length, width, perimeter and area of ​​the triangular calibration samples can be calculated in the same way by replacing the morphological index values ​​of the rectangular calibration samples in the formulas with the morphological index values ​​of the triangular or circular calibration samples through formulas (4) to (7). LT , CV WT , CV PT , CV AT , and the coefficient of variation CV of the length, width, circumference, and area of ​​the circular calibration sample LC , CV WC , CV PC , CV AC ; 3) Determine whether the image morphology quality meets the requirements of online imaging acquisition: The coefficient of variation threshold is set to 5%. If all calculated coefficients of variation are within 5%, it indicates that the image morphology quality meets the requirements of online imaging acquisition.

2. The image morphology quality determination method for industrial visual inspection according to claim 1, characterized in that: in, In step three, each of the standard-sized graphics is placed on three areas respectively, and corresponding true-color images are acquired.

Images