Pre-filling pipe glue seam visual inspection system and method

Abstract

The invention belongs to the technical field of device detection, and provides a visual detection system and method for a glue line of a pre-filling pipe, and the method comprises the following steps: firstly, carrying out double-image label detection on the pre-filling pipe, selecting an image without a label from the images to carry out glue line width detection, and under the condition that the double images of the pre-filling pipe have the labels, carrying out visual detection on the glue line width of the pre-filling pipe; the method comprises the following steps: analyzing historical glue line detection data, judging whether the existence of a label in a pre-filled pipe image can influence glue line width detection or not, correcting the glue line width in the pre-filled pipe image, processing and outputting based on the corrected glue line width to obtain a target glue line width, and detecting the glue line of the pre-filled pipe according to the target glue line width, so as to detect the glue line of the pre-filled pipe. According to the method, the corrected glue line width is processed and analyzed, the glue line width of the double images of the pre-filled pipe and the label area proportion are combined, the target glue line width is obtained through comprehensive calculation, and the accuracy of the glue line of the pre-filled pipe is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of device detection, in particular to a pre-filled tube glue seam visual detection system and method. Background Art

[0002] There are many deficiencies in the production of prefilled tubes (prefilled syringes). For example, the rotary rod labeling machine on the prefilled tube packaging line may not be in place when rotating the push rod due to the height difference of the product stopper after filling. Products with incomplete rotation of the rod may cause the push rod to fall off during transportation after being boxed. To prevent the push rod from loosening and falling during transportation, the prefilled tube (prefilled syringe) needs to be tested for the fit between the push rod and the stopper during the production and manufacturing process. However, a label is usually affixed to one side of the outer surface of the prefilled tube. During the transmission inspection process, some images blocked by the label will be collected, which will affect the measurement of the gap between the push rod and the stopper, thereby causing misjudgment of the visual inspection of the prefilled tube seam.

[0003] To this end, the present invention provides a pre-filled tube glue seam visual inspection system and method. Summary of the invention

[0004] In order to make up for the deficiencies of the prior art, at least one technical problem raised in the background technology is solved.

[0005] The technical solution adopted by the present invention to solve the technical problem is: a pre-filled tube glue seam visual inspection system, comprising: Label recognition module: When the pre-filled tube is being conveyed, the industrial camera set on the side of the conveyor belt is used to collect the image of the pre-filled tube, and the label recognition is performed on the image. According to the label recognition result, it is judged whether the label of the pre-filled tube is normal, and a processing signal is generated, wherein the processing signal includes a normal label signal and an analysis signal; The first glue seam identification module: based on the normal signal of the label, it determines whether the glue seam of the pre-filled tube is qualified by identifying the glue seam width in the unstored image of the label; Label impact analysis module: Based on the analysis signal, the label data and glue seam absolute error data of the labels in the image when glue seam detection is performed on multiple pre-filled tubes in the historical detection cycle are obtained according to the glue seam detection analysis report; and the label data and glue seam absolute error data are correlated and analyzed to determine whether the glue seam detection is affected by the label. If so, an impact signal is generated; Glue seam width correction module: Based on the impact signal, the glue seam width in the image of the pre-filled tube is corrected to obtain the corrected glue seam width KD 修1 and KD 修2 ; The second glue seam recognition module: based on the corrected glue seam width KD 修1 and KD 修2, the processing output is obtained to obtain the target glue seam width, and the glue seam of the pre-filled tube is tested for qualification according to the target glue seam width.

[0006] As a further technical solution of the present invention, the processing signal is generated in the following manner: If, in the images of the two pre-filled tubes, the label can be identified in the image of one pre-filled tube, but not in the image of the other pre-filled tube, the image of the pre-filled tube in which the label cannot be identified is marked as a label-not-stored image; and a label-normal signal is generated at the same time; If the label is identified in the images of both prefilled tubes, an analysis signal is generated.

[0007] As a further technical solution of the present invention, the process of judging whether the glue seam of the pre-filling tube is qualified is as follows: If the glue seam width of the pre-filled tube is greater than the standard glue seam width, it means that the glue seam of the pre-filled tube is unqualified; If the glue seam width of the pre-filled tube is less than or equal to the standard glue seam width, it means that the glue seam of the pre-filled tube is qualified.

[0008] As a further technical solution of the present invention, the process of determining whether the glue seam detection is affected by the label is as follows: If the glue seam width absolute error is greater than the preset glue seam width absolute error, the label image with abnormal glue seam width is marked as an image with an error exceeding the standard; Count the number of images with excessive errors and compare them with the number of images with labels to get the ratio of excessive errors, which is marked as WCb. According to the label area ratio in each error-exceeding image and the absolute error of the glue seam width, an absolute error change curve is constructed, and the absolute error change curve is divided into several continuous error sub-curves with equal horizontal lengths, and processed and analyzed to obtain the error increase ratio DZb and the fitting spacing Jn; By formula: The label influence value BY is obtained, where CD represents the horizontal length of the absolute error change curve; s1, s2 and s3 are all preset proportional coefficients; If the label influence value BY is greater than or equal to the label influence threshold, an influence signal is generated.

[0009] As a further technical solution of the present invention, the error increasing ratio DZb is obtained as follows: The slope K of the error sub-curve is calculated based on the coordinates of the two end points of the error sub-curve. i ; If the slope of the error sub-curve K i If it is greater than 0, the error sub-curve is marked as an increasing sub-curve The number of increasing sub-curves is counted and the ratio is processed with the total number of error sub-curves to obtain the error increasing ratio, which is marked as DZb.

[0010] As a further technical solution of the present invention, the fitting interval Jn is obtained in the following manner: The endpoints of all increasing sub-curves are integrated into an endpoint group, and the endpoint group is closed using the least square method to obtain the endpoint fitting line; A number of acquisition points are evenly set on each incremental sub-curve, and the shortest vertical distance from the acquisition point to the endpoint fitting line is measured to obtain the fitting spacing Jn of the acquisition points, where n is the number of the acquisition points.

[0011] As a further technical solution of the present invention, the corrected glue seam width KD 修1 and KD 修2 The way to obtain is: The slopes of all increasing sub-curves are summed and averaged to obtain the correction coefficient XZ; Obtain the area proportions MJ1 and MJ2 of the labels in the two images of the pre-filled tube respectively; By formula: Calculate the corrected glue seam width KD in the two images of the pre-filled tube 修1 and KD 修2 , where KD1 and KD2 are the initial glue gap widths in the two images of the pre-filled tube.

[0012] As a further technical solution of the present invention, the target glue seam width is obtained by: Get the corrected glue seam width KD 修1 and KD 修2 , processing to generate a first calculation signal and a second calculation signal; Based on the first calculation signal, the corrected glue seam width KD 修1 and KD 修2 Sum and average the values ​​to get the target glue seam width MKD; Based on the second calculation signal, according to the area proportions MJ1 and MJ2 of the labels in the two images of the pre-filled tube, the label area proportions MJB1 and MJB2 of the two images of the pre-filled tube are calculated; in, ; According to the label area ratio MJB1 and MJB2 of the two images of the pre-filled tube and the corrected glue seam width KD 修1 and KD 修2 , through the formula: The target glue seam width MKD is calculated.

[0013] As a further technical solution of the present invention, the first calculation signal and the second calculation signal are generated in the following manner: The corrected glue seam width KD 修1 and KD 修2 Performing difference processing and taking the absolute value of the difference to obtain a corrected glue seam width difference, and comparing the corrected glue seam width difference with a preset corrected glue seam width difference; If the corrected glue seam width difference is less than or equal to the preset corrected glue seam width difference, a first calculation signal is generated; If the corrected glue seam width difference is greater than the preset corrected glue seam width difference, a second calculation signal is generated.

[0014] A method for visually inspecting a glue seam of a pre-filled tube, comprising: Step 1: When the pre-filled tube is being conveyed, an industrial camera arranged on the side of the conveyor belt is used to collect an image of the pre-filled tube, and the image is subjected to label recognition. Whether the pre-filled tube label is normal is determined according to the label recognition result, and a processing signal is generated, wherein the processing signal includes a label normal signal and an analysis signal; Step 2: Based on the normal signal of the label, the glue seam width in the unstored image of the label is identified to determine whether the glue seam of the pre-filled tube is qualified; Step 3: Based on the analysis signal, obtain the label data and the glue seam absolute error data of the label in the image when glue seam detection is performed on multiple pre-filled tubes in the historical detection period according to the glue seam detection analysis report; and perform correlation analysis on the label data and the glue seam absolute error data to determine whether the glue seam detection is affected by the label. If so, generate an impact signal; Step 4: Based on the impact signal, correct the glue seam width in the image of the pre-filled tube to obtain the corrected glue seam width KD 修1 and KD 修2 ; Step 5: Based on the corrected glue seam width KD 修1 and KD 修2 , the processing output is obtained to obtain the target glue seam width, and the glue seam of the pre-filled tube is tested for qualification according to the target glue seam width.

[0015] The beneficial effects of the present invention are as follows: 1. The present invention first performs dual-image label detection on the pre-filled tube, which is beneficial to the qualified detection of the label on the one hand, and is beneficial to selecting an image without a label for glue seam width detection on the other hand. In addition, when both dual images of the pre-filled tube have labels, the present invention determines whether the existence of the label in the pre-filled tube image will affect the glue seam width detection through analysis of historical glue seam detection data, so as to facilitate the subsequent optimization of the glue seam width detection of the pre-filled tube.

[0016] 2. Based on the influencing signal, the glue seam width in the image of the pre-filled tube is corrected, and based on the corrected glue seam width, the glue seam is identified to obtain the target glue seam width, and the glue seam of the pre-filled tube is detected according to the target glue seam width to determine whether the glue seam of the pre-filled tube is qualified. The present invention processes and analyzes the corrected glue seam width, combines the glue seam widths of the double images of the pre-filled tube and the label area ratio, and comprehensively calculates the target glue seam width, thereby improving the accuracy of the glue seam of the pre-filled tube. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The present invention will be further described below in conjunction with the accompanying drawings.

[0018] Figure 1 It is a module diagram of a pre-filled tube glue seam visual inspection system according to an embodiment of the present invention;

[0019] Figure 2 It is a flowchart of the steps of a method for visually inspecting the glue seam of a pre-filled tube according to an embodiment of the present invention;

[0020] Figure 3 It is a schematic diagram of a pre-filled tube detection process in a pre-filled tube glue seam visual detection system according to an embodiment of the present invention;

[0021] Figure 4 It is a schematic diagram of pre-filled tube label detection in a pre-filled tube glue seam visual detection system according to an embodiment of the present invention;

[0022] Figure 5 It is a schematic diagram of pre-filled tube glue seam detection in a pre-filled tube glue seam visual detection system described in an embodiment of the present invention. DETAILED DESCRIPTION

[0023] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods. Example 1

[0024] like Figure 1 As shown, a pre-filled tube glue seam visual inspection system according to an embodiment of the present invention includes: Label recognition module: When the pre-filled tube is being conveyed, the industrial camera set on the side of the conveyor belt is used to collect the image of the pre-filled tube, and the label recognition is performed on the image. According to the label recognition result, it is judged whether the label of the pre-filled tube is normal, and a processing signal is generated, wherein the processing signal includes a normal label signal and an analysis signal; It should be noted that, see Figure 3As shown, the number of industrial cameras set on the side of the conveyor belt is 2, and the number of images of prefilled tubes (prefilled syringes) collected is 2. The purpose is: since the labels of the prefilled tubes are in different positions, if a single camera is set, it can prevent the label from being on the back of the prefilled tube, causing the industrial camera to be unable to collect the label, resulting in misjudgment of label recognition; Specifically, label recognition is performed in the image of the pre-filled tube; see Figure 4 (Left is with label, right is without label) as shown; If the labels cannot be identified in the images of the two pre-filled tubes, it means that the labels of the pre-filled tubes are abnormal, and the pre-filled tubes are screened out and the labels are reset, including but not limited to: subsidized labels and replacement labels; If, in the images of the two pre-filled tubes, a label can be identified in the image of one pre-filled tube, but not in the image of the other pre-filled tube, the image of the pre-filled tube in which the label cannot be identified is marked as a label-not-stored image; and a label-normal signal is generated at the same time, and the label-normal signal and the label-not-stored image are transmitted to the first glue seam recognition module; If the label can be identified in the images of both pre-filled tubes, an analysis signal is generated and sent to the label impact analysis module; The first glue seam identification module: based on the normal signal of the label, it determines whether the glue seam of the pre-filled tube is qualified by identifying the glue seam width in the unstored image of the label; Specifically, the glue seam width of the pre-filled tube in the unstored image of the label is measured and compared with the standard glue seam width (2mm); see Figure 5 (The left one shows a qualified glue seam of 1.45mm, and the right one shows a failed glue seam of 2.18mm); If the glue seam width of the pre-filled tube is greater than the standard glue seam width, it means that the glue seam of the pre-filled tube is unqualified; If the glue seam width of the pre-filled tube is less than or equal to the standard glue seam width, it means that the glue seam of the pre-filled tube is qualified; It should be noted that the glue gap of the pre-filling tube indicates the distance between the push rod and the rubber plug; Label impact analysis module: Based on the analysis signal, the label data and glue seam absolute error data of the labels in the image when glue seam detection is performed on multiple pre-filled tubes in the historical detection cycle are obtained according to the glue seam detection analysis report; and the label data and glue seam absolute error data are correlated and analyzed to determine whether the glue seam detection is affected by the label. If so, an impact signal is generated and sent to the glue seam width correction module; Among them, the label data includes the label area ratio, and the glue seam absolute error data includes the glue seam width absolute error; It should be noted that the label area ratio indicates the ratio of the label area to the label image area in the label presence image, and the glue seam width absolute error is the absolute value of the error between the actual glue seam width and the glue seam width measured in the image; it should also be noted that the label presence image indicates that there is a label in the image of the pre-filled tube; Specifically, the label area ratio and the absolute error of the glue seam width are obtained according to the glue seam detection analysis report; If the glue seam width absolute error is greater than the preset glue seam width absolute error, it means that the glue seam width is abnormal, and the label image with the abnormal glue seam width is marked as an image with an error exceeding the standard; If the glue seam width absolute error is less than or equal to the preset glue seam width absolute error, it means that the glue seam width is normal and no operation is performed; In the images with labels, the number of images with errors exceeding the standard is counted and the ratio is processed with the number of images with labels to obtain the ratio of the number of errors exceeding the standard and marked as WCb; With the label area ratio as the X-axis and the glue seam width absolute error as the Y-axis, an XY two-dimensional rectangular coordinate system is constructed to obtain the label area ratio and the glue seam width absolute error in each error-exceeding image, and mark and connect them in the XY two-dimensional rectangular coordinate system to obtain the absolute error change curve; Divide the absolute error variation curve into a number of continuous error sub-curves with equal horizontal lengths; Get the coordinates of the two end points of the error sub-curve in the XY two-dimensional rectangular coordinate system (X i1 , Y i1 ) and (X i2 , Y i2 ) Among them, (X i1 , Y i1 ) represents the coordinates of the initial endpoint of the error sub-curve, (X i2 , Y i2 ) represents the coordinates of the end point of the error sub-curve, i is the number of error sub-curves, according to the slope formula: Calculate the slope K of the error sub-curve i ; If the slope of the error sub-curve K i If it is greater than 0, the error sub-curve is marked as an increasing sub-curve If the slope of the error sub-curve K i If it is less than or equal to 0, the error sub-curve is marked as a non-increasing sub-curve; Count the number of increasing sub-curves and compare them with the total number of error sub-curves to obtain the error increasing ratio, which is marked as DZb; Based on the increasing sub-curve, the two end points of all increasing sub-curves are integrated into an end point group, and the end point group is closed using the least square method to obtain the end point fitting line; A number of collection points are evenly set on each incremental sub-curve, and the shortest vertical distance from the collection point to the endpoint fitting line is measured to obtain the fitting spacing Jn of the collection points, where n is the number of the collection points; The obtained error exceeding standard quantity ratio WCb, error increasing ratio DZb and fitting spacing Jn of acquisition points are processed by the formula: The label influence value BY is obtained, where CD represents the horizontal length of the absolute error change curve; s1, s2 and s3 are all preset proportional coefficients, s1 is 1.326, s2 is 1.145, and s3 is 1.015; It should be noted that the label influence value BY means: the label influence value BY is obtained by combining the error exceeding the standard number ratio WCb, the error increasing ratio DZb and the fitting spacing Jn of the acquisition points for data processing. The error exceeding the standard number ratio WCb reflects the proportion of the number of pre-filled tube images whose glue seam width absolute error exceeds the standard glue seam width absolute error in the historical monitoring period. The larger the number, the greater the degree to which the existence of the label affects the glue seam width detection. The error increasing ratio DZb reflects that as the area ratio of the label in the image of the pre-filled tube increases, the absolute error of the glue seam width increases. The area ratio of the label in the image of the pre-filled tube and the absolute error of the glue seam width show a positive growth influence relationship. The fitting spacing Jn reflects whether the area ratio of the label in the image and the absolute error of the glue seam width show a proportional growth relationship. The smaller the fitting spacing Jn, the linear proportional growth influence relationship between the area ratio of the label in the image and the absolute error of the glue seam width is. Compare the label influence value BY with the label influence threshold; If the label impact value BY is greater than or equal to the label impact threshold, an impact signal is generated and sent to the glue seam width correction module; If the label influence value BY is less than the label influence threshold, a non-influence signal is generated; Based on the non-influencing signal, the glue seam widths in the images of the two pre-filled tubes when the analysis signal is generated are identified, and the sum is taken to obtain the average value to obtain the target glue seam width; The technical solution of the embodiment of the present invention is as follows: the present invention first performs dual-image label detection on the pre-filled tube, which is beneficial to the qualified detection of the label on the one hand, and is beneficial to the selection of the image without the label for glue seam width detection on the other hand. In addition, when both the dual images of the pre-filled tube have labels, the present invention determines whether the existence of the label in the pre-filled tube image will affect the glue seam width detection through the analysis of historical glue seam detection data, so as to facilitate the subsequent optimization of the glue seam width detection of the pre-filled tube. Example 2

[0025] like Figure 1As shown, based on Example 1, the present invention optimizes the glue seam width detection when the label affects the glue seam width detection. Therefore, a pre-filled tube glue seam visual detection system according to an embodiment of the present invention includes: Glue seam width correction module: Based on the impact signal, the glue seam width in the image of the pre-filled tube is corrected to obtain the corrected glue seam width KD 修1 and KD 修2 ; Specifically, the slopes of all increasing sub-curves are summed and averaged to obtain the correction coefficient XZ; Obtain the area proportions MJ1 and MJ2 of the labels in the two images of the pre-filled tube respectively; By formula: Calculate the corrected glue seam width KD in the two images of the pre-filled tube 修1 and KD 修2 , where KD1 and KD2 are the initial glue gap widths in the two images of the pre-filled tube; It should be noted that the correction coefficient XZ reflects the proportional relationship between the absolute value of the glue seam width error and the label area. It represents the product of the area ratio of the label in the pre-filled tube image and the correction coefficient, which reflects the correction value of the glue seam width in the pre-filled tube image. The glue seam width in the pre-filled tube image is corrected on the basis of the correction value, and the corrected glue seam width is obtained. It should also be noted that the presence of the label makes the detection instrument more sensitive or accurate to the remaining area. This situation is relatively rare. In actual glue seam detection, the presence of the label usually causes a negative error in the glue seam width. Therefore, a positive correction is made on the glue seam width of the image; The second glue seam recognition module: based on the corrected glue seam width KD 修1 and KD 修2 , processing the output to obtain the target glue seam width, and performing a qualified test on the glue seam of the pre-filled tube according to the target glue seam width; Specifically, obtain the corrected glue seam width KD 修1 and KD 修2 , and perform difference processing, and take the absolute value of the difference to obtain a corrected glue seam width difference, and compare the corrected glue seam width difference with a preset corrected glue seam width difference; If the corrected glue seam width difference is less than or equal to the preset corrected glue seam width difference, a first calculation signal is generated; If the corrected glue seam width difference is greater than the preset corrected glue seam width difference, a second calculation signal is generated; Based on the first calculation signal, the corrected glue seam width KD 修1 and KD 修2 Sum and average the values ​​to get the target glue seam width MKD; It should be noted that the first calculation signal indicates that the corrected glue seam width difference is close to the preset corrected glue seam width difference, and the corrected glue seam width KD 修1 and KD 修2 The target glue seam width is output by summing and averaging. On the contrary, the second calculation signal indicates that the corrected glue seam width difference is far away from the preset corrected glue seam width difference. In order to output the target glue seam width more accurately, it is necessary to combine the area ratio of the label in the two images of the pre-filled tube and the corrected glue seam width KD 修1 and KD 修2 , fusion output target glue seam width; Based on the second calculation signal, according to the area proportions MJ1 and MJ2 of the labels in the two images of the pre-filled tube, the label area proportions MJB1 and MJB2 of the two images of the pre-filled tube are calculated; in, ; According to the label area ratio MJB1 and MJB2 of the two images of the pre-filled tube and the corrected glue seam width KD 修1 and KD 修2 Calculate the target glue seam width MKD; Specifically, through the formula: ; It should be noted that the above formula The meaning is: after the correction of the glue seam width KD 修1 and KD 修2 When the deviation is large, the influence coefficient of the corrected glue seam width is determined according to the ratio of the label area in the two images of the pre-filled tube, and the final target glue seam width is output in combination with the influence coefficient and the corrected glue seam width in the two images of the pre-filled tube, thereby improving the accuracy of the pre-filled tube glue seam detection; It should also be noted that the label area ratio MJB1 and the corrected glue seam width KD 修1 Corresponding to the image of the same pre-filled tube, the label area ratio MJB2 and the corrected glue seam width KD 修2 The images corresponding to the same pre-filled tube; Compare the target glue seam width MKD with the standard glue seam width; If the target glue seam width MKD is greater than the standard glue seam width, the glue seam of the pre-filled tube is unqualified; If the target glue seam width MKD is less than or equal to the standard glue seam width, the glue seam of the pre-filled tube is qualified; The technical solution of the embodiment of the present invention is: based on the influencing signal, the glue seam width in the image of the pre-filled tube is corrected, based on the corrected glue seam width, the glue seam is identified to obtain the target glue seam width, and the glue seam of the pre-filled tube is detected according to the target glue seam width to determine whether the glue seam of the pre-filled tube is qualified. The present invention processes and analyzes the corrected glue seam width, combines the glue seam width of the double images of the pre-filled tube and the label area ratio, and comprehensively calculates the target glue seam width, thereby improving the accuracy of the glue seam of the pre-filled tube. Example 3

[0026] like Figure 2 As shown, a method for visually inspecting a glue seam of a pre-filled tube according to an embodiment of the present invention includes: Step 1: When the pre-filled tube is being conveyed, an industrial camera arranged on the side of the conveyor belt is used to collect an image of the pre-filled tube, and the image is subjected to label recognition. Whether the pre-filled tube label is normal is determined according to the label recognition result, and a processing signal is generated, wherein the processing signal includes a label normal signal and an analysis signal; Step 2: Based on the normal signal of the label, the glue seam width in the unstored image of the label is identified to determine whether the glue seam of the pre-filled tube is qualified; Step 3: Based on the analysis signal, obtain the label data and the glue seam absolute error data of the label in the image when glue seam detection is performed on multiple pre-filled tubes in the historical detection period according to the glue seam detection analysis report; and perform correlation analysis on the label data and the glue seam absolute error data to determine whether the glue seam detection is affected by the label. If so, generate an impact signal; Step 4: Based on the impact signal, correct the glue seam width in the image of the pre-filled tube to obtain the corrected glue seam width KD 修1 and KD 修2 ; Step 5: Based on the corrected glue seam width KD 修1 and KD 修2 , the processing output is obtained to obtain the target glue seam width, and the glue seam of the pre-filled tube is tested for qualification according to the target glue seam width.

[0027] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A pre-filled tube glue seam visual inspection system, characterized by: include: Label recognition module: When the pre-filled tube is being conveyed, the industrial camera set on the side of the conveyor belt is used to collect the image of the pre-filled tube, and the label recognition is performed on the image. According to the label recognition result, it is judged whether the label of the pre-filled tube is normal, and a processing signal is generated, wherein the processing signal includes a normal label signal and an analysis signal; The first glue seam identification module: based on the normal signal of the label, it determines whether the glue seam of the pre-filled tube is qualified by identifying the glue seam width in the unstored image of the label; Label impact analysis module: Based on the analysis signal, the label data and glue seam absolute error data of the labels in the image when glue seam detection is performed on multiple pre-filled tubes in the historical detection cycle are obtained according to the glue seam detection analysis report; and the label data and glue seam absolute error data are correlated and analyzed to determine whether the glue seam detection is affected by the label. If so, an impact signal is generated; Glue seam width correction module: Based on the impact signal, the glue seam width in the image of the pre-filled tube is corrected to obtain the corrected glue seam width KD 修1 and KD 修2 ; The second glue seam recognition module: based on the corrected glue seam width KD 修1 and KD 修2 , the processing output is obtained to obtain the target glue seam width, and the glue seam of the pre-filled tube is tested for qualification according to the target glue seam width.

2. A pre-filled tube glue seam visual inspection system according to claim 1, characterized in that: The processing signal is generated in the following manner: If, in the images of the two pre-filled tubes, the label can be identified in the image of one pre-filled tube, but not in the image of the other pre-filled tube, the image of the pre-filled tube in which the label cannot be identified is marked as a label-not-stored image; and a label-normal signal is generated at the same time; If the label is identified in the images of both prefilled tubes, an analysis signal is generated.

3. A pre-filled tube glue seam visual inspection system according to claim 1, characterized in that: The process of judging whether the glue seam of the pre-filling tube is qualified is as follows: If the glue seam width of the pre-filled tube is greater than the standard glue seam width, it means that the glue seam of the pre-filled tube is unqualified; If the glue seam width of the pre-filled tube is less than or equal to the standard glue seam width, it means that the glue seam of the pre-filled tube is qualified.

4. A pre-filled tube glue seam visual inspection system according to claim 1, characterized in that: The process of determining whether the glue seam detection is affected by the label is as follows: If the glue seam width absolute error is greater than the preset glue seam width absolute error, the label image with abnormal glue seam width is marked as an image with an error exceeding the standard; Count the number of images with excessive errors and compare them with the number of images with labels to get the ratio of excessive errors, which is marked as WCb. According to the label area ratio in each error-exceeding image and the absolute error of the glue seam width, an absolute error change curve is constructed, and the absolute error change curve is divided into several continuous error sub-curves with equal horizontal lengths, and processed and analyzed to obtain the error increase ratio DZb and the fitting spacing Jn; By formula: The label influence value BY is obtained, where CD represents the horizontal length of the absolute error change curve; s1, s2 and s3 are all preset proportional coefficients; If the label influence value BY is greater than or equal to the label influence threshold, an influence signal is generated.

5. A pre-filled tube glue seam visual inspection system according to claim 4, characterized in that: The error increasing ratio DZb is obtained as follows: The slope K of the error sub-curve is calculated based on the coordinates of the two end points of the error sub-curve. i ; If the slope of the error sub-curve K i If it is greater than 0, the error sub-curve is marked as an increasing sub-curve The number of increasing sub-curves is counted and the ratio is processed with the total number of error sub-curves to obtain the error increasing ratio, which is marked as DZb.

6. A pre-filled tube glue seam visual inspection system according to claim 4, characterized in that: The fitting spacing Jn is obtained as follows: The endpoints of all increasing sub-curves are integrated into an endpoint group, and the endpoint group is closed using the least square method to obtain the endpoint fitting line; A number of acquisition points are evenly set on each incremental sub-curve, and the shortest vertical distance from the acquisition point to the endpoint fitting line is measured to obtain the fitting spacing Jn of the acquisition points, where n is the number of the acquisition points.

7. A pre-filled tube glue seam visual inspection system according to claim 1, characterized in that: The corrected glue seam width KD 修1 and KD 修2 The way to obtain is: The slopes of all increasing sub-curves are summed and averaged to obtain the correction coefficient XZ; Obtain the area proportions MJ1 and MJ2 of the labels in the two images of the pre-filled tube respectively; By formula: Calculate the corrected glue seam width KD in the two images of the pre-filled tube 修1 and KD 修2 , where KD1 and KD2 are the initial glue gap widths in the two images of the pre-filled tube.

8. A pre-filled tube glue seam visual inspection system according to claim 1, characterized in that: The target glue seam width is obtained as follows: Get the corrected glue seam width KD 修1 and KD 修2 , processing to generate a first calculation signal and a second calculation signal; Based on the first calculation signal, the corrected glue seam width KD 修1 and KD 修2 Sum and average the values ​​to get the target glue seam width MKD; Based on the second calculation signal, according to the area proportions MJ1 and MJ2 of the labels in the two images of the pre-filled tube, the label area proportions MJB1 and MJB2 of the two images of the pre-filled tube are calculated; wherein, ; According to the label area ratio MJB1 and MJB2 of the two images of the pre-filled tube and the corrected glue seam width KD 修1 and KD 修2 , through the formula: The target glue seam width MKD is calculated.

9. A pre-filled tube glue seam visual inspection system according to claim 8, characterized in that: The first calculation signal and the second calculation signal are generated in the following manner: The corrected glue seam width KD 修1 and KD 修2 Performing difference processing and taking the absolute value of the difference to obtain a corrected glue seam width difference, and comparing the corrected glue seam width difference with a preset corrected glue seam width difference; If the corrected glue seam width difference is less than or equal to the preset corrected glue seam width difference, a first calculation signal is generated; If the corrected glue seam width difference is greater than the preset corrected glue seam width difference, a second calculation signal is generated.

10. A method for visual inspection of glue seams of pre-filled tubes, characterized in that: include: Step 1: When the pre-filled tube is being conveyed, an industrial camera arranged on the side of the conveyor belt is used to collect an image of the pre-filled tube, and the image is subjected to label recognition. Whether the pre-filled tube label is normal is determined according to the label recognition result, and a processing signal is generated, wherein the processing signal includes a label normal signal and an analysis signal; Step 2: Based on the normal signal of the label, the glue seam width in the unstored image of the label is identified to determine whether the glue seam of the pre-filled tube is qualified; Step 3: Based on the analysis signal, obtain the label data and the glue seam absolute error data of the label in the image when glue seam detection is performed on multiple pre-filled tubes in the historical detection period according to the glue seam detection analysis report; and perform correlation analysis on the label data and the glue seam absolute error data to determine whether the glue seam detection is affected by the label. If so, generate an impact signal; Step 4: Based on the impact signal, correct the glue seam width in the image of the pre-filled tube to obtain the corrected glue seam width KD 修1 and KD 修2 ; Step 5: Based on the corrected glue seam width KD 修1 and KD 修2 , the processing output is obtained to obtain the target glue seam width, and the glue seam of the pre-filled tube is tested for qualification according to the target glue seam width.

Images