Defect identification method for non-woven fabric

Abstract

The invention relates to a non-woven fabric defect identification method, which comprises the following steps of: 1, performing signal acquisition on a moving non-woven fabric to be detected by adopting a line camera to obtain a line video signal; step 2, performing shadow correction on the line video signal; step 3, comparing the line video signal after shadow correction with a set threshold range, and screening bright pixel points and / or dark pixel points; step 3, performing 2D processing on the line video signal after shadow correction, and synthesizing a plurality of adjacent line video signals into a frame image to obtain a 2D video signal; step 4, carrying out noise reduction processing on the 2D video signal by adopting low-pass filtering; step 5, carrying out coloring processing on the 2D video signal after noise reduction processing; and 6, identifying the sizes of the color marking blocks, comparing the sizes with a set threshold Hm, and screening out the color marking blocks which are greater than or equal to the threshold Hm, namely defects. According to the method, the defects of the traditional plain non-woven fabric can be identified, and the defects of the punched and embossed non-woven fabric can also be accurately identified.

Description

Technical Field

[0001] The present invention belongs to the technical field of non-woven fabric quality inspection, and particularly relates to a method for identifying defects in non-woven fabrics. Background Art

[0002] The non-woven fabric industry, especially the non-woven fabric industry for sanitary products, has developed rapidly in recent years, and consumers' demand for product cleanliness is also getting higher and higher.

[0003] With the rapid development of the non-woven fabric industry, hot air non-woven fabrics have gradually developed from the initial unembossed smooth non-woven fabrics to embossed plain non-woven fabrics, punched or punched and embossed plain non-woven fabrics. The appearance of embossed patterns and punched holes on non-woven fabrics undoubtedly interferes with their defect detection, making it impossible for traditional defect detection methods to accurately identify the defects of punched and embossed non-woven fabrics, thus reducing the product quality and greatly increasing the probability of consumer complaints.

[0004] Therefore, it is crucial to develop a defect identification method for non-woven fabrics that can not only be used for traditional plain non-woven fabrics but also be compatible with non-woven fabrics to improve the ex-factory quality of non-woven fabrics. Summary of the Invention

[0005] The purpose of the present invention is to overcome the defects in the prior art and provide a method for identifying defects in non-woven fabrics, which can not only identify the defects of traditional plain non-woven fabrics but also accurately identify the defects of punched and embossed non-woven fabrics.

[0006] To achieve the above purpose, the technical solutions adopted by the present invention are as follows:

[0007] A method for identifying defects in non-woven fabrics includes the following steps:

[0008] Step 1: Use a line camera to collect signals from the moving non-woven fabric to be detected, and process them through a SMASH processor to obtain a line video signal.

[0009] Step 2: Perform shadow correction on the line video signal in Step 1.

[0010] Step 3: Compare the pixel brightness values of the line video signal after shadow correction with the set threshold range H1 - H2, and screen out the pixel points outside the threshold range H1 - H2 to obtain bright pixel points and / or dark pixel points.

[0011] Step 3: Perform 2D processing on the line video signal after shadow correction, synthesize multiple adjacent line video signals into a frame image to obtain a 2D video signal.

[0012] Step 4: Perform noise reduction processing on the 2D video signal in Step 3 using low-pass filtering.

[0013] Step 5: Perform coloring processing on the denoised 2D video signal, using different colors to mark bright pixel points and dark pixel points, and obtain a color-marked block with defects.

[0014] Step 6: Identify the size of the color-marked block and compare it with the set threshold Hm, and filter out the color-marked blocks greater than or equal to the threshold Hm, that is, defects.

[0015] As a further technical solution, the shadow correction includes:

[0016] Use a line camera to collect signals from a moving non-woven fabric sample (without defects), and process them through a SMASH processor to obtain a sample line video signal as a positive function; then calculate the difference between the gray value of the positive function and the standard flat value to obtain an inverse function.

[0017] Use the inverse function to perform shadow correction on the line video signal obtained in Step 1.

[0018] As a further technical solution, the line video signal is drawn with the pixel as the abscissa and the pixel gray value as the ordinate.

[0019] As a further technical solution, the pattern and texture of the non-woven fabric sample without defects are the same as those of the non-woven fabric to be detected in Step 1.

[0020] As a further technical solution, the low-pass filtering uses a Butterworth low-pass filter.

[0021] As a further technical solution, the size of the defect is expressed by the area of the defect or the length of the diagonal of the defect.

[0022] As a further technical solution, the non-woven fabric includes any one of ordinary plain non-woven fabric, embossed plain non-woven fabric, perforated plain non-woven fabric, and perforated embossed plain non-woven fabric.

[0023] As a further technical solution, the defects include cotton knots, holes, sparse meshes, black dots, and hairs.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0025] The present invention uses an inverse function to perform shadow correction on the collected line video signal, uses low-pass filtering technology to denoise the synthesized 2D video signal, and screens defects according to the size of the defects. It can not only identify the defects of traditional plain non-woven fabrics, but also accurately identify the defects of perforated embossed non-woven fabrics, greatly improving the ex-factory quality of perforated embossed non-woven fabrics. Brief Description of the Drawings

[0026] Figure 1 It is a picture of a non-woven fabric sample in an embodiment of the present invention;

[0027] Figure 2 It is the line video signal diagram of the non-woven fabric to be measured in an embodiment of the present invention;

[0028] Figure 3 It is the inverse function diagram in an embodiment of the present invention;

[0029] Figure 4 It is the line video signal diagram after shadow correction in an embodiment of the present invention;

[0030] Figure 5 It is the synthesized 2D video signal diagram in an embodiment of the present invention;

[0031] Figure 6 It is the schematic diagram of filtering in an embodiment of the present invention;

[0032] In Figure 6 A: Original 2D video signal diagram, B: New 2D video signal diagram;

[0033] Figure 7 It is the video signal diagram obtained with different filtering sizes in an embodiment of the present invention;

[0034] In Figure 7 A: 4×4, 2D video information diagram; B: 8×8, 2D video signal diagram; C: 4×4, line video signal diagram; D: 8*8, line video signal diagram;

[0035] Figure 8 The 2D video signal diagram after color marking in an embodiment of the present invention. Specific embodiments

[0036] Next, the technical solutions of the present invention will be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are some, but not all, of the embodiments of the present invention. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0037] The following further describes the present invention in detail with reference to the accompanying drawings.

[0038] A method for identifying defects in non-woven fabrics can be applied to ordinary plain non-woven fabrics, embossed plain non-woven fabrics, perforated plain non-woven fabrics or perforated embossed plain non-woven fabrics, and can identify defects such as neps, holes, sparse meshes, black dots, hair, etc. In this embodiment, taking the embossed and perforated plain non-woven fabric as shown in Figure 1 as an example, the specific steps are as follows:

[0039] Step (1), obtaining the inverse function:

[0040] First, a line camera is used to collect signals from a moving non-woven fabric sample (without defects), and after being processed by a SMASH processor, a line video signal of the sample is obtained as a positive function;

[0041] Then, the difference between the gray value of the positive function and the standard flat value is calculated to obtain an inverse function (see Figure 2 );

[0042] Among them, the standard flat value is the standard calibrated gray value. The calibrated gray value generally ranges from 1 to 256. Therefore, the standard flat value is taken as half of 256, that is, 128.

[0043] The abscissa of the line video signal and the inverse function is the pixel point, and the ordinate is the gray value.

[0044] Step (2), signal acquisition:

[0045] A line camera is used to collect signals from the moving non-woven fabric to be detected, and after being processed by a SMASH processor, a line video signal is obtained (see Figure 3 );

[0046] The shadow is a regular change in the line video signal caused by the optical distortion of the lens. This change will lead to misjudgment of defects. Therefore, shadow correction is required before defect identification;

[0047] Step (2), shadow correction:

[0048] The inverse function obtained in step (1) is used to perform shadow correction on the line video signal obtained in step (2) to obtain a shadow-corrected line video signal (see Figure 4 );

[0049] Step 3, screening of unqualified pixel points: The brightness values of the pixel points of the shadow-corrected line video signal are compared with the set threshold range H1 - H2 (the threshold range set in this embodiment is 80 - 195), and the pixel points exceeding the set threshold range H1 - H2 are screened out to obtain bright pixel points and / or dark pixel points;

[0050] Step 3, synthesis: The shadow-corrected line video signal is subjected to 2D processing, and multiple adjacent line video signals are synthesized into a frame image to obtain a 2D video signal; Figure 5 Is the frame image (20 frames) synthesized from the line video signal;

[0051] Step 4, filtering and noise reduction:

[0052] The 2D video signal in step 3 is subjected to noise reduction processing using low-pass filtering; specifically, a Butterworth low-pass filter is used for noise reduction processing.

[0053] The Low - pass filtering: is a filtering method that allows low - frequency signals to pass through while blocking and weakening high - frequency signals exceeding a set critical value; among them, the Butterworth low - pass filter replaces the original line video signal with a line video signal obtained by re - averaging the pixel gray values, to achieve noise reduction; the specific processing is as follows:

[0054] As Figure 6 shown, the low - pass filtering technology is used to Figure 6 sum the pixel gray values in each unit (2×2, 4×4, 4×16, 64×64, …; all are powers of 2) in the original 2D video signal in A and evenly distribute them to each pixel, to obtain Figure 6 the new 2D video signal in B; the larger the filter size we choose, the stronger the averaging effect, the stronger the ability to suppress changes, and the smoother the boundaries of the defects appear; for example Figure 7 in, the boundary of the defect in B with 8×8 filtering is smoother than the boundary of the defect in A with 4×4 filtering, where C and D are the line video signal diagrams of the rows where the defects are located in their respective figures;

[0055] The purpose of the low - pass filtering denoising in the present invention is to reduce the false detection caused by high - contrast background textures;

[0056] Step 5: Perform coloring processing on the 2D video signal after noise reduction processing, use different colors to mark bright pixel points and dark pixel points, to obtain a color - marked block with defects; in this embodiment, Figure 8 as shown, the bright pixel points are colored red and the dark pixel points are marked blue;

[0057] The purpose of the present invention to color - mark the bright pixel point and dark pixel point defects is to distinguish the types of defects and identify the regions of the defects;

[0058] Step 6: Identify the size of the color - marked block and compare it with the set threshold Hm, and screen out the color - marked blocks greater than or equal to the threshold Hm, that is, defects;

[0059] Among them, the size of the identified color - marked block can be represented by the area and / or the length of the diagonal; through threshold comparison, smaller color - marked blocks (i.e., defect blocks) that hardly affect the product quality can be eliminated, while blocks with obvious defects, that is, defects, can be screened out.

[0060] In this embodiment, it is represented by the area, and the set threshold Hm is 0.5mm 2 , then the color - marked blocks greater than or equal to 0.5mm 2 are identified as defects.

[0061] The above-described embodiments are only preferred embodiments of the present invention and not an exhaustive list of the feasible embodiments of the present invention. Any obvious modifications made by those of ordinary skill in the art without departing from the principle and spirit of the present invention should be considered to be included within the scope of protection of the claims of the present invention.

Claims

1. A method for identifying defects in nonwoven fabrics, characterized in that: The steps include: Step 1: Use a line camera to collect signals from the moving non-woven fabric to be detected, and process them through a SMASH processor to obtain a line video signal; Step 2, performing shadow correction on the line video signal of step 1; Step 3: Compare the pixel brightness value of the line video signal after shadow correction with the set threshold range H1-H2, filter out the pixels exceeding the threshold range H1-H2, and obtain bright pixels and / or dark pixels; Step 3, performing 2D processing on the line video signal after shadow correction, synthesizing a plurality of adjacent line video signals into a frame image, and obtaining a 2D video signal; Step 4: low-pass filter the 2D video signal in step 3 to perform noise reduction processing; Step 5: coloring the 2D video signal after the noise reduction process, using different colors to mark bright pixels and dark pixels, and obtaining color-marked blocks with defects; Step 6: Identify the size of the color-marked block and compare it with the set threshold value Hm, and select the color-marked blocks that are greater than or equal to the threshold value Hm, i.e., defects.

2. A method for identifying defects in nonwoven fabrics according to claim 1, characterized in that: The shadow correction comprises: A line camera is used to collect signals of the moving nonwoven fabric sample, and the sample line video signal is obtained by processing it with a SMASH processor as a positive function. Then the difference between the gray value of the positive function and the standard average value is calculated to obtain the inverse function. The inverse function is used to perform shading correction on the line video signal obtained in step 1.

3. A method for identifying defects in nonwoven fabrics according to claim 2, characterized in that: The line video signal is plotted with pixels as the horizontal coordinate and the pixel gray value as the vertical coordinate.

4. The method for identifying defects in nonwoven fabrics according to claim 2, characterized in that: The pattern and texture of the defect-free non-woven fabric sample are the same as those of the non-woven fabric to be tested in step 1.

5. The method for identifying defects in nonwoven fabrics according to claim 1, characterized in that: The low-pass filtering adopts a Butterworth low-pass filter.

6. The method for identifying defects in nonwoven fabrics according to claim 1, characterized in that: The defect size is expressed as the defect area or the length of the defect diagonal.

7. A method for identifying defects in nonwoven fabrics according to claim 1, characterized in that: The nonwoven fabric includes any one of ordinary plain nonwoven fabric, embossed plain nonwoven fabric, perforated plain nonwoven fabric and perforated embossed plain nonwoven fabric.

8. The method for identifying defects in nonwoven fabrics according to claim 1, characterized in that: The defects include neps, holes, thin webs, black spots, and hair.

Images