A defect assessment method and system for door panel wood

Abstract

The present invention relates to the field of data processing, and in particular to a defect assessment method and system for door panel wood, comprising the following steps: Step 1: obtaining image data of door panel board; Step 2: classifying tree knots into dead knots and live knots according to the generation process of tree knots and the influence on the board, and obtaining the area corresponding to the tree knots, and weighing and evaluating each tree knot area according to the darker veneer structure of the outer ring of the dead knot and the texture difference between the two tree knots, and obtaining the significance of the decay effect; Step 3: judging the defect of the door panel board according to the significance of the decay effect. That is, the scheme of the present invention can judge the quality of the wood board board and obtain the specific location corresponding to the defect, which is detailed and accurate.

Description

Technical Field

[0001] The invention relates to the field of data processing, and in particular to a defect assessment method and system for door panel wood. Background Art

[0002] Door panels in my country are mainly divided into solid wood panels, blister panels, melamine panels, UV panels, etc. Among them, solid wood panels are popular among consumers due to their natural environmental protection, durability, heat insulation and sound insulation. In order to meet the huge market demand for wood, my country's annual demand for wood exceeds 800 million cubic meters. In order to improve the utilization rate of wood, reduce the impact of wood defects on the quality of panels, and achieve sustainable development of door panel processing, it is necessary to conduct defect detection on panels.

[0003] In order to ensure the normal use of the board, non-destructive testing is usually performed on the board. Stress wave testing applies a certain amount of mechanical stress to the wood to identify whether there are defects inside the wood. The result is accurate, but the specific location of the internal defects cannot be obtained. Ultrasonic testing and X-ray testing are often greatly affected by the high requirements for the medium and instruments when detecting board defects. The existing machine vision methods that are more effective for door panel defects often require the use of neural networks, but the training process of neural networks is relatively complicated, the sample volume required is large, and the training time is slow. Therefore, there is an urgent need for a method of identifying defects that does not require a lot of training and is easy and fast to use. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of the present invention is to provide a defect assessment method and system for door panel wood. The technical solutions adopted are as follows:

[0005] The present invention provides a method for evaluating defects of door panel wood, comprising the following steps:

[0006] Step 1: Obtain a door panel image and preprocess the image to obtain a grayscale image;

[0007] Step 2: Segment the grayscale image according to the generation process of the tree knot and its impact on the board to obtain the tree knot type to be determined area; use the Canny operator to obtain the edge image of each tree knot type to be determined area, and perform circle fitting on the edge pixels of the outermost circle in the edge image. When the goodness of fit is greater than the set value, the tree knot type to be determined area is considered to be a tree knot area;

[0008] The knot area is divided to obtain a division threshold, grayscale values ​​less than the division threshold are screened out, the proportion of phase pixels in each knot area is determined, the highest proportion is selected as the veneer wood grain annular area, and the dark veneer significance coefficient of the veneer wood grain annular area is calculated;

[0009] The looseness of the ring texture is calculated based on the annular area of ​​the wood grain of the veneer, and then the looseness of the ring texture is obtained;

[0010] The significance of the decay effect is obtained according to the significance coefficient of the dark veneer and the looseness of the ring texture;

[0011] Step 3: Determine the defects of the door panel material based on the significance of the corrosion effect.

[0012] Preferably, the method for obtaining the area to be determined for the type of tree knot is: use the OTSU law to segment the grayscale image to obtain the part of the image that is suspected to be a tree knot; assign the grayscale values ​​corresponding to all pixels contained in the suspected tree knot part to be 1, and assign the grayscale values ​​corresponding to the remaining pixels to be 0, to obtain a binary image; use opening operation and area filling to process the binary image; mark the pixels with corresponding values ​​of 1 in the obtained image, and obtain the corresponding pixels in the grayscale image according to the positions of the pixels, and the area composed of these pixels is the area to be determined for the type of tree knot.

[0013] Preferably, the dark veneer significance coefficient is:

[0014]

[0015] Among them, g1 is the mean gray value of the pixel points in the wood grain annular area of ​​the veneer, g2 is the mean gray value of the pixel points in the wood grain annular area on one side of the wood grain annular area; g3 is the mean gray value of the pixel points in the wood grain annular area on the other side of the wood grain annular area, r is the radius of the circle obtained by fitting each pixel point in the wood grain annular area of ​​the veneer with a circle, R is the radius of the circle fitted by the pixel points of the outermost edge in the edge image, d(x1), d(x2), ..., d(x n2 ) are the distances from all pixels in the wood grain annular area to the edge of the area, n1 is the number of wood grain annular areas, n2 is the number of pixels in the wood grain annular area, and min() is the minimum value.

[0016] Preferably, the looseness of the basic construction ring texture is:

[0017]

[0018] Among them, h is the gray value corresponding to any pixel point in the wood grain annular area, h1, h2, h3, h4 are the gray values ​​of the corresponding pixels in the four neighborhoods of any pixel point, and h max and h min are the maximum and minimum grayscale values ​​of the pixels in the wood grain annular area, respectively, and n3 is the number of pixels in the wood grain annular area.

[0019] Preferably, the ring texture looseness is the average of the base-constructed ring texture looseness of all knot areas.

[0020] Preferably, the decay impact significance is the product of the dark veneer significance coefficient and the looseness of the ring texture.

[0021] The present invention also provides a defect assessment system for door panel wood, comprising a memory and a processor, wherein the processor is used to execute the steps of implementing the above-mentioned defect assessment method for door panel wood stored in the memory.

[0022] Beneficial effects of the present invention:

[0023] The solution of the present invention has simple requirements on the environment, instruments and manual operation, does not require a large number of samples for training, and is convenient and quick to use. The judgment process uses the specific formation mechanism of dead knots and live knots and the unique dark veneer structure and texture as characteristics to judge the specific division of tree knots on the door panel. While judging the quality of the wood panel, the specific location corresponding to the defect can be obtained, which is detailed and accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0025] Figure 1 is a flow chart of a defect assessment method for door panel wood of the present invention;

[0026] Figure 2 is a grayscale image of the door panel joint;

[0027] Figure 3 It is a grayscale image of the dead knot of the door panel;

[0028] Figure 4 is the grayscale image of the door panel joint after denoising;

[0029] Figure 5 It is the grayscale image of the dead knot of the door panel after denoising;

[0030] Figure 6 It is a binary image of the suspected live knot part;

[0031] Figure 7 It is a binary image of the suspected dead knot part of the tree;

[0032] Figure 8 It is a binary image of the area to be determined for the tree knot type corresponding to the active knot;

[0033] Fig. 9 It is a binary image of the area to be determined for the type of tree knot corresponding to the dead knot. DETAILED DESCRIPTION

[0034] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the scheme according to the present invention, its specific implementation, structure, features and effects in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0035] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0036] The present invention provides an embodiment of a defect assessment method for door panel wood, please refer to Figure 1 As shown, the following steps are included:

[0037] Step 1: Obtain the door panel image and preprocess the image to obtain a grayscale image.

[0038] To ensure the stability of obtaining the board image, the door board made of logs is placed at the fixed position of the conveyor belt at the same length. A strip light source is used to ensure the stability of the light intensity at the position where the image is taken. The CCD camera is fixed above the position where the image is to be taken. An image of the door board is obtained at the same interval. The shooting interval is the interval time when the adjacent door boards appear at the center of the camera shooting area. The door board image is an RGB image. The RGB image is converted into a grayscale image, such as Figure 2 and Figure 3 As shown, the median filter is used to denoise the grayscale image, as Figure 4 and Figure 5 shown.

[0039] Step 2: Segment the grayscale image according to the generation process of the tree knot and its impact on the board to obtain the tree knot type to be determined area; use the Canny operator to obtain the edge image of each tree knot type to be determined area, and perform circle fitting on the edge pixels of the outermost circle in the edge image. When the goodness of fit is greater than the set value, the tree knot type to be determined area is considered to be a tree knot area;

[0040] The knot area is divided to obtain a division threshold, grayscale values ​​less than the division threshold are screened out, the proportion of phase pixels in each knot area is determined, the highest proportion is selected as the veneer wood grain annular area, and the dark veneer significance coefficient of the veneer wood grain annular area is calculated;

[0041] The looseness of the ring texture is calculated based on the annular area of ​​the wood grain of the veneer, and then the looseness of the ring texture is obtained;

[0042] The significance of the decay influence is obtained according to the significance coefficient of the dark veneer and the looseness of the ring texture.

[0043] During the growth of trees, bud meristems will continue to grow and grow branches in order to grow more leaves and get more sunlight. The parts where branches meet the trunk are tightly connected and have a hard texture, which is called a knot. Normally constructed knots are harder in texture, have less impact on the board, and have more complex wood textures, which can provide a unique effect for wood decoration. Only when the density is too high do they need to be removed. These knots are called live knots in the board. However, some branches fall off naturally before being cut down, and are not connected to the surrounding wood tissue. Some have rotted or become soft, which will reduce the strength of the entire board and affect the service life of the board. They need to be removed. These knots are called dead knots in the board.

[0044] Since door panels are made of logs, there will inevitably be parts where branches connect to the trunk, as well as live and dead knots. Therefore, when inspecting door panels for defects, it is necessary to identify the parts of the panels where knots appear, distinguish whether each knot is live or dead, and remove dead knots and live knots that are too dense. The specific steps are as follows:

[0045] Use OTSU's law to segment the grayscale image and obtain the part of the image that may be a tree knot. If this algorithm is used alone, there will still be some isolated points or small areas corresponding to darker textures in the image. Assign the value corresponding to all pixels in the part that may be a tree knot to 1, and assign the value corresponding to the remaining pixels to 0, and obtain a binary image, such as Figure 6 and Figure 7 As shown. The binary image is opened to remove obvious black spots and thorns that affect subsequent analysis. The obtained image is region filled (algorithm in morphological processing) to fill the empty parts in the region of the image. The pixels with a corresponding value of 1 in the obtained image are marked, and the corresponding pixels are obtained according to the positions of the pixels in the grayscale image. The area composed of these pixels is the area that may be a tree knot, which is called the tree knot type to be determined area, such as Figure 8 and Fig. 9 shown.

[0046] The canny operator is used to obtain the edge image of each tree knot type to be determined. The edge image is a binary image. Since trees produce annual rings during their growth, there will be multiple edges in the edge image of each tree knot type to be determined. The outermost edge in the edge image is the edge of the area. The pixel points contained in the edge are fitted with the circle to obtain the goodness of fit R.2 , the radius of the fitted circle is recorded as R. When the goodness of fit is greater than a1 (a1 is a constant coefficient, which serves as a threshold, and the empirical value is 0.74), the area is considered to be a knot, and the following analysis is continued to determine the specific classification of the knot. Otherwise, it is directly determined that the area does not correspond to a knot and no further analysis is performed. Since there may be interference such as wood grain on the log board, it may also be identified as a knot type to be determined by the above process, but wood grain is generally not circular, so a threshold is set for the above judgment to eliminate interference such as wood grain.

[0047] 1. Starting from the edge of the area, the area between every two edges in the edge image is formed into a whole, which is called a wood grain ring area. The number of wood grain ring areas is recorded as n1. Since the texture of the live knot is sound and tight, the annual rings are fine and clear, and the number of corresponding wood grain ring areas is large. The tissue corresponding to the dead knot is dead and the texture is loose, so the annual rings are blurred or no longer exist, and the number of corresponding wood grain ring areas is small. Due to the formation mechanism of the dead knot, there is a circle of darker veneer around the knot eye, the color of the veneer changes dramatically with other positions, and the color and texture of different wood grain ring areas are quite different. The live knot does not have this layer of veneer structure, only the color is slightly darker than the center, and the texture of different wood grain ring areas is uniform. Based on this, these differences are evaluated.

[0048] Use the OTSU law to divide the grayscale image of the area to be measured for each tree knot type to obtain the division threshold. Filter out the pixels with grayscale values ​​less than the threshold in each wood grain ring area, obtain the proportion of these pixels in each wood grain ring area, select the wood grain ring area with the highest proportion, record it as the veneer wood grain ring area, record the mean grayscale value of the pixels in this area as g1, and record the mean grayscale value of the pixels in the two wood grain ring areas on both sides of the wood grain ring area as g2 and g3 respectively (if there are other wood grain ring areas on only one side of the wood grain ring area, g2 and g3 correspond to the mean grayscale value of the pixels in the wood grain ring area). Fit each pixel in the veneer wood grain ring area with a circle, and record the radius of the fitted circle as r. The number of pixels in the wood grain annular area is n2. Take each pixel in the wood grain annular area and calculate the Euclidean distance with each pixel in the edge of the area. The minimum value of these distances is recorded as the distance from the pixel to the edge of the area. The distances from all pixels in the area to the edge of the area are recorded as The dark veneer significance coefficient dwv is constructed using the above related indicators.

[0049]

[0050] In the formula, min{} is a value-taking function, which is used to take the minimum value in the brackets.

[0051] The more obvious the difference between the grayscale value of the pixel point in the veneer wood grain ring area and the grayscale value in the adjacent wood grain ring area (that is, the more obvious the existence of the veneer wood grain ring area is), the closer the veneer wood grain ring area is to the edge of the area to be determined for the knot type, the larger the dark veneer significance coefficient dwv corresponding to the area to be determined for the knot type, that is, the veneer wood grain ring area is more likely to correspond to a dead knot.

[0052] 2. (1) The texture of live knots is hard and firm, with cells closely arranged and relatively fine; while the texture of dead knots is loose and the interior is relatively rough. Analyze each wood grain ring area separately, take each pixel point in the area, record the grayscale value corresponding to the pixel point as h, and record the grayscale values ​​of the corresponding pixels in the four neighborhoods of the pixel point as h1, h2, h3, and h4 respectively. The maximum grayscale value of the pixel points in the wood grain ring area is recorded as h max , the minimum gray value is recorded as h min , the number of pixels is recorded as n3. The looseness of the ring texture ltk is constructed based on the above related indicators.

[0053]

[0054] When the gray value of each pixel in the wood grain annular area is closer to that of the pixels in the four neighborhoods, the looseness of the ring texture ltk corresponding to the wood grain annular area is smaller, and the structure in the area is more compact.

[0055] (2) The looseness of the ring texture rtl is:

[0056]

[0057] Where rtl is the looseness of the ring texture, and each tree knot type to be measured area corresponds to a looseness of the ring texture. The sum in the formula is the sum of all the looseness of the ring texture in the area to be measured for the tree knot type.

[0058] The smaller the ring texture looseness ltk corresponding to each wood grain annular area in the knot type to be determined area is, the smaller the ring texture looseness rtl corresponding to the knot type to be determined area is, that is, the area is more likely to correspond to a live knot.

[0059] 3. The dark veneer significance coefficient DWV is based on the unique feature of the dark veneer structure around the dead knot, and judges the obviousness of this feature in the area to be tested of the knot type; the looseness of the ring texture RTL is based on the ring-shaped wood grain of the annual rings in the knot, and the fineness of the texture in each wood grain ring area is measured according to the looseness of the ring texture, and then the fineness of the texture in the entire knot type to be tested area is obtained; the significance of decay influence SSE is constructed based on the two.

[0060] sse=dwv×rtl

[0061] When the dark veneer significance coefficient dwv is larger and the ring texture looseness rtl is larger, the decay influence significance sse corresponding to the area to be tested of the knot type is larger, and the area is more likely to be a dead knot.

[0062] At this point, a corresponding decay significance level SSE can be obtained for each location where a knot appears on the door panel.

[0063] Step 3: Determine the defect of the door panel according to the significance of the decay effect corresponding to each knot area.

[0064] The decay significance sse corresponding to each tree knot is obtained by using the above process for multiple (experience value is 50) door panels with dead knots and live knots, and the types of these tree knots (including dead knots and live knots) are manually classified. The median of the minimum decay significance corresponding to the dead knot and the maximum decay significance corresponding to the live knot is taken as the threshold t for automatically classifying the tree knot types. In each subsequent automatic judgment process, if the decay significance corresponding to the obtained tree knot is substituted into the above rule and the threshold t changes, the value corresponding to the threshold is continuously updated according to the newly obtained data to ensure the accuracy of the threshold selection.

[0065] If there is an area identified as a dead knot in the door panel, the door panel is marked as "there is a dead knot and needs to be removed", and the position corresponding to the dead knot and the mark are output; if only live knots are identified in the same door panel, but the number of live knots is greater than or equal to t1 (t1 is the threshold coefficient, and the empirical value is 6), the door panel is marked as "there are too many live knots and a specific solution needs to be determined", and the position corresponding to each live knot and the mark are output (if the above two situations are identified in the same door panel at the same time, the marks corresponding to both situations need to be output); if only live knots are identified in the same door panel, but the number of live knots is less than t1, the door panel is marked as "the quality is within the qualified range" and the mark is output.

[0066] The present invention also provides a defect assessment system for door panel wood, comprising a memory and a processor, wherein the processor is used to execute a method embodiment of the method for implementing the defect assessment method for door panel wood stored in the memory. Since the method embodiment of the method for defect assessment method for door panel wood has been introduced above, it will not be repeated here.

[0067] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for evaluating defects in door panel wood, characterized in that: The following steps are involved: Step 1: Obtain a door panel image and preprocess the image to obtain a grayscale image; Step 2: Segment the grayscale image according to the generation process of the tree knot and its impact on the board to obtain the tree knot type to be determined area; use the Canny operator to obtain the edge image of each tree knot type to be determined area, and perform circle fitting on the edge pixels of the outermost circle in the edge image. When the goodness of fit is greater than the set value, the tree knot type to be determined area is considered to be a tree knot area; The tree knot area is divided to obtain a division threshold, gray values ​​less than the division threshold are screened out, the proportion of pixels in each tree knot area whose gray values ​​are less than the division threshold is determined, the highest proportion is selected as the veneer wood grain annular area, and the dark veneer significance coefficient of the veneer wood grain annular area is calculated; The looseness of the ring texture is calculated based on the annular area of ​​the wood grain of the veneer, and then the looseness of the ring texture is obtained; The significance of the decay effect is obtained according to the significance coefficient of the dark veneer and the looseness of the ring texture; Step 3: Determine the defect of the door panel according to the significance of the decay effect; The method for obtaining the area to be determined for the type of tree knot is as follows: the grayscale image is segmented using the OTSU law to obtain the part of the image that is suspected to be a tree knot; the grayscale values ​​corresponding to all the pixels in the part suspected to be a tree knot are assigned to 1, and the grayscale values ​​corresponding to the remaining pixels are assigned to 0 to obtain a binary image; the binary image is processed by opening operation and area filling; the pixels corresponding to the value of 1 in the obtained image are marked, and the corresponding pixels are obtained in the grayscale image according to the positions of the pixels, and the area composed of these pixels is the area to be determined for the type of tree knot; The significant coefficient of the dark veneer is: ； in, The significant coefficient of dark wood veneer is is the mean gray value of the pixels in the wood grain annular area of ​​the veneer, is the average grayscale value of the pixel points in the wood grain annular area on one side adjacent to the wood grain annular area; is the mean gray value of the pixel points in the wood grain annular area on the other side of the wood grain annular area, r is the radius of the circle obtained by fitting the pixels in the wood grain annular area of ​​the veneer with the circle, R is the radius of the circle fitting of the pixel points at the outermost edge of the edge image, , , …, are the distances from all pixels in the wood grain annular area to the edge of the area, is the number of wood grain annular areas, is the number of pixels in the wood grain annular area of ​​the veneer, and min() is the minimum value function; The looseness of the basic construction ring texture is: ； in, The looseness of the ring texture is constructed based on h, and h is the gray value corresponding to any pixel point in the ring area of ​​the veneer wood grain. , , , are the grayscale values ​​of the corresponding pixels in the four neighborhoods of any pixel, are the maximum and minimum grayscale values ​​of the pixels in the wood grain annular area, is the number of pixels in the wood grain annular area of ​​the veneer; The looseness of the ring texture is the average looseness of the basic construction of the ring texture of all the knot areas; The significance of the decay effect is the product of the significance coefficient of the dark veneer and the looseness of the ring texture.

2. A defect assessment system for door panel wood, characterized in that: It comprises a memory and a processor, wherein the processor is used to execute the steps stored in the memory to implement the defect assessment method for door panel wood according to claim 1.

Images