A method of classifying the quality of a dermal surface and related apparatus

Abstract

The application provides a dermal surface quality grading method and related equipment, the method comprising: acquiring image data of a target hide and dividing a main hide area of the target hide; determining a correction parameter based on location information of the target hide; correcting the main hide area according to the correction parameter to obtain an actual main hide area; calculating the utilization rate of the actual main hide area; and judging the surface quality grade of the target hide based on the utilization rate and a preset dermal surface quality reference table. Instead of the traditional manual dermal surface quality grading process, the application can greatly save labor costs, improve grading stability, standardize, digitize and simplify the traditional process, and improve the overall work efficiency of the blue wet hide grading process.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of dermis grading, in particular to a dermis surface quality grading method and related equipment. BACKGROUND

[0002] In the field of leather processing, the skin peeled off the slaughtered cattle and sheep is generally referred to as fur. Fur is prone to rot, so to prevent it from rotting, losing hair, and decaying pores and surface, it is generally pickled with salt, which is called salted wet skin. Salted wet skin can be stored for 3-5 months. Fur and salted wet skin are collectively referred to as rawhide. Rawhide is treated with lime, sulfide, chlorine powder, nitric acid, and chromium powder for oil removal, fat removal, hair removal, and chromium tanning, and will be blue and have moisture, so it is generally referred to as blue wet skin. It is a semi-finished product in the processing of cattle and sheep skin, but it has not been further processed such as drying, conditioning, and dyeing.

[0003] In the leather production and manufacturing process, blue wet skin is a semi-finished product of leather, and leather raw materials can be stored, transported, and traded for a long time after being made into blue wet skin. Because of the natural growth of the dermis and the damage caused by the processing process, the dermis must be graded according to the number and density of surface damage to determine which type of leather product the blue wet skin is suitable for processing, and different processing techniques are developed according to the final product form. Therefore, different grades determine what products the blue wet skin can make, and also determine the price of the products, so the detection and grading of blue wet skin is particularly important. Currently, the entire leather industry relies on manual detection and grading of blue wet skin, which is labor-intensive, low in efficiency, and unstable in grading results due to the subjectivity of manual work. The same worker may get different results when sorting the same skin multiple times. SUMMARY

[0004] The present application provides a dermis surface quality grading method and related equipment, which can use visual scanning technology to obtain image data of the required dermis, and use deep learning to detect defects. The application creatively uses computer simulation of manual grading methods to segment different types of defects in different positions, and obtains accurate layout and grading results. The existing non-quantifiable manual grading method based on experience is innovated in a quantifiable way.

[0005] The technical solution of the present application is a dermis surface quality grading method, which comprises:

[0006] S101, acquiring image data of a target skin, and dividing a main skin area of the target skin and an outer contour of the target skin;

[0007] S102, determining a correction parameter based on position information of the target skin;

[0008] S103, correcting the main skin area according to the correction parameter and the outer contour to obtain an actual main skin area;

[0009] S104, calculating a utilization rate of the actual main skin area;

[0010] S105, judging a surface quality grade of the target skin material based on the utilization rate and a preset dermis surface quality reference table.

[0011] Optionally, the step of dividing the main skin area of the target skin material includes:

[0012] adopting a simulated manual grading method to divide the target skin material to obtain a main skin area represented by a dashed line grid.

[0013] Optionally, the step of adopting the simulated manual grading method to divide the target skin material to obtain a main skin area represented by a dashed line grid includes:

[0014] determining a number of the dashed line grids;

[0015] adopting the simulated manual grading method to divide the target skin material based on the number of the dashed line grids to obtain a main skin area represented by a dashed line grid.

[0016] Optionally, the step of determining the number of the dashed line grids includes:

[0017] which can be obtained by the following formula:

[0018]

[0019] a number of rows and a number of columns of the main skin area grid, wherein the number of rows of the main skin area grid is denoted as a, the number of columns of the main skin area grid is denoted as β, and the number of square feet of the blue wet hide is denoted as θ.

[0020] Optionally, the step of determining the correction parameter based on the location information of the target skin material includes:

[0021] determining a main skin area width to total width ratio Wp and a main skin area height to total height ratio Hp based on the location information of the target skin material and a preset location-parameter reference table, wherein the preset location-parameter reference table includes different locations of the skin material and corresponding correction parameters thereof, and the correction parameters include the main skin area width to total width ratio and the main skin area height to total height ratio.

[0022] Optionally, the step of correcting the main skin area according to the correction parameter and the outer contour to obtain an actual main skin area includes:

[0023] performing an opening operation on the outer contour to obtain a first target area;

[0024] Performing a closing operation on the initial target region to obtain a second target region;

[0025] Obtaining a circumscribed rectangle of the second target region, a center point, and a height and a width of the circumscribed rectangle;

[0026] Generating a target rectangle based on the center point, the height of the circumscribed rectangle, the width of the circumscribed rectangle, and the correction parameter;

[0027] Obtaining a corrosion contour by performing a corrosion operation on the dermis contour;

[0028] Obtaining an actual main skin area as an intersection of the target rectangle and the corrosion contour.

[0029] Optionally, the step of calculating the utilization rate of the actual main skin area comprises:

[0030] If the area in the dashed grid has defects, the defective degree of the dashed grid is calculated as a1, and if there are no defects in the dashed grid, the defective degree of the grid is calculated as 0;

[0031]

[0032] When the dashed grid is located at the upper left corner, the upper right corner, the lower left corner, and the lower right corner of the target skin, the defective degree of the grid when the area in the dashed grid has defects is a2;

[0033]

[0034] The defective degree is 0 when there are no defects in the area in the dashed grid;

[0035] Assigning a weight value of b1 to a serious defect and a weight value of b2 to a non-serious defect to obtain the utilization rate L of the target skin,

[0036] L = (1 - (b1*c1*a1 + c2*a1 + c3*a2)) * 100%

[0037] Wherein, c1 is the number of serious defect grids other than the four corners, c2 is the number of non-serious defect grids, and c3 is the number of defect grids in the four corners.

[0038] In a second aspect, the present application provides a dermis surface quality grading device, the device comprising:

[0039] A data acquisition module for acquiring image data of a target skin and dividing a main skin area of the target skin and an outer contour of the target skin;

[0040] A determination module for determining a correction parameter based on position information of the target skin;

[0041] a correction module configured to correct the main skin area according to the correction parameter and the outer contour to obtain an actual main skin area;

[0042] a calculation module configured to calculate a utilization rate of the actual main skin area;

[0043] a judgment module configured to judge a surface quality grade of the target skin material based on the utilization rate and a preset dermal surface quality reference table.

[0044] In a third aspect, the present application provides an electronic device, comprising a processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, and the machine readable instructions are executed by the processor to perform the steps of the dermal surface quality grading method as described above.

[0045] In a fourth aspect, the present application provides a computer readable storage medium, the computer readable storage medium stores a computer program, and the computer program is executed by the processor to perform the steps of the dermal surface quality grading method as described above.

[0046] The image data of the required dermis can be obtained by using visual scanning technology, and defects can be detected by using deep learning. The computer simulated artificial grading method is used to segment different types of defects in different positions to obtain accurate layout and grading results. The existing experience-based non-quantifiable artificial grading method is innovated in a quantifiable way. BRIEF DESCRIPTION OF DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the present application, the drawings required in the embodiments will be briefly introduced below. Obviously, other drawings can also be obtained by those skilled in the art without creative labor.

[0048] Figure 1 A flow chart of a dermal surface quality grading method in an embodiment of the present application;

[0049] Figure 2 A scene diagram of a dermal surface quality grading method in an embodiment of the present application;

[0050] Figure 3 A scene diagram of a dermal surface quality grading method in an embodiment of the present application;

[0051] Figure 4 A scene diagram of a dermal surface quality grading method in an embodiment of the present application;

[0052] Figure 5A scene diagram of a method for grading a leather surface quality according to an embodiment of the present application;

[0053] Figure 6 A scene diagram of a method for grading a leather surface quality according to an embodiment of the present application;

[0054] Figure 7 A scene diagram of a method for grading a leather surface quality according to an embodiment of the present application;

[0055] Figure 8 A scene diagram of a method for grading a leather surface quality according to an embodiment of the present application;

[0056] Figure 9 A scene diagram of a method for grading a leather surface quality according to an embodiment of the present application. DETAILED DESCRIPTION

[0057] The embodiments will be described in detail below with reference to the drawings. When the description below refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following embodiments do not represent all implementations consistent with the present application. Rather, they are merely examples of systems and methods consistent with some aspects of the present application as detailed in the appended claims.

[0058] As shown in Figure 1 , the present application provides a method for grading a leather surface quality, comprising:

[0059] acquiring image data of a target hide and dividing a main hide area of the target hide and an outer contour of the target hide;

[0060] determining a correction parameter based on location information of the target hide;

[0061] correcting the main hide area according to the correction parameter and the outer contour to obtain an actual main hide area;

[0062] calculating a utilization rate of the actual main hide area;

[0063] judging a surface quality grade of the target hide based on the utilization rate and a preset leather surface quality reference table.

[0064] Optionally, the step of dividing the main hide area of the target hide comprises:

[0065] dividing the target hide by using a simulated manual grading method to obtain a main hide area represented by a dashed grid.

[0066] Optionally, the step of dividing the target hide by using a simulated manual grading method to obtain a main hide area represented by a dashed grid comprises:

[0067] determining the number of the dashed-line grids;

[0068] dividing the target hide based on the number of the dashed-line grids by using a simulated manual grading method to obtain a main hide area represented by the dashed-line grids.

[0069] Optionally, the step of determining the number of the dashed-line grids comprises:

[0070] The number of the dashed-line grids can be obtained by the following formula:

[0071]

[0072] the number of grid rows and the number of grid columns of the main hide area, wherein, the number of grid rows of the main hide area is denoted as a, the number of grid columns is denoted as b, and the square footage of the blue wet hide is denoted as q.

[0073] Optionally, the step of determining the correction parameter based on the location information of the target hide comprises:

[0074] determining a ratio of the main hide area width to the total width Wp and a ratio of the main hide area height to the total height Hp based on the location information of the target hide and a preset location-parameter comparison table, wherein, the preset location-parameter comparison table comprises different locations of hides and corresponding correction parameters thereof, and the correction parameters comprise the ratio of the main hide area width to the total width and the ratio of the main hide area height to the total height.

[0075] Optionally, the step of correcting the main hide area based on the correction parameter and the outer contour to obtain an actual main hide area comprises:

[0076] performing an opening operation on the outer contour to obtain a first target area;

[0077] performing a closing operation on the initial target area to obtain a second target area;

[0078] obtaining a circumscribed rectangle of the second target area, a center point, a height of the circumscribed rectangle, and a width of the circumscribed rectangle;

[0079] generating a target rectangle based on the center point, the height of the circumscribed rectangle, the width of the circumscribed rectangle, and the correction parameter;

[0080] obtaining an erosion contour by performing an erosion operation on the dermis contour;

[0081] obtaining an intersection of the target rectangle and the erosion contour as the actual main hide area.

[0082] Optionally, the step of calculating the utilization rate of the actual main hide area comprises:

[0083] If the area in the dotted grid has defects, the defective degree of the dotted grid is calculated as a1, and if the area in the dotted grid has no defects, the defective degree of the grid is calculated as 0;

[0084]

[0085] When the dotted grid is located at the upper left corner, the upper right corner, the lower left corner and the lower right corner of the target leather, the defective degree of the grid is a2 when the area in the dotted grid has defects;

[0086]

[0087] The defective degree is 0 when the area in the dotted grid has no defects;

[0088] The weight value of the serious defect is b1, and the weight value of the non-serious defect is b2, and the utilization rate L of the target leather is obtained,

[0089] L = (1-(b1*c1*a1+c2*a1+c3*a2))*100%

[0090] Wherein, c1 is the number of serious defect grids other than the four corners, c2 is the number of non-serious defect grids, and c3 is the number of defect grids in the four corners.

[0091] In a second aspect, the present application provides a dermal surface quality grading device, which comprises:

[0092] A data acquisition module is configured to acquire image data of a target leather and divide a main skin area of the target leather and an outer contour of the target leather;

[0093] A determination module is configured to determine a correction parameter based on position information of the target leather;

[0094] A correction module is configured to correct the main skin area according to the correction parameter and the outer contour to obtain an actual main skin area;

[0095] A calculation module is configured to calculate a utilization rate of the actual main skin area;

[0096] A judgment module is configured to judge a surface quality grade of the target leather based on the utilization rate and a preset dermal surface quality reference table.

[0097] In a third aspect, the present application provides an electronic device, which comprises a processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, and the machine readable instructions are executed by the processor to perform the steps of the dermal surface quality grading method as described above.

[0098] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the dermal surface quality grading method described above.

[0099] For example, taking wet blue sheepskin as an example, visual inspection technology (which is not limited to image processing technologies such as line scan cameras and area scan cameras) can be used to obtain... Figure 2 The image data shows the sheepskin outline with thick solid lines, defects with thin solid lines, and a dashed grid representing the main skin area divided using a simulated manual grading method. Figure 2 The main leather area shown is located vertically between the front and hind legs of the sheepskin, and the left and right sides are automatically generated using a computer algorithm based on the sheepskin's outline after being shrunk inwards. In actual production, sheepskins vary in shape and have different edge textures, so the main leather area must be adjusted accordingly. Figure 3 After obtaining the sheepskin outline P1, first perform an opening operation on P1 to remove the more protruding parts of P1 to obtain P2. Figure 4 (As shown); then perform a closing operation on P2 to fill in the concave part of P2 to obtain P3 (as shown); Figure 5 (As shown). Then in Figure 6 The bounding rectangle R1 of P3 is visible, with its center point C1. The height H and width W of R1 are obtained. The parameters are set as follows: the ratio of the main skin area width to the total width is Wp, and the ratio of the main skin area height to the total height is Hp. Then the main skin area width is Wz = W * Wp; the main skin area height is Hz = H * Hp. A rectangle Rz is generated through the center point C1, the main skin area width Wz, and the main skin area height Hz. Figure 7 (As shown). At this point, the rectangular main skin area Rz may be outside the sheepskin outline or may include the sheepskin trimming area. The eroded outline P4 is obtained by performing an erosion operation on the sheepskin outline. Figure 8 (Middle dashed outline), erosion value K. Find the intersection M of Rz and P4, i.e. Figure 9 The shaded area shown is the actual main skin area. In actual production, the size of the main skin area is modified by adjusting the values ​​of Wp, Hp, and K according to the actual requirements of the main skin area.

[0100] Because the area outside the main hide zone has a very low utilization rate during leather processing, and trimming is required before processing the hide, this grading method is only used within the main hide zone. Since sheepskin sizes range from 4 to 12 square feet, the number of dotted grid lines dividing the main hide zone also varies. Let α be the number of grid rows in the main hide zone, β be the number of grid columns, and θ be the sheepskin size in square feet. The following formula can be used to obtain:

[0101]

[0102] For example, the size of the grid of the sheepskin of 5.5 square feet is 4 rows and 4 columns; the size of the grid of the sheepskin of 7 square feet is 5 rows and 4 columns; the size of the grid of the sheepskin of 9 square feet is 6 rows and 5 columns.

[0103] In other application scenarios, such as the shoe uppers, luggage, cars, sofas and other industries involved in the wet blue of cowhide, the number of grids in the main skin area can define the values of a and b according to the size of the cutting piece of different products, and the case of sofa skin is more special. Since the cutting piece area used by the sofa skin is large and is a complete 80*80cm square, a=2 and b=2 can be set. In the above other scenarios, the length of the cutting piece is X and the width is Y. According to the following formula, the following formula can be obtained:

[0104]

[0105] As shown in Figure 2 If there is any defect in a grid, the defective degree of the grid is calculated as a1,

[0106]

[0107] If there is no defect in a grid, the defective degree of the grid is calculated as 0. The upper left, upper right, lower left and lower right corners are close to the leg area, so the four corners are less used in product manufacturing. Therefore, when the grid of the four corners has a defect, the defective degree of the grid is a2, When there is no defect, the defective degree is 0, which can exclude the influence of defects in unimportant areas on grading.

[0108] According to existing experience, the influence of different types of defects on grading is different. Deep learning is used to classify defects into serious defects and non-serious defects. Serious defects include open wounds, pinholes, burr holes, rotten surfaces and moss scars; non-serious defects include scratches and healed wounds. According to the grading requirements, the weight of serious defects is b1, and b1 can be assigned a value of 1-2 before the number, such as 1.2, 1.5, etc., according to the type of product in the later processing stage and customer requirements. The weight of non-serious defects is b2, and b2=1. In this way, the utilization rate L of each sheepskin can be obtained according to the number of grids occupied by serious defects c1, the number of grids occupied by non-serious defects c2, and the number of grids occupied by defects in the four corners c3,

[0109] L=(1-(b1*c1*a1+c2*a1+c3*a2))*100%

[0110] Generally, sheepskin blue wet skin is divided into A, B, C, D four grades, according to different skin can be given to each level of different threshold, for example, commonly used domestic skin A level of utilization is 85%-100%, B level of utilization is 70%-85%, C level of utilization is 45%-75%, D level of utilization is 0-45%. Thus according to the different L value computer can quickly and accurately divided into each level of skin. In actual production, often according to the different requirements of the final product to some special defects skin for downgrading processing. For example Figure 2 There are long long scratch, large area of open wound, etc. Therefore, some special rules can be added: 1, the number of broken surface injury more than X, the utilization rate is reduced w; 2, the area of broken surface injury is greater than Y, the grid weight factor is A; 3, the perimeter of broken surface injury is Z, the grid weight factor is B. A = 3, B = 2 can be set. The values of X, Y and Z can be set according to different product requirements.

[0111] The above detailed description of the embodiments of the present application, but the content is only the preferred embodiments of the present application, can not be considered for limiting the scope of the present application. Any equivalent changes and improvements made within the scope of the present application, should still belong to the patent scope of the present application.

Claims

1. A method of classifying the quality of a dermal surface, characterized in that, The method comprises the following steps: acquiring image data of a target hide and dividing a main hide area of the target hide and an outer contour of the target hide; determining a correction parameter based on location information of the target hide; the step of determining the correction parameter based on the location information of the target hide comprises: determining a main hide area width to total width ratio Wp and a main hide area height to total height ratio Hp based on the location information of the target hide and a preset location-parameter comparison table, wherein the preset location-parameter comparison table comprises different locations of hides and corresponding correction parameters thereof, and the correction parameters comprise a main hide area width to total width ratio and a main hide area height to total height ratio; correcting the main hide area according to the correction parameter and the outer contour to obtain an actual main hide area; the step of correcting the main hide area according to the correction parameter and the outer contour to obtain an actual main hide area comprises: performing an opening operation on the outer contour to obtain a first target area; performing a closing operation on the initial target area to obtain a second target area; acquiring an outer rectangle, a center point of the second target area, a height of the outer rectangle and a width of the outer rectangle; generating a target rectangle based on the center point, the height of the outer rectangle, the width of the outer rectangle and the correction parameter; obtaining an erosion contour by performing an erosion operation on the dermis contour; obtaining an intersection of the target rectangle and the erosion contour as the actual main hide area; calculating a utilization rate of the actual main hide area; judging a surface quality grade of the target hide based on the utilization rate and a preset dermis surface quality comparison table.

2. The dermal surface quality grading method of claim 1, wherein, the step of dividing the main hide area of the target hide comprises: dividing the target hide by using a simulated artificial grading method to obtain a main hide area represented by a dashed line grid.

3. The dermal surface quality grading method of claim 2, wherein, the step of dividing the target hide by using the simulated artificial grading method to obtain a main hide area represented by a dashed line grid comprises: determining a number of the dashed line grid; dividing the target hide by using the simulated artificial grading method based on the number of the dashed line grid to obtain a main hide area represented by a dashed line grid.

4. The dermal surface quality grading method of claim 2, wherein, the step of determining the number of the dashed line grid comprises: the number of the dashed line grid can be obtained by the following formula: a number of rows of the main hide area grid and a number of columns of the main hide area grid, wherein the number of rows of the main hide area grid is denoted as a, the number of columns of the main hide area grid is denoted as b, and a square foot number of the dermis size is denoted as c.

5. The dermal surface quality grading method of claim 2, wherein, the step of calculating the utilization rate of the actual main hide area comprises: if there is a defect in the area in the dashed line grid, a defective degree of the dashed line grid is calculated as a1, and if there is no defect in the area in the dashed line grid, the defective degree of the grid is calculated as 0; when the dashed line grid is located at the upper left corner, the upper right corner, the lower left corner and the lower right corner of the target hide, if the area in the dashed line grid has a defect, a defective degree of the grid is a2; if the area in the dashed line grid has no defect, the defective degree is 0; a weight value of a serious defect is b1, a weight value of a non-serious defect is b2, and a utilization rate L of the target hide is obtained, L=(1-(b1*c1*a1+b2*c2*a1+c3*a2))*100% Wherein, c1 is the number of grids occupied by the serious defects except for the four corners, c2 is the number of grids occupied by the non-serious defects, and c3 is the number of grids occupied by the defects in the four corners.

6. A dermal surface quality grading device obtained by the method of claim 1, wherein, The device comprises: a data acquisition module, configured to acquire image data of a target skin material, and divide a main skin area of the target skin material and an outer contour of the target skin material; a determination module, configured to determine a correction parameter based on position information of the target skin material; a correction module, configured to correct the main skin area based on the correction parameter and the outer contour, to obtain an actual main skin area; a calculation module, configured to calculate a utilization rate of the actual main skin area; a judgment module, configured to judge a surface quality grade of the target skin material based on the utilization rate and a preset dermal surface quality reference table.

7. An electronic device, comprising: comprise: a processor, a memory and a bus, the memory stores machine readable instructions executable by the processor, when the electronic device is running, the processor and the memory communicate through the bus, the machine readable instructions are executed by the processor to execute the steps of the dermal surface quality grading method in any one of claims 1 to 5.

8. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is executed by the processor to execute the steps of the dermal surface quality grading method in any one of claims 1 to 5.

Images