CN110607405A - Leather inner contour recognition cutting device and method based on machine vision industrial application - Google Patents

Abstract

The invention discloses a leather inner contour recognition cutting device and method based on machine vision industrial application. Meanwhile, the method for identifying and cutting the inner contour of the leather comprises the following steps: (1) collecting leather images: the leather is conveyed to a detection station by a conveyor belt, and the image acquisition is carried out by a camera on the detection station. (2) Leather image processing: segmenting a leather image, carrying out image smoothing treatment and HSV channel separation, and finally extracting an inner contour (3) through a threshold segmentation algorithm based on region growth to cut the leather: and cutting the inner contour of the leather through coordinate system conversion according to the extracted inner contour edge. The invention solves the practical problems of high cost, low detection efficiency and the like of the existing manual detection, enhances the production automation degree and greatly improves the production efficiency and the market competitiveness.

Description

Leather inner contour recognition cutting device and method based on machine vision industrial application

Technical Field

The invention relates to the technical field of leather identification and measurement, in particular to a device and a method for identifying and cutting an inner contour of leather based on machine vision industrial application.

Background

At present, leather is an indispensable daily necessity in daily life of people, and the demand volume is on a trend of rising year by year. The traditional method for cutting the inner contour of the leather mainly utilizes the visual and subjective judgment ability of people, adopts a manual cutting method to cut, is easily influenced by factors such as the vision, emotion, fatigue, light and the like of the people, and has high cost and common effect. The difficulty of detecting the inner contour of the leather lies in extracting the inner contour leather from the background of the ground color leather. Therefore, the accurate and quick leather inner contour recognition cutting device and method have great engineering application value.

Accordingly, further improvements and improvements are needed in the art.

Disclosure of Invention

The invention aims to overcome the defects of the prior art and provides a device and a method for recognizing and cutting an inner leather contour based on machine vision industrial application, which have high automation degree and high efficiency.

The purpose of the invention is realized by the following technical scheme:

the recognition cutting device mainly comprises a rack, a cutting device for cutting leather, a conveying belt for conveying the leather to the cutting device, a transmission device for driving the conveying belt to rotate, an image acquisition device for acquiring leather data on the conveying belt, light source equipment for polishing/supplementing light, an industrial computer for analyzing the image data and a touch screen for inputting and adjusting system parameters.

Specifically, the conveyor belt is horizontally arranged on the rack and connected with the transmission device. The transmission device is arranged on the rack and drives the conveying belt to rotate. The image acquisition device is fixed above the rack and positioned at one end of the conveyor belt, and the acquisition end of the image acquisition device is installed downwards. The light source equipment comprises four lighting lamps and a bracket. The support is fixed on the frame, and the light-emitting lamps are arranged on the support and distributed around the image acquisition device. The cutting device is arranged on the rack, is positioned at the other end of the conveying belt and comprises a cutting module for cutting leather and a servo moving module for driving the cutting module to do linear reciprocating motion. And two ends of the servo moving module are fixedly connected with the racks on two sides of the conveyor belt. The cutting module is arranged on the servo moving module, and the servo moving module drives the cutting module to transversely move on the conveyor belt. The touch screen is fixed on the frame. The industrial computer is electrically connected with the transmission device, the image acquisition device, the lighting lamp, the servo moving module and the cutting module respectively.

As a preferable scheme of the invention, the touch screen adopts an industrial display screen.

In a preferred embodiment of the present invention, the image capturing device is a camera.

The other purpose of the invention is realized by the following technical scheme:

the method for recognizing and cutting the inner contour of the leather based on the machine vision industrial application mainly comprises the following specific steps:

step S1: collecting leather images: the conveyer belt is sent the leather product to assigned position department, and the camera on the image acquisition device carries out image acquisition to the leather product, polishes from four directions through four light of polishing to acquire the leather image.

Step S2: leather image segmentation: separating the leather ImageS in HSV space to obtain ImageS ImageH, ImageS and ImageV on each channel; selecting an ImageS with better display effect, and separating the leather from the background by adopting a fixed threshold segmentation method based on gray level characteristics to obtain a leather region image ROI _ 1.

Specifically, the concept of the morphological dilation algorithm in step S2 is as follows:

the expansion operation is one of the bases of morphological treatment, and mainly plays a role in connecting cracks in a target, and the operation formula is as follows:

in the formula, A represents an original binary image; b represents a morphologically processed structural element;a map showing B with respect to its origin, and x showing the amount of displacement;

as can be seen from equation (2), the expansion of A by B is the set of all displacements x that do not null at the intersection with A, so that at least one element of B and A overlaps; therefore, the dilation process can be viewed as a convolution process, except that dilation is based on set operations, convolution is based on arithmetic operations, but the processing of the two is similar; the process of expansion is as follows:

a. scanning each pixel of image a with structuring element B;

b. carrying out AND operation on the structural elements and the binary image covered by the structural elements;

c. if both are 0, the pixel of the resulting image is 0, otherwise it is 1.

Step S3: and (3) image smoothing processing: and smoothing the image by adopting a mean filtering and noise reduction algorithm to eliminate the noise in the detected area.

Step S4: detecting the inner contour of the leather: separating the leather region image ROI _1 in an HSV space to obtain ImageH, ImageS and ImageV of the ROI _1 image on each channel; and selecting ImageS with better display effect, and segmenting by adopting a threshold segmentation method based on region growth so as to obtain the inner contour XLD _1 of the leather.

Specifically, the idea of the region growing algorithm in step S4 is as follows:

the basic formula for region growing is:

let R denote the entire image region, then the segmentation can be seen as a process of dividing the region R into n connected sub-regions R1, R2.. Rn, and the following conditions need to be satisfied:

the steps for realizing the region growing are as follows:

(1) sequentially scanning the image, finding the 1 st pixel which is not yet attributed, and setting the pixel as (x0, y 0);

(2) considering the 8 neighborhood pixels (x, y) of (x0, y0) centered at (x0, y0), if (x, y) satisfies the growth criteria, (x, y) is merged (within the same region) with (x0, y0) while (x, y) is pushed onto the stack;

(3) taking a pixel from the stack and returning it as (x0, y0) to step (2);

(4) when the stack is empty, returning to the step (1);

(5) and (5) repeating the steps (1) to (4) until each point in the image has attribution, and ending the growth.

Step S5: cutting leather: generating a DXF file according to the extracted inner contour of the leather, and then generating a path coordinate point through the DXF file; the DXL file generated by the image can be synchronous with the coordinate system of the leather cutting machine only by external calibration, so that external two-dimensional hand-eye calibration is carried out firstly, the leather is cut by synchronizing the cutting coordinate of the air knife and the coordinate of the vision system after the calibration is finished, and the leather is taken away by a conveyor belt after the cutting is finished.

As a preferred embodiment of the present invention, the idea of the threshold segmentation algorithm in step S2 and step S4 is:

the threshold segmentation operation may be defined as shown in equation (1):

S＝{(r,c)∈K|g_min≤g(r,c)≤g_max} (1)

in the formula, K represents a coordinate set of an input image; (r, c) represents the coordinates of a certain pixel; g (r, c) represents the gray value of the pixel; gmin and gmax represent the maximum threshold and minimum threshold, respectively;

in the threshold segmentation, the gray value in the image is in a certain designated gray value range (0, 2)^bAll points within-1) that satisfy equation (1) are selected into the output region S, where b is the bit depth, serving as a threshold adjustment.

As a preferred embodiment of the present invention, the idea of the HSV separation algorithm in step S4 is as follows:

the RGB and CMY color models are both hardware-oriented, while the HSV (hue validation value) color model is user-oriented, the basic formula for RGB to HSV conversion is:

R'＝R/255

G'＝G/255

B'＝B/255

Cmax＝max(R',G',B')

Cmin＝min(R',G',B')

△＝Cmax-Cmin (4)

color tone:

saturation degree:

lightness:

V＝C max (7)

the three-dimensional representation of the HSV model evolves from the RGB cube, where the hexagonal shape of the cube is seen looking from the white to the black vertices of RGB along the diagonal of the cube, the hexagonal boundaries representing colors, the horizontal axis representing purities, and the lightness measured along the vertical axis.

Compared with the prior art, the invention also has the following advantages:

(1) the device and the method for recognizing and cutting the inner leather contour based on the machine vision industrial application adopt the advanced machine vision technology and the leather cutting robot, and can accurately and quickly detect, recognize and cut the inner leather contour. The device solves the technical problems of low judgment efficiency, high cost and the like in the traditional manual cutting, and is successfully applied to a leather cutting production line. The invention also designs a system hardware platform aiming at the detection and identification of the inner contour of the leather in a maximum optimization way according to different characteristics of the leather.

(2) The invention provides a leather inner contour recognition cutting device and method based on machine vision industrial application, and provides an algorithm based on threshold segmentation, HSV separation and region growth, wherein the algorithm can recognize the leather inner contour and carry out edge extraction. The algorithm has the advantages of high calculation speed, strong environmental adaptability and high detection and identification rate, and can effectively solve the problem of detection and identification of the inner contour of the leather.

(3) According to the device and the method for recognizing and cutting the inner contour of the leather based on the machine vision industrial application, a device for cutting the leather on a plane is developed according to the mode of recognizing the inner contour of the leather, and the device and the method can accurately cut according to the extracted edge of the inner contour so as to obtain the inner contour of the leather.

Drawings

Fig. 1 is a front view of an inner profile recognition cutting device for leather based on machine vision industrial application provided by the present invention.

Fig. 2 is a perspective view of the inner contour recognition cutting device for leather based on machine vision industrial application provided by the present invention.

Fig. 3 is a schematic view of the inner profile of the leather provided by the present invention.

FIG. 4 is a schematic diagram of the result of the inner contour detection of leather provided by the present invention.

The reference numerals in the above figures illustrate:

the method comprises the following steps of 1-an image acquisition device, 2-light source equipment, 3-a transmission device, 4-a servo moving module, 5-a cutting module, 6-a conveyor belt and 7-a touch screen.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer and clearer, the present invention is further described below with reference to the accompanying drawings and examples.

Example 1:

as shown in fig. 1 and 2, the present embodiment discloses an inner leather contour recognition and cutting device based on machine vision industrial application, which mainly comprises a frame, a cutting device for cutting leather, a conveyor belt 6 for conveying leather to the cutting device, a transmission device 3 for driving the conveyor belt 6 to rotate, an image acquisition device 1 for acquiring leather data on the conveyor belt 6, a light source device 2 for lighting/light supplementing, an industrial computer for analyzing image data, and a touch screen 7 for inputting and adjusting system parameters.

Specifically, the conveyor belt 6 is horizontally arranged on the frame and connected with the transmission device 3. The transmission device 3 is arranged on the frame and drives the conveyor belt 6 to rotate. The image acquisition device 1 is fixed above the frame and positioned at one end of the conveyor belt 6, and the acquisition end of the image acquisition device is installed downwards. The light source device 2 comprises four floodlights and a support. The support is fixed on the frame, and the light-emitting lamps are arranged on the support and distributed around the image acquisition device 1. The cutting device is arranged on the rack, is positioned at the other end of the conveyor belt 6 and comprises a cutting module 5 for cutting leather and a servo moving module 4 for driving the cutting module 5 to do linear reciprocating motion. And two ends of the servo moving module 4 are fixedly connected with the racks on two sides of the conveyor belt 6. The cutting module 5 is mounted on the servo moving module 4, and the servo moving module 4 drives the cutting module 5 to move transversely on the conveyor belt 6. The touch screen 7 is fixed on the frame. The industrial computer is respectively and electrically connected with the transmission device 3, the image acquisition device 1, the lighting lamp, the servo moving module 4 and the cutting module 5.

As a preferred embodiment of the present invention, the touch screen 7 is an industrial display screen.

In a preferred embodiment of the present invention, the image capturing device 1 is a camera.

With reference to fig. 3 and 4, the present embodiment further discloses a method for identifying and cutting an inner contour of leather based on machine vision industrial application, which mainly includes the following specific steps:

step S1: collecting leather images: the conveyer belt 6 sends the leather product to assigned position department, and the camera on the image acquisition device 1 carries out image acquisition to the leather product, polishes from four directions through four lights of polishing to acquire the leather image.

Step S2: leather image segmentation: separating the leather ImageS in HSV space to obtain ImageS ImageH, ImageS and ImageV on each channel; selecting an ImageS with better display effect, and separating the leather from the background by adopting a fixed threshold segmentation method based on gray level characteristics to obtain a leather region image ROI _ 1.

Specifically, the concept of the morphological dilation algorithm in step S2 is as follows:

the expansion operation is one of the bases of morphological treatment, and mainly plays a role in connecting cracks in a target, and the operation formula is as follows:

in the formula, A represents an original binary image; b represents a morphologically processed structural element;a map showing B with respect to its origin, and x showing the amount of displacement;

as can be seen from equation (2), the expansion of A by B is the set of all displacements x that do not null at the intersection with A, so that at least one element of B and A overlaps; therefore, the dilation process can be viewed as a convolution process, except that dilation is based on set operations, convolution is based on arithmetic operations, but the processing of the two is similar; the process of expansion is as follows:

a. scanning each pixel of image a with structuring element B;

b. carrying out AND operation on the structural elements and the binary image covered by the structural elements;

c. if both are 0, the pixel of the resulting image is 0, otherwise it is 1.

Step S3: and (3) image smoothing processing: and smoothing the image by adopting a mean filtering and noise reduction algorithm to eliminate the noise in the detected area.

Step S4: detecting the inner contour of the leather: separating the leather region image ROI _1 in an HSV space to obtain ImageH, ImageS and ImageV of the ROI _1 image on each channel; and selecting ImageS with better display effect, and segmenting by adopting a threshold segmentation method based on region growth so as to obtain the inner contour XLD _1 of the leather.

Specifically, the idea of the region growing algorithm in step S4 is as follows:

the basic formula for region growing is:

let R denote the entire image region, then the segmentation can be seen as a process of dividing the region R into n connected sub-regions R1, R2.. Rn, and the following conditions need to be satisfied:

the steps for realizing the region growing are as follows:

(1) sequentially scanning the image, finding the 1 st pixel which is not yet attributed, and setting the pixel as (x0, y 0);

(2) considering the 8 neighborhood pixels (x, y) of (x0, y0) centered at (x0, y0), if (x, y) satisfies the growth criteria, (x, y) is merged (within the same region) with (x0, y0) while (x, y) is pushed onto the stack;

(3) taking a pixel from the stack and returning it as (x0, y0) to step (2);

(4) when the stack is empty, returning to the step (1);

(5) and (5) repeating the steps (1) to (4) until each point in the image has attribution, and ending the growth.

Step S5: cutting leather: generating a DXF file according to the extracted inner contour of the leather, and then generating a path coordinate point through the DXF file; because the DXL file generated by the image can be synchronous with the coordinate system of the leather cutting machine only by external calibration, the external two-dimensional hand-eye calibration is carried out firstly, the leather is cut by the coordinate of synchronous air knife cutting and the coordinate of the vision system after the calibration is finished, and the leather is taken away by the conveyor belt 6 after the cutting is finished.

As a preferred embodiment of the present invention, the idea of the threshold segmentation algorithm in step S2 and step S4 is:

the threshold segmentation operation may be defined as shown in equation (1):

S＝{(r,c)∈K|g_min≤g(r,c)≤g_max} (1)

in the formula, K represents a coordinate set of an input image; (r, c) represents the coordinates of a certain pixel; g (r, c) represents the gray value of the pixel; gmin and gmax represent the maximum threshold and minimum threshold, respectively;

in the threshold segmentation, the gray value in the image is in a certain designated gray value range (0, 2)^bAll points within-1) that satisfy equation (1) are selected into the output region S, where b is the bit depth, serving as a threshold adjustment.

As a preferred embodiment of the present invention, the idea of the HSV separation algorithm in step S4 is as follows:

the RGB and CMY color models are both hardware-oriented, while the HSV (hue validation value) color model is user-oriented, the basic formula for RGB to HSV conversion is:

R'＝R/255

G'＝G/255

B'＝B/255

Cmax＝max(R',G',B')

Cmin＝min(R',G',B')

△＝C max-C min (4)

color tone:

saturation degree:

lightness:

V＝C max (7)

the three-dimensional representation of the HSV model evolves from the RGB cube, where the hexagonal shape of the cube is seen looking from the white to the black vertices of RGB along the diagonal of the cube, the hexagonal boundaries representing colors, the horizontal axis representing purities, and the lightness measured along the vertical axis.

The above embodiments are preferred embodiments of the present invention, but the present invention is not limited to the above embodiments, and any other changes, modifications, substitutions, combinations, and simplifications which do not depart from the spirit and principle of the present invention should be construed as equivalents thereof, and all such changes, modifications, substitutions, combinations, and simplifications are intended to be included in the scope of the present invention.

Claims (8)

1. The device is characterized by comprising a rack, a cutting device for cutting leather, a conveying belt for conveying the leather to the cutting device, a transmission device for driving the conveying belt to rotate, an image acquisition device for acquiring leather data on the conveying belt, light source equipment for polishing/supplementing light, an industrial computer for analyzing the image data and a touch screen for inputting and adjusting system parameters;

the conveying belt is horizontally arranged on the rack and is connected with the transmission device; the transmission device is arranged on the rack and drives the conveying belt to rotate; the image acquisition device is fixed above the rack and positioned at one end of the conveyor belt, and the acquisition end of the image acquisition device is installed downwards; the light source equipment comprises four lighting lamps and a bracket; the bracket is fixed on the frame, and the lighting lamps are arranged on the bracket and distributed around the image acquisition device; the cutting device is arranged on the rack, is positioned at the other end of the conveying belt, and comprises a cutting module for cutting leather and a servo moving module for driving the cutting module to do linear reciprocating motion; two ends of the servo moving module are fixedly connected with the racks on two sides of the conveyor belt; the cutting module is arranged on the servo moving module, and the servo moving module drives the cutting module to transversely move on the conveyor belt; the touch screen is fixed on the rack; the industrial computer is electrically connected with the transmission device, the image acquisition device, the lighting lamp, the servo moving module and the cutting module respectively.

2. The inner leather contour recognition cutting device based on machine vision industrial application as claimed in claim 1, characterized in that the touch screen is an industrial grade display screen.

3. The apparatus for recognizing and cutting the inner contour of leather based on machine vision industrial application as claimed in claim 1, wherein said image capturing means is provided as a camera.

4. The method for recognizing and cutting the inner contour of the leather based on the machine vision industrial application is characterized by comprising the following steps of:

step S1: collecting leather images: the leather product is conveyed to a designated position by a conveying belt, a camera on the image acquisition device acquires images of the leather product, and the leather product is polished from four directions by four light-emitting lamps, so that a leather image is acquired;

step S2: leather image segmentation: separating the leather ImageS in HSV space to obtain ImageS ImageH, ImageS and ImageV on each channel; selecting an ImageS with better display effect, and separating the leather from the background by adopting a fixed threshold segmentation method based on gray level characteristics to obtain a leather region image ROI _ 1;

step S3: and (3) image smoothing processing: smoothing the image by adopting a mean filtering noise reduction algorithm to eliminate the noise in the detected area;

step S4: detecting the inner contour of the leather: separating the leather region image ROI _1 in an HSV space to obtain ImageH, ImageS and ImageV of the ROI _1 image on each channel; selecting ImageS with better display effect, and adopting a threshold segmentation method based on region growth to segment so as to obtain an inner contour XLD _1 of the leather;

step S5: cutting leather: generating a DXF file according to the extracted inner contour of the leather, and then generating a path coordinate point through the DXF file; the DXL file generated by the image can be synchronous with the coordinate system of the leather cutting machine only by external calibration, so that external two-dimensional hand-eye calibration is carried out firstly, the leather is cut by synchronizing the cutting coordinate of the air knife and the coordinate of the vision system after the calibration is finished, and the leather is taken away by a conveyor belt after the cutting is finished.

5. The apparatus for identifying and cutting inner contour of leather based on machine vision industrial application as claimed in claim 4, wherein the idea of the threshold segmentation algorithm of step S2 and step S4 is as follows:

the threshold segmentation operation may be defined as shown in equation (1):

S＝{(r,c)∈K|g_min≤g(r,c)≤g_max} (1)

in the formula, K represents a coordinate set of an input image; (r, c) represents the coordinates of a certain pixel; g (r, c) represents the gray value of the pixel; gmin and gmax represent the maximum threshold and minimum threshold, respectively;

in the threshold segmentation, the gray value in the image is in a certain designated gray value range (0, 2)^bAll points within-1) that satisfy equation (1) are selected into the output region S, where b is the bit depth, serving as a threshold adjustment.

6. The apparatus for recognizing and cutting inner contour of leather based on machine vision industrial application as claimed in claim 4, wherein the concept of morphological dilation algorithm in step S2 is:

the expansion operation is one of the bases of morphological treatment, and mainly plays a role in connecting cracks in a target, and the operation formula is as follows:

in the formula, A represents an original binary image; b represents a morphologically processed structural element;a map showing B with respect to its origin, and x showing the amount of displacement;

as can be seen from equation (2), the expansion of A by B is the set of all displacements x that do not null at the intersection with A, so that at least one element of B and A overlaps; therefore, the dilation process can be viewed as a convolution process, except that dilation is based on set operations, convolution is based on arithmetic operations, but the processing of the two is similar; the process of expansion is as follows:

a. scanning each pixel of image a with structuring element B;

b. carrying out AND operation on the structural elements and the binary image covered by the structural elements;

c. if both are 0, the pixel of the resulting image is 0, otherwise it is 1.

7. The apparatus for recognizing and cutting inner contour of leather based on machine vision industrial application as claimed in claim 4, wherein the idea of the area growing algorithm in said step S4 is:

the basic formula for region growing is:

let R denote the entire image region, then the segmentation can be seen as a process of dividing the region R into n connected sub-regions R1, R2.. Rn, and the following conditions need to be satisfied:

the steps for realizing the region growing are as follows:

(1) sequentially scanning the image, finding the 1 st pixel which is not yet attributed, and setting the pixel as (x0, y 0);

(2) considering the 8 neighborhood pixels (x, y) of (x0, y0) centered at (x0, y0), if (x, y) satisfies the growth criteria, (x, y) is merged (within the same region) with (x0, y0) while (x, y) is pushed onto the stack;

(3) taking a pixel from the stack and returning it as (x0, y0) to step (2);

(4) when the stack is empty, returning to the step (1);

(5) and (5) repeating the steps (1) to (4) until each point in the image has attribution, and ending the growth.

8. The inside contour recognition cutting device for leather based on machine vision industrial application as claimed in claim 4, wherein the idea of HSV separation algorithm in step S4 is as follows:

the RGB and CMY color models are both hardware-oriented, while the HSV (hue validation value) color model is user-oriented, the basic formula for RGB to HSV conversion is:

R'＝R/255

G'＝G/255

B'＝B/255

Cmax＝max(R',G',B')

Cmin＝min(R',G',B')

△＝Cmax-Cmin (4)

color tone:

saturation degree:

lightness:

V＝Cmax (7)

the three-dimensional representation of the HSV model evolves from the RGB cube, where the hexagonal shape of the cube is seen looking from the white to the black vertices of RGB along the diagonal of the cube, the hexagonal boundaries representing colors, the horizontal axis representing purities, and the lightness measured along the vertical axis.