CN105954301A - Bottleneck quality detection method based on machine vision - Google Patents

Abstract

The invention discloses a machine vision-based bottle mouth quality inspection method, comprising the following steps: collecting the image of the bottle mouth to be inspected and converting it into a grayscale image; calculating the gradient vector of each pixel in the grayscale image to obtain the Detect the edge image of the bottle mouth grayscale image; segment the edge according to the gray value threshold range; use the area as a feature to segment the inner and outer rings of the bottle mouth to be detected; calculate the center coordinates and radii of the inner and outer rings respectively, Calculate the average value of the center coordinates to obtain the center coordinates of the bottle mouth to be detected, and set the radius value range according to the outer and inner ring radii; obtain the parameter equation of the circle according to the center coordinates of the bottle mouth to be detected and the radius value range, The parametric equation scans the ring circularly, calculates the average gray value, and draws the average gray value curve; analyzes the average gray value curve, and if the change range of the ring is within a certain range, it is determined that the ring is not damaged. The invention improves the efficiency of bottle mouth quality detection.

Description

A method of bottle mouth quality inspection based on machine vision

technical field

The invention relates to the technical field of quality detection, in particular to a machine vision-based bottle finish quality detection method.

Background technique

Wine, beverage, medicine, food and other manufacturing industries have adopted a large number of filling production lines in production, and most of them use glass bottles as product packaging. However, glass bottles are inevitably polluted and damaged during production and transportation, especially in industries such as beer that need to use recyclable glass bottles. Therefore, glass bottles must go through cleaning, testing and other processes before entering the filling process. In order to overcome the harm caused by foreign matter and damage, it is necessary to carry out meticulous inspection on the glass bottle before filling, which is called full bottle inspection in the industry. This detection is usually performed manually in a dark room under light.

The visual detection robot mainly uses the theory and technology of machine vision to detect the quality of empty bottles on the filling production line. As a comprehensive frontier subject, machine vision has attracted widespread attention in recent years, and it is one of the research hotspots, and its research and application are quite active.

Existing bottleneck detection can only rely on manual detection, which not only has low efficiency, but also has low precision, which seriously affects the efficiency of the production line.

Contents of the invention

In view of this, the purpose of the present invention is to overcome the above-mentioned defects of the prior art, provide a kind of bottle mouth quality detection method based on machine vision, solve the problem of relying on manual detection accuracy and low efficiency in the past, and improve the detection efficiency of the domestic manufacturing industry. technical content.

The present invention solves the above problems by the following technical means:

A method for detecting the quality of bottle finish based on machine vision, comprising the steps of:

S1. Collect the image of the bottle mouth to be detected and convert it into a grayscale image;

S2, using the Sobel edge detection operator to calculate the gradient vector of each pixel of the grayscale image, to obtain the edge image of the grayscale image of the bottle mouth to be detected;

S3, setting the threshold range of the gray value, and segmenting the edges according to the threshold range of the gray value;

S4, according to the segmented area, use the area as a feature to segment the inner ring and outer ring of the bottle mouth to be detected;

S5. Use the center of gravity method to calculate the center coordinates and radii of the inner ring and the outer ring respectively, average the center coordinates of the outer ring and the inner ring to obtain the center coordinates of the bottle mouth to be detected, and set the radius according to the radius of the outer ring and the inner ring. range of values;

S6. Obtain the parametric equation of the circle according to the coordinates of the center of the bottle mouth to be detected and the value range of the radius, scan the circle according to the parametric equation of the circle, calculate the average gray value, and draw the average gray value curve;

S7. Analyze the average gray value curve. If the variation range of the ring is within a certain range, it is determined that the ring is not damaged; otherwise, it is determined that the ring is damaged, and it is classified as an unqualified product.

Further, in step S1, the image of the bottle mouth to be detected is collected by using the lighting detection system. The lighting detection system includes the bottle mouth to be detected, an LED light source, a baffle, and a CCD industrial camera, and the baffle corresponds to the bottle mouth to be detected. There is an opening at the center, and the LED light source irradiates the light to the mouth of the bottle to be inspected. After reflection, it enters the CCD industrial camera through the opening to obtain the image of the mouth of the bottle to be inspected.

Further, in step S2, the specific method of calculating the gradient vector of each pixel of the grayscale image using the Sobel edge detection operator is as follows:

The Sobel edge detection operator contains two sets of 3*3 matrices, which are horizontal convolution factors and vertical convolution factor

Let A represent the original image, Gx and Gy represent the gray value of the image detected by horizontal and vertical edges respectively, and the formula is as follows:

${\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; A</span>}$

${\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; A</span>}$

The gradient value of each pixel of the image is calculated by the following formula:

Among them, an approximation without the square root is used for efficiency:

|G| = |G _x | + |G _y |,

Calculate the gradient direction with the following formula:

$\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; r</span>}\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; t</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; no</span>}\frac{{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}{{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}}$

If the above angle θ is equal to zero, it means that the image has a vertical edge at this point, and the left side is darker than the right side.

Further, in step S3, the gray value threshold range is (25, 255).

Further, in step S4, a threshold is set, and if the area is greater than the threshold, it is considered that there is a crack at the mouth of the bottle to be inspected, and it is directly judged as a failure.

Further, the threshold is 60000.

Further, in step S5, the specific process is as follows:

Calculate the smallest rectangle circumscribed by the outer ring to obtain the coordinates of the four points above, below, left, and right on the outer ring, and then calculate the midpoint, which is the coordinates of the center of the outer ring, by the center of gravity method:

The formula for calculating the radius of the outer ring:

Wherein, x ₂ is the abscissa of the right point of the outer ring, x ₁ is the abscissa of the left point of the outer ring, y ₂ is the ordinate of the lower edge of the outer ring, and y ₁ is the ordinate of the upper edge of the outer ring;

According to the above method, the coordinates of the center of the inner ring can be obtained in the same way: (center _{X (inner)} , center _{Y (inner)} ) and _inner ring radius R;

Calculate the average value of the center coordinates of the outer ring and the inner ring to obtain the center coordinates of the bottle mouth to be detected: (center _X , center _Y );

Set the radius value range as R according to the outer and inner ring radii.

Further, in step S6, the parametric equation of the circle is:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; center</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&psi;</span>}\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; center</span>}}_{\mathrm{<span\; class="notranslate">\; Y</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&psi;</span>}\end{array}\right.$

Among them, R is [R _inner –3, R _outer +3], and the central angle ψ is [0,360°];

I(x,y) is the gray value corresponding to the coordinate point of the ring (x,y), then:

mean _I is the average gray value.

Further, in step S7, if the variation range of the ring is within 40 pixels, it is determined that the ring is not damaged.

The bottle mouth quality detection method based on machine vision of the present invention solves the problems of low accuracy and efficiency of relying on manual detection in the past, and improves the technical content of detection in the domestic manufacturing industry. At the same time, the successful implementation of this method has good application prospects in the fields of industrial robot hand-eye system, logistics transportation, packaging industry, optical detection and processing.

Description of drawings

Fig. 1 is the flowchart of the bottleneck quality detection method based on machine vision of the present invention;

Fig. 2 is the structural representation of the lighting detection system adopted in the bottle finish quality detection method based on machine vision of the present invention;

Fig. 3 adopts the grayscale image of the bottle mouth collected in Fig. 2 in the embodiment of the present invention;

FIG. 4 is an average grayscale graph corresponding to each bottle mouth grayscale image in FIG. 3 .

detailed description

In order to make the above objects, features and advantages of the present invention more comprehensible, the technical solutions of the present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be pointed out that the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all those skilled in the art can obtain all without creative work. Other embodiments all belong to the protection scope of the present invention.

As shown in Figure 1, a machine vision-based bottle mouth quality detection method includes the following steps:

S1. Collect the image of the bottle mouth to be detected and convert it into a grayscale image;

As shown in Figure 2, the image of the bottle mouth to be detected is collected using the lighting detection system. There is an opening, and the LED light source irradiates the light to the mouth of the bottle to be inspected. After reflection, it enters the CCD industrial camera through the opening, and the image of the mouth of the bottle to be inspected is obtained and converted into a grayscale image, as shown in Figure 3.

S2, using the Sobel edge detection operator to calculate the gradient vector of each pixel of the grayscale image, to obtain the edge image of the grayscale image of the bottle mouth to be detected;

Technically, it is a discrete difference operator, which is used to calculate the approximate value of the gray level of the image brightness function. Using this operator at any point of the image will generate the corresponding grayscale vector or its normal vector;

The Sobel edge detection operator contains two sets of 3*3 matrices, which are horizontal convolution factors and vertical convolution factor Convolve it with the image plane to obtain the approximate value of the horizontal and vertical brightness difference respectively;

If A represents the original image, Gx and Gy represent the gray value of the image detected by horizontal and vertical edges respectively, the formula is as follows:

${\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; A</span>}$

${\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; A</span>}$

The gradient value of each pixel of the image is calculated by the following formula: Usually, an approximation without the square root is used for efficiency:

|G|=|G _x +G _y |,

The gradient direction can then be calculated using the following formula:

$\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; r</span>}\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; t</span>}\mathrm{<span\; class="notranslate">\; a</span>}\mathrm{<span\; class="notranslate">\; no</span>}\frac{{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}{{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}}$

If the above angle θ is equal to zero, it means that the image has a vertical edge, and the left side is darker than the right side.

S3, setting the threshold range of the gray value, and segmenting the edges according to the threshold range of the gray value;

The gray value threshold range is (25, 255).

S4, according to the segmented area, use the area as a feature to segment the inner ring and outer ring of the bottle mouth to be detected;

According to the segmented area, the area is used as a feature to segment the inner and outer rings of the bottle mouth. Generally speaking, under the same lighting detection system for the same batch of bottles, the area of the ring has a fixed range, that is, the area of the edge also has a fixed range. When there is a crack at the mouth of the bottle, the inner ring and the outer ring after edge segmentation will be connected together, as can be seen from b in Figure 3; here you can set a threshold to judge, if the area is greater than the threshold, it will be regarded as a crack , can be directly judged as failing. According to the experimental data, the area threshold is set to 60000 here.

S5. Use the center of gravity method to calculate the center coordinates and radii of the inner ring and the outer ring respectively, average the center coordinates of the outer ring and the inner ring to obtain the center coordinates of the bottle mouth to be detected, and set the radius according to the radius of the outer ring and the inner ring. range of values;

After extracting the edge of the ring, calculate the smallest rectangle circumscribing the outer ring to obtain the coordinates of the upper, lower, left, and right points of the outer ring, and then use the center of gravity method to calculate the midpoint, which is the coordinates of the center of the outer ring:

The formula for calculating the radius of the outer ring:

Wherein, x ₂ is the abscissa of the right point of the outer ring, x ₁ is the abscissa of the left point of the outer ring, y ₂ is the ordinate of the lower edge of the outer ring, and y ₁ is the ordinate of the upper edge of the outer ring;

According to the above method, the coordinates of the center of the inner ring can be obtained in the same way: (center _{X (inner)} , center _{Y (inner)} ) and _inner ring radius R;

Calculate the average value of the center coordinates of the outer ring and the inner ring to obtain the center coordinates of the bottle mouth to be detected: (center _X , center _Y );

Set the radius value range as R according to the outer and inner ring radii.

S6. Obtain the parametric equation of the circle according to the coordinates of the center of the bottle mouth to be detected and the value range of the radius, scan the circle according to the parametric equation of the circle, calculate the average gray value, and draw the average gray value curve;

Since the image of the bottle mouth can be roughly regarded as a standard ring, the circle is scanned according to the parameter equation of the circle, the average gray value is calculated, and the average gray value curve is drawn, as shown in Figure 4.

The parametric equation of a circle is:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; center</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; cos</span>}\mathrm{<span\; class="notranslate">\; \&psi;</span>}\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; center</span>}}_{\mathrm{<span\; class="notranslate">\; Y</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; sin</span>}\mathrm{<span\; class="notranslate">\; \&psi;</span>}\end{array}\right.$

Among them, R is (R _inside –3, R _outside +3), and the central angle ψ is (0,360°);

I(x,y) is the gray value corresponding to the coordinate point of the ring (x,y), then:

mean _I is the average gray value.

S7. Analyze the average gray value curve. If the variation range of the ring is within a certain range, it is determined that the ring is not damaged; otherwise, it is determined that the ring is damaged, and it is classified as an unqualified product.

If the variation range of the ring is within 40 pixels, it is determined that the ring is not damaged.

A in Figure 3 is a standard donut diagram. It can be seen from its average grayscale curve in Fig. 4 a that its variation range is 16 pixels, which is a qualified product; while the other four donut diagrams The range of change is relatively large. It can be seen that the rings in those pictures are all damaged, so the corresponding bottle mouths can be listed as unqualified products.

The bottle mouth quality detection method based on machine vision of the present invention solves the problems of low accuracy and efficiency of relying on manual detection in the past, and improves the technical content of detection in the domestic manufacturing industry. At the same time, the successful implementation of this method has good application prospects in the fields of industrial robot hand-eye system, logistics transportation, packaging industry, optical detection and processing.

The above-mentioned embodiments only express several implementation modes of the present invention, and the descriptions thereof are relatively specific and detailed, but should not be construed as limiting the patent scope of the present invention. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present invention, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention should be based on the appended claims.

Claims (9)

1. a bottle finish quality detection method based on machine vision, is characterized in that, comprises the steps: S1. Collect the image of the bottle mouth to be detected and convert it into a grayscale image; S2, using the Sobel edge detection operator to calculate the gradient vector of each pixel of the grayscale image, to obtain the edge image of the grayscale image of the bottle mouth to be detected; S3, setting the threshold range of the gray value, and segmenting the edges according to the threshold range of the gray value; S4, according to the segmented area, use the area as a feature to segment the inner ring and outer ring of the bottle mouth to be detected; S5. Use the center of gravity method to calculate the center coordinates and radii of the inner ring and the outer ring respectively, average the center coordinates of the outer ring and the inner ring to obtain the center coordinates of the bottle mouth to be detected, and set the radius according to the radius of the outer ring and the inner ring. range of values; S6. Obtain the parametric equation of the circle according to the coordinates of the center of the bottle mouth to be detected and the value range of the radius, scan the circle according to the parametric equation of the circle, calculate the average gray value, and draw the average gray value curve; S7. Analyze the average gray value curve. If the variation range of the ring is within a certain range, it is determined that the ring is not damaged; otherwise, it is determined that the ring is damaged, and it is classified as an unqualified product. 2. the bottle finish quality detection method based on machine vision according to claim 1, is characterized in that, in step S1, utilizes the image of the bottle finish to be detected to be collected by an illumination detection system, and said illumination detection system comprises a bottleneck to be detected, LED light source, baffle, and CCD industrial camera, the baffle corresponds to an opening at the mouth of the bottle to be detected, the LED light source irradiates the light to the mouth of the bottle to be detected, and enters the CCD industrial camera through the opening through reflection to obtain the image of the bottle mouth to be detected image. 3. the bottleneck quality detection method based on machine vision according to claim 2, is characterized in that, in step S2, utilizes the gradient vector concrete method of each pixel point of Sobel edge detection operator to calculate gray scale image as follows: The Sobel edge detection operator contains two sets of 3*3 matrices, which are horizontal convolution factors and vertical convolution factor Let A represent the original image, Gx and Gy represent the gray value of the image detected by horizontal and vertical edges respectively, and the formula is as follows: ${\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; A</span>}$ ${\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; 1</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 2</span>}& <span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}\end{array}\right]<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; A</span>}$ The gray value of each pixel of the image is calculated by the following formula: Among them, an approximation without the square root is used for efficiency: |G|=|G _x +G _y |, Calculate the gradient direction with the following formula: $\mathrm{<span\; class="notranslate">\; \&theta;</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; arctan</span>}\frac{{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}{{\mathrm{<span\; class="notranslate">\; G</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}}$ If the above angle θ is equal to zero, it means that the image has a vertical edge, and the left side is darker than the right side. 4. The machine vision-based bottle finish quality detection method according to claim 3, characterized in that, in step S3, the gray value threshold range is (25, 255). 5. the bottleneck quality detection method based on machine vision according to claim 4, is characterized in that, in step S4, set a threshold value, if area is greater than threshold value, it is deemed that there is a crack in the bottleneck to be detected, and it is directly judged as not pass. 6 . The machine vision-based bottle finish quality detection method according to claim 5 , wherein the threshold is 60,000. 7. the bottleneck quality detection method based on machine vision according to claim 6, is characterized in that, in step S5, concrete process is as follows: Calculate the smallest rectangle circumscribed by the outer ring to obtain the coordinates of the four points above, below, left, and right on the outer ring, and then calculate the midpoint, which is the coordinates of the center of the outer ring, by the center of gravity method: The formula for calculating the radius of the outer ring: Wherein, x ₂ is the abscissa of the right point of the outer ring, x ₁ is the abscissa of the left point of the outer ring, y ₂ is the ordinate of the lower edge of the outer ring, and y ₁ is the ordinate of the upper edge of the outer ring; According to the above method, the coordinates of the center of the inner ring can be obtained in the same way: (center _{X (inner)} , center _{Y (inner)} ) and _inner ring radius R; Calculate the average value of the center coordinates of the outer ring and the inner ring to obtain the center coordinates of the bottle mouth to be detected: (center _X , center _Y ); Set the radius value range as R according to the outer and inner ring radii. 8. the bottleneck quality detection method based on machine vision according to claim 7, is characterized in that, in step S6, the parametric equation of circle is: $\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; x</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; center</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; c</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; \&psi;</span>}\\ \mathrm{<span\; class="notranslate">\; the\; y</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; center</span>}}_{\mathrm{<span\; class="notranslate">\; Y</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; *</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; no</span>}\mathrm{<span\; class="notranslate">\; \&psi;</span>}\end{array}\right.$ Among them, R is (R _inside –3, R _outside +3), and the central angle ψ is (0,360°); I(x,y) is the gray value corresponding to the coordinate point of the ring (x,y), then: mean _I is the average gray value. 9. The machine vision-based bottleneck quality detection method according to claim 8, characterized in that, in step S7, if the variation range of the ring is within 40 pixels, it is determined that the ring is not damaged.