Hippophae rhamnoides fruit screening equipment based on machine vision

Abstract

The invention relates to sea buckthorn fruit screening equipment based on machine vision, which comprises a conveying table, a first conveying belt and a second conveying belt which are wound on the conveying table, screening holes which are uniformly distributed in the second conveying belt, an image acquisition and analysis module which is arranged on the conveying table and faces the screening holes in the second conveying belt, and a rejector which is arranged on the conveying table, the image acquisition and analysis module is used for acquiring detected surfaces in the screening holes and analyzing the detected surfaces to find detected surface defects, the rejector is used for rejecting unqualified detected surfaces falling from the second conveying belt according to feedback of the image acquisition and analysis module, and the moving directions of the first conveying belt and the second conveying belt are opposite; according to the sea buckthorn fruit screening equipment based on machine vision, a detected complete surface image is obtained in an active driving rotation mode, then scars are determined by means of an automatic analysis mode, the scars are removed according to the scars, and the quality of detected products can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of automated screening technology, and in particular to a sea buckthorn fruit screening device based on machine vision. Background Art

[0002] Sea buckthorn is widely used in the field of sand control because of its characteristics such as cold resistance, drought resistance, saline-alkali resistance and economic value, because it can make sand control projects have economic attributes, because sea buckthorn fruit can be made into economic commodities (food, drinks, medicinal materials).

[0003] The production process is roughly divided into harvesting, cleaning, screening and grading, and processing. Currently, most harvesting methods are carried out in a low-temperature environment. The purpose is to use low-temperature freezing to reduce the difficulty of harvesting and ensure the integrity of the fruit. Cleaning and screening and grading are also mechanically automated with the help of bubble cleaning machines, vibrating screening machines, and air separation machines.

[0004] However, after screening and grading, the bad sea buckthorn fruits need to be removed. This is mainly because the skin of sea buckthorn fruit is thin and easily injured by pressing and squeezing. These types of scars are not obvious. If they are sold directly as commodities or used for processed foods, they will directly affect the final product quality and price. Summary of the invention

[0005] The present application provides a sea buckthorn fruit screening device based on machine vision, which obtains a complete surface image of the tested product by actively driving the rotation, and then uses an automated analysis method to determine the scars and remove them accordingly, which can effectively improve the quality of the tested product.

[0006] The above-mentioned purpose of the present application is achieved through the following technical solutions: The present application provides a sea buckthorn fruit screening device based on machine vision, including: Transport platform; The first conveyor belt and the second conveyor belt are both wound on the conveyor platform, and the first conveyor belt is located on the inner side of the second conveyor belt; Screening holes are evenly distributed on the second conveyor belt; An image acquisition and analysis module is arranged on the transport platform and faces the screening hole on the second transport belt. The image acquisition and analysis module is used to acquire the image to be measured in the screening hole and analyze the surface to be measured to find surface defects of the image to be measured. A rejector is arranged on the transport platform, and is used to reject unqualified test objects dropped from the second transport belt according to the feedback of the image acquisition and analysis module; The moving directions of the first conveyor belt and the second conveyor belt are opposite.

[0007] In a possible implementation of the present application, the image acquisition and analysis module discovers that the surface defects being measured include: Collect an image of the coverage area and extract the object to be measured in the image, which is recorded as the image to be measured; Performing grayscale processing on the image to be tested to obtain a grayscale image; Count the distribution of grayscale values ​​in the grayscale image, delete the areas corresponding to grayscale values ​​that exceed the set ratio, and obtain suspected surface defects; Identify the type of suspected surface defects, including adhesion, discoloration, wrinkles and depressions; The suspected surface defects are screened according to the type of the suspected surface defects and the area of ​​the suspected surface defects to obtain the surface defects.

[0008] In a possible implementation of the present application, it also includes: Decomposing the image to be tested into a monochrome image, a red image to be tested, a green image to be tested and a blue image to be tested; Performing grayscale processing on the monochrome image to obtain a red grayscale measured image, a green grayscale measured image, and a blue grayscale measured image; Look for suspected surface defects on the red grayscale test image, the green grayscale test image, and the blue grayscale test image.

[0009] In a possible implementation of the present application, determining the type of suspected surface defects includes: Determine height variations of suspected surface defects; When there is no height change in the suspected surface defect, the type of the suspected surface defect is determined to be discoloration; When the suspected surface defect has a height change and the height change is located outside the suspected surface defect, the type of the suspected surface defect is determined to be adhesion; When the suspected surface defect has a height change and the height change is located inside the suspected surface defect, the type of the suspected surface defect is determined to be a wrinkle; When the suspected surface defect has a height change and an overall color change or the overall color remains unchanged, the type of the suspected surface defect is determined to be a depression.

[0010] In a possible implementation of the present application, determining the height change of the suspected surface defect includes: A tracking point is determined on the suspected surface defect, and the color value of the tracking point is the maximum color value or the minimum color value within the corresponding range of the suspected surface defect; A tracking range is established based on the tracking point, where the tracking point is located at the center of the tracking range; The tracking points are tracked dynamically and the color changes within the tracking range are recorded. If there are color changes within the tracking range, suspected surface defects including height changes are determined.

[0011] In a possible implementation of the present application, the shape of the tracking range is a rectangle; the length direction of the tracking range is consistent with the color change direction of the suspected surface defect.

[0012] In a possible implementation of the present application, determining a tracking point on a suspected surface defect includes: Establish an extraction domain according to the maximum color value or the minimum color value within the corresponding range of the suspected surface defect; Use the extraction domain to extract suspected surface defects and obtain an extraction map; The content on the extracted graph is used as tracking points.

[0013] In a possible implementation of the present application, extracting the content on the graph as a tracking point includes: Create a closed circle so that the content on the extracted image is located inside the closed circle; Uses the edges, center, and middle of the closed circle as tracking points.

[0014] In a possible implementation of the present application, after extracting the contents on the graph as tracking points, the shape of each tracking point is also recorded.

[0015] In a possible implementation of the present application, the tracking points are used to draw a tracking grid, and the number of sides of each grid in the tracking grid is the same.

[0016] The beneficial effects of this application are: The machine vision-based sea buckthorn fruit screening equipment provided in the present application uses a continuous rolling method to obtain the surface image of the test object and then analyzes the obtained surface image. During the analysis process, the defective areas and types of defective areas on the test surface can be determined, and unqualified test objects can be eliminated based on this, so as to obtain the test objects (sea buckthorn fruits) that meet the requirements and have better consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a structural schematic diagram of a sea buckthorn fruit screening device provided by the present application, in which the arrows indicate the moving direction of the measured object.

[0018] Figure 2 This is a schematic diagram of the relative position of a first conveyor belt and a second conveyor belt provided by the present application, in which the arrow indicates the direction of rotation.

[0019] Figure 3 This is a schematic diagram of the distribution of screening holes on a second conveyor belt provided in the present application.

[0020] Figure 4 This is a schematic flowchart of the steps for discovering surface defects provided by the present application.

[0021] Figure 5 This is a schematic diagram of decomposing a measured image provided by the present application.

[0022] Figure 6 It is a schematic diagram of a tracking range provided by this application.

[0023] Figure 7 This is a schematic diagram of the principle of color change within a tracking range provided by the present application.

[0024] In the figure, 1. transport platform, 2. first transport belt, 3. second transport belt, 4. screening hole, 5. image acquisition and analysis module, 6. rejector. DETAILED DESCRIPTION

[0025] The technical solution in this application is further described in detail below in conjunction with the accompanying drawings.

[0026] The machine vision-based sea buckthorn fruit screening equipment disclosed in this application is deployed after the screening and grading process in the production process, with the purpose of removing bad fruits. In combination with the content mentioned in the background technology, the harvested sea buckthorn fruits are further stated to be selected by wind (to remove dust and impurities) and washed with water (to remove surface floating soil), and then enter the screening and grading process. In this process, the sea buckthorn fruits are graded according to diameter by layered vibration, generally graded into three to four grades, and the particle size of the sea buckthorn fruits in each grade is within a set range. However, the grading process cannot remove bad fruits, and the grading process may also cause potential surface damage. Therefore, after the grading process, the bad sea buckthorn fruits need to be removed.

[0027] See also Figure 1 In some examples, the machine vision-based sea buckthorn fruit screening equipment disclosed in the present application includes a transport platform 1, a first conveyor belt 2, a second conveyor belt 3, an image acquisition and analysis module 5 and a rejector 6, the first conveyor belt 2 and the second conveyor belt 3 are both wound around the transport platform 1, and the first conveyor belt 2 is located on the inner side of the second conveyor belt 3.

[0028] The first conveyor belt 2 and the second conveyor belt 3 are both driven by motors. Figure 2 The moving directions of the two conveyor belts are opposite. The purpose of moving in opposite directions is to drive the object to be tested (sea buckthorn fruit) in the screening hole 4 to rotate, because only in this way can the entire surface of the object to be tested be exposed to the field of view of the image acquisition and analysis module 5.

[0029] The upper two sides of the second conveyor belt 3 (close to the two ends of the first conveyor belt 2) are supported by the first conveyor belt 2. The driving motor of the first conveyor belt 2 is located on the left side of the first conveyor belt 2, and the driving motor of the second conveyor belt 3 is located below the middle position of the second conveyor belt 3.

[0030] See also Figure 3 The screening holes 4 are evenly distributed on the second conveyor belt 3 and arranged in a matrix form of MxN, where M and N are both natural numbers greater than zero.

[0031] The image acquisition and analysis module 5 is installed on the transport platform 1, and is used to collect the image to be tested in the screening hole 4 and analyze the surface to be tested to find surface defects of the image to be tested. In some examples, the image acquisition and analysis module 5 is mainly composed of two parts: an image sensor and a graphics processor. The image sensor generally uses a high-speed camera.

[0032] It should be noted that the machine vision-based sea buckthorn fruit screening device disclosed in this application generally needs to be deployed in a closed space, and an artificial light source is deployed in the closed space (to improve the uniformity of light). The first conveyor belt 2 and the second conveyor belt 3 are made of non-reflective materials, and their colors need to be clearly different from the color of the sea buckthorn fruit. Generally, black is used to make the sea buckthorn fruit clearly displayed in the image.

[0033] The rejector 6 is arranged on the transport platform 1. The rejector 6 is used to reject unqualified test objects dropped from the second transport belt 3 according to the feedback of the image acquisition and analysis module 5. The specific implementation method is as follows: The screening holes 4 are arranged in a matrix arrangement, so digital marks can be made on one side of the surface of the second conveyor belt 3. At this time, the row number on the screening hole 4 can be determined, and then the coordinates of the screening hole 4 can be obtained through the column number of the screening hole 4.

[0034] That is, if the fruit in a screening hole 4 is identified as bad fruit, it has a clear number (x, y), refer to Figure 3 The numbers (1,5,10) in the table represent the row numbers. Now, by reading the numbers from left to right or from right to left, we can get (x, y).

[0035] At the same time, by properly controlling the rotation speed of the second conveyor belt 3, a row of sea buckthorn fruits can be dropped in each dropping process, and the rejector 6 at the corresponding column number is started to remove the bad fruit.

[0036] In some examples, the rejector 6 is composed of a cylinder and a swing arm, and the cylinder drives the swing arm to rotate to push the bad fruits away from the predetermined falling trajectory.

[0037] In other examples, the rejector 6 may use a linear motor.

[0038] At the same time, the rejector 6 may be equipped with a camera or a proximity sensor to detect the measured position, so as to give the rejector 6 a clear start time.

[0039] For some examples, see Figure 4The specific steps of the image acquisition and analysis module 5 to find the surface defects under test are as follows: S101, collecting an image of the coverage area and extracting the object to be measured in the image, which is recorded as the image to be measured; S102, performing grayscale processing on the image to be tested to obtain a grayscale image; S103, counting the distribution of grayscale values ​​in the grayscale image, deleting the areas corresponding to grayscale values ​​that account for more than a set ratio, and obtaining suspected surface defects; S104, determine the type of suspected surface defects, including adhesion, discoloration, wrinkles and depressions; S105, screening the suspected surface defects according to the types of the suspected surface defects and the areas of the suspected surface defects to obtain the surface defects.

[0040] In step S101, firstly, an image of the coverage area is acquired and the object to be measured in the image is extracted, which is recorded as the image to be measured. Combining the content in the previous article, it can be seen that because there is an obvious color difference between the object to be measured and the background, extraction can be achieved here only by using the distinguishing color, or it can be described as deleting the conveyor belt in the image of the coverage area acquired.

[0041] In step S102, it is necessary to perform grayscale processing on the image to be measured. The image obtained at this time is called a grayscale image. The purpose of grayscale processing is to reduce the amount of data processing.

[0042] In step S103, the distribution of grayscale values ​​in the grayscale image is counted, and the areas corresponding to grayscale values ​​exceeding a set ratio are deleted to obtain suspected surface defects. The purpose here is to eliminate the normal areas under test.

[0043] It should be noted that the fruits tested in this step mainly have a small amount or even a trace amount of surface defects. Bad fruits and unripe fruits have been removed in the previous processing steps. The fruits in this step have the obvious characteristics of being basically mature and having a uniform surface color.

[0044] Here, the ratio is generally set to 15%-20%, and the area less than this ratio is regarded as a suspected surface defect. It should be noted that for the corresponding area of ​​the suspected surface defect on the grayscale image, the content of the corresponding area is not deleted, but is retained.

[0045] In step S104, it is necessary to determine the type of suspected surface defects. In this application, the types are divided into four categories, namely, attachment, discoloration, wrinkles and depressions. For the sea buckthorn fruits that have been washed and screened, the attachment is mainly black spots, which are formed during the growth of the sea buckthorn fruits and are generally not used as a criterion for quality evaluation. However, for some orders, the number of black spots will be restricted, which is mainly related to the appearance issue.

[0046] Discoloration, wrinkles and depressions are manifestations of damage to the surface of sea buckthorn fruit.

[0047] Finally, in step S105, the suspected surface defects are screened according to the types of the suspected surface defects and the areas of the suspected surface defects to obtain the surface defects.

[0048] For example, two standards are generally used for screening. The first standard is the cumulative value, for example, the area value of the black spot is not allowed to exceed a pre-set value. The second standard is the maximum value, for example, the area value of the largest black spot among all black spots is not allowed to exceed a pre-set value.

[0049] Of course, both standards can be used simultaneously.

[0050] In some examples, the following steps are added: The image to be tested is decomposed into monochrome images, red image to be tested, green image to be tested and blue image to be tested, such as Figure 5 As shown; Performing grayscale processing on the monochrome image to obtain a red grayscale measured image, a green grayscale measured image, and a blue grayscale measured image; Look for suspected surface defects on the red grayscale test image, the green grayscale test image, and the blue grayscale test image.

[0051] The above method is to perform color separation processing on the tested image before grayscale processing is performed. After obtaining the red tested image, the green tested image and the blue tested image, grayscale processing is performed on the red tested image, the green tested image and the blue tested image respectively.

[0052] The advantages of this approach are as follows: The gradient calculation of monochrome images is more direct. For example, the Sobel operator or Canny edge detector can quickly locate the contour of the object and avoid interference between color channels. In addition, the homogeneity of grayscale images (such as pixel brightness consistency) facilitates region-based segmentation algorithms.

[0053] Suspected surface defects are composed of three colors: red, green and blue. The proportions of these three colors are different. For example, in some cases, the display may be more obvious when only one color is used for observation, or a monochrome image can be enhanced to obtain a higher recognition rate.

[0054] In some examples, the type of suspected surface defect is determined as follows: Determine height variations of suspected surface defects; When there is no height change in the suspected surface defect, the type of the suspected surface defect is determined to be discoloration; When the suspected surface defect has a height change and the height change is located outside the suspected surface defect, the type of the suspected surface defect is determined to be adhesion; When the suspected surface defect has a height change and the height change is located inside the suspected surface defect, the type of the suspected surface defect is determined to be a wrinkle; When the suspected surface defect has a height change and an overall color change or the overall color remains unchanged, the type of the suspected surface defect is determined to be a depression.

[0055] In the above method, the type of suspected surface defects is determined by color change and height change. The specific rules are: Discoloration: No height change is present for suspected surface defects; Adhesion: There is a height change of the suspected surface defect, and the height change is located outside the suspected surface defect; Wrinkles: suspected surface defects have height changes, and the height changes are located inside the suspected surface defects; Depression: Suspected surface defect with a change in height, an overall color change, or an overall unchanged color.

[0056] In some examples, the height variation of suspected surface defects is determined by: S201, determining a tracking point on a suspected surface defect, wherein a color value of the tracking point is a maximum color value or a minimum color value within a range corresponding to the suspected surface defect; S202, establishing a tracking range based on the tracking point, where the tracking point is located at the center of the tracking range; S203, dynamically track the tracking point and record the color change within the tracking range. If there is a color change within the tracking range, it is determined that the suspected surface defect includes a height change.

[0057] In step S201 to step S203, it is first necessary to determine a tracking point on the suspected surface defect, and the color value of the tracking point is the maximum color value or the minimum color value within the corresponding range of the suspected surface defect.

[0058] In some possible implementations, the tracking point is a combination of multiple pixel points (MxN matrix).

[0059] See also Figure 6 Then, a tracking range is established based on the tracking point. The tracking point is located at the center of the tracking range. The tracking range generally uses a circular structure. This is because if there is a color change around the tracking point, the direction of the color change is uncertain. The circular structure ensures full coverage of the circumferential direction when the direction cannot be determined.

[0060] Finally, the tracking point is dynamically tracked and the color changes within the tracking range are recorded. If there is a color change within the tracking range, it is determined that the suspected surface defect includes height change. The reason for the change here is that as the measured object rotates, the lighting conditions on its surface will also change.

[0061] It should be clarified here that there may be two situations for color changes within the tracking range. The first situation is that the color change within the tracking range is consistent with the color change outside the tracking range. The second situation is that the color change within the tracking range is inconsistent with the color change outside the tracking range.

[0062] When the second situation occurs, it means that the suspected surface defects include height changes, such as Figure 7 As shown, conversely, the suspected surface defect does not include height changes.

[0063] In some possible implementations, the shape of the tracking range is a rectangle, and the length direction of the tracking range is consistent with the color change direction of the suspected surface defect. That is, after the color of the suspected surface defect begins to change, the color change direction of the suspected surface defect can be determined. At this time, adjusting the shape of the tracking range can, on the one hand, achieve accurate tracking of color changes, and on the other hand, reduce the area of ​​the tracking range, so as to achieve the purpose of reducing data processing volume and improving data processing speed.

[0064] The specific method for determining the tracking point on the suspected surface defect is as follows: S301, establishing an extraction domain according to the maximum color value or the minimum color value within the range corresponding to the suspected surface defect; S302, using the extraction domain to extract the suspected surface defects to obtain an extraction map; S303: extracting the contents on the graph as tracking points.

[0065] In step S301 to step S303, an extraction domain is first established according to the maximum color value or the minimum color value in the corresponding range of the suspected surface defect, and the maximum color value or the minimum color value is located at the midpoint or endpoint in the corresponding extraction domain.

[0066] Then use the extraction domain to extract the suspected surface defects to obtain an extraction map. At this time, the contents obtained are all pixel points that have a relationship with the maximum color value or the minimum color value (extraction domain). Finally, the contents on these extraction maps are used as tracking points.

[0067] Furthermore, the tracking point area was mentioned in the previous article. The content on the extracted image includes three situations: greater than the tracking point area, equal to the tracking point area, and less than the tracking point area. The content corresponding to the area less than the tracking point area is deleted, and the content greater than the tracking point area and equal to the tracking point area is retained. For the content greater than the tracking point area, a part of it is selected at its edge to be used as the tracking point.

[0068] Then proceed as follows: Create a closed circle so that the content on the extracted image is located inside the closed circle; Uses the edges, center, and middle of the closed circle as tracking points.

[0069] The purpose of the above two steps is to ensure the uniformity of the distribution of tracking points. The middle position is generally 40%-70% of the radius of the closed circle.

[0070] If the above method cannot be implemented, the processing is performed by adjusting the extraction domain. The length of the extraction domain is generally controlled at around 15-30. When adjusting, the length of the extraction area can be adjusted, and the position of the maximum color value or the minimum color value in the extraction domain can also be adjusted, for example, from the midpoint to the endpoint or other position.

[0071] In some examples, after extracting the contents on the image as tracking points, the shape of each tracking point is also recorded, in order to facilitate continuous tracking of the tracking points.

[0072] Furthermore, it also includes drawing a tracking grid using the tracking points, and the number of sides of each grid in the tracking grid is the same, and its purpose is also to facilitate continuous tracking of the tracking points.

[0073] It should be understood that the shape and relative position of the tracking point will basically not change during the continuous movement in a small range. In the present application, multiple images that appear sequentially in a time series are processed separately, and after the processing is completed, the matching of the tracking points is involved.

[0074] The matching work of this application is determined by the shape and relative position of the tracking point. When matching, the shape and relative position also need to set an error range. The error range of the relative position is generally 1%-2%, and the error range of the shape is generally 1-2 pixels.

[0075] The embodiments of this specific implementation method are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Therefore, all equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. Sea buckthorn fruit screening equipment based on machine vision, characterized in that: include: Transport platform (1); The first conveyor belt (2) and the second conveyor belt (3) are both wound around the conveyor platform (1), and the first conveyor belt (2) is located on the inner side of the second conveyor belt (3); Screening holes (4) are evenly distributed on the second conveyor belt (3); An image acquisition and analysis module (5) is disposed on the transport platform (1) and faces the screening hole (4) on the second transport belt (3), the image acquisition and analysis module (5) being used to acquire images of the object to be tested in the screening hole (4) and analyze the surface of the object to be tested to find defects on the surface of the object to be tested; A rejector (6) is disposed on the transport platform (1), and is used to reject unqualified objects to be tested that fall from the second transport belt (3) according to feedback from the image acquisition and analysis module (5); The moving directions of the first conveyor belt (2) and the second conveyor belt (3) are opposite.

2. The machine vision-based seabuckthorn fruit screening device according to claim 1, characterized in that: The image acquisition and analysis module (5) finds that the surface defects under test include: Collect an image of the coverage area and extract the object to be measured in the image, which is recorded as the image to be measured; Performing grayscale processing on the image to be tested to obtain a grayscale image; Count the distribution of grayscale values ​​in the grayscale image, delete the areas corresponding to grayscale values ​​that exceed the set ratio, and obtain suspected surface defects; Identify the type of suspected surface defects, including adhesion, discoloration, wrinkles and depressions; The suspected surface defects are screened according to the type of the suspected surface defects and the area of ​​the suspected surface defects to obtain the surface defects.

3. The machine vision-based seabuckthorn fruit screening device according to claim 2, characterized in that: Also includes: Decomposing the image to be tested into a monochrome image, a red image to be tested, a green image to be tested and a blue image to be tested; Performing grayscale processing on the monochrome image to obtain a red grayscale measured image, a green grayscale measured image, and a blue grayscale measured image; Look for suspected surface defects on the red grayscale image, the green grayscale image, and the blue grayscale image.

4. The machine vision-based seabuckthorn fruit screening device according to claim 2, characterized in that: Identify suspected surface defects including: Determine height variations of suspected surface defects; When there is no height change in the suspected surface defect, the type of the suspected surface defect is determined to be discoloration; When the suspected surface defect has a height change and the height change is located outside the suspected surface defect, the type of the suspected surface defect is determined to be adhesion; When the suspected surface defect has a height change and the height change is located inside the suspected surface defect, the type of the suspected surface defect is determined to be a wrinkle; When the suspected surface defect has a height change and an overall color change or the overall color remains unchanged, the type of the suspected surface defect is determined to be a depression.

5. The machine vision-based seabuckthorn fruit screening device according to claim 4, characterized in that: Height changes that identify suspected surface defects include: A tracking point is determined on the suspected surface defect, and the color value of the tracking point is the maximum color value or the minimum color value within the corresponding range of the suspected surface defect; A tracking range is established based on the tracking point, where the tracking point is located at the center of the tracking range; The tracking points are tracked dynamically and the color changes within the tracking range are recorded. If there are color changes within the tracking range, suspected surface defects including height changes are determined.

6. The machine vision-based seabuckthorn fruit screening device according to claim 5, characterized in that: The shape of the tracking range is rectangular; the length direction of the tracking range is consistent with the color change direction of the suspected surface defect.

7. The machine vision-based seabuckthorn fruit screening device according to claim 5, characterized in that: Identifying tracking points on suspected surface defects includes: Establish an extraction domain according to the maximum color value or the minimum color value within the corresponding range of the suspected surface defect; Use the extraction domain to extract suspected surface defects and obtain an extraction map; The content on the extracted graph is used as tracking points.

8. The machine vision-based seabuckthorn fruit screening device according to claim 7, characterized in that: Extracting the contents of the graph as tracking points includes: Create a closed circle so that the content on the extracted image is located inside the closed circle; Uses the edges, center, and middle of the closed circle as tracking points.

9. The machine vision-based seabuckthorn fruit screening device according to claim 7, characterized in that: After extracting the contents on the image as tracking points, the shape of each tracking point is also recorded.

10. The machine vision-based seabuckthorn fruit screening device according to claim 8, characterized in that: It also includes drawing a tracking grid using tracking points, where each grid cell has the same number of sides.

Images