Tea rolling control method based on artificial intelligence

Abstract

The invention discloses a tea rolling control method based on artificial intelligence. The tea rolling control method comprises the following steps: detecting the damping condition of a motor by detecting the feedback current of the grinding disc motor of a rolling machine; the rotating speed of a grinding disc motor is collected; the downward cover plate pressure of the pressure cover plate is directly measured through the pressure sensor; the current feeding rate is calculated according to the collected information; a grinding disc image is collected; performing image processing on the grinding disc image, and calculating to obtain a rolling ratio; calculating rolling efficiency; and adjusting the rolling efficiency and the feeding speed according to the rolling efficiency obtained by calculation. According to the method, visual image processing is introduced, the method capable of effectively evaluating the tea leaf curling degree is designed, the rolling ratio of each tea leaf area on the rolling disc is counted through the method, the gland pressure and real-time damping are obtained through self feedback of a pressure sensor and a rolling disc motor, the current feeding rate is estimated, and the tea leaf curling degree is evaluated. And feeding can be controlled by adjusting a control system of the stepping motor.

Description

Technical Field

[0001] This invention relates to the field of tea processing technology, specifically to a tea rolling control method based on artificial intelligence. Background Technology

[0002] In tea production, some tea leaves need to be rolled into rolls after the initial processing (kill-green) to facilitate further drying. Current tea production has evolved from purely manual rolling to machine rolling. However, because various brands of rolling machines (such as the patented CN112273489A - Tea Rolling Machine) require manual intervention to adjust the rolling speed and feed rate, reliable quality control of the rolled tea leaves is lacking. If the rolling process is too slow, the feed material will be squeezed, interfering with the rolling process. Therefore, rolling efficiency needs to be matched with the millstone speed, and adjustments need to be made according to the actual situation to ensure stable performance of the rolling machine. During the rolling process, the current state of the tea leaves needs to be observed manually to judge the degree of curling, resulting in low rolling efficiency and poor quality. Summary of the Invention

[0003] To address the aforementioned technical problems, the present invention aims to provide an artificial intelligence-based tea rolling control method. This method incorporates a camera to analyze the degree of curling of the tea leaves, and further improves the curling ratio through control to ensure the quality of the tea leaves. It can also simultaneously control the rolling efficiency to achieve a balance between the feeding rate and the rolling efficiency, thereby ensuring the overall production speed.

[0004] This invention provides an artificial intelligence-based method for controlling the kneading of tea leaves, comprising the following steps:

[0005] S1. Detect the damping of the motor by detecting the built-in feedback current of the grinding disc motor of the kneading machine.

[0006] S2. Collect the rotational speed r of the millstone motor;

[0007] S3. Directly measure the downward pressure F of the pressure cover plate using a pressure sensor. i ;

[0008] S4. Calculate the current feed rate J based on the collected information. i ;

[0009] S5. Acquire images of the millstone;

[0010] S6. Perform image processing on the millstone image and calculate the kneading ratio N;

[0011] S7. Calculate the kneading efficiency X i ;

[0012] S8. Based on the calculated kneading efficiency X iAdjust the kneading efficiency D i and feed rate J i .

[0013] Preferably, step S1 specifically includes:

[0014] When the motor driving the grinding disc is working, its current will fluctuate within the rated range. When the load increases, that is, when the kneading damping is too large, the corresponding phase current will also increase accordingly. When the load decreases, the corresponding phase current will also decrease accordingly.

[0015] The current information is converted into voltage information by a sampling resistor and collected in real time.

[0016] The real-time damping of the millstone is represented by this voltage information, denoted as p. i ;

[0017] Real-time damping p i Stored in array format according to time series.

[0018] Preferably, step S3 specifically includes:

[0019] The feed inlet applies pressure directly to the tea leaves through a rotating pressure cover, forcing them into the grinding disc. This process is controlled by a stepper motor, which maintains the pressure F of the cover. i Controllable and adjustable, cover pressure F i Stored as an array.

[0020] Preferably, step S4 specifically includes:

[0021] The formula for calculating the feed rate is as follows:

[0022]

[0023] Among them, J i+1 F is the feed rate of the tea rolling machine at time i+1. i+1 This corresponds to the pressure of the cover plate at time i+1, p i It is the real-time damping of the millstone, and r is the rotational speed of the millstone motor.

[0024] Preferably, the feed rate J is collected over a period of time. i+1 Normalize it so that its value range is [0, 1], which facilitates subsequent processing.

[0025] Preferably, step S6 specifically includes:

[0026] 6.1 Ring Selection: The outer ring of the circular millstone image is selected as the detection area, and the tea distribution area on the millstone is detected. Each tea area is marked with a box of the same size, and the number of boxes is denoted as M1. Specifically, since the tea leaves and the millstone have obvious color differences, the distribution of tea leaves is detected by color recognition. The camera and the pre-built algorithm in OpenCV 3.4 are used to perform binarization, ellipse fitting, and center point localization on the dark green object to complete the box selection and detect the tea distribution area on the millstone. Each tea area is marked with a box of the same size, and the number of boxes is denoted as M1.

[0027] 6.2 For each tea-producing region, the following treatment shall be performed:

[0028] 6.2.1 Mark boxes to identify the shape of tea leaf distribution within the boxes;

[0029] 6.2.2 Based on the distribution shape of the tea leaves, find their maximum diameter R;

[0030] 6.2.2.1 Select the tea leaf distribution shape for easier processing;

[0031] 6.2.2.2 Contour extraction, identifying the largest contour;

[0032] 6.2.2.3 Traverse all lines that pass through the contour at the center point and record the two coordinates where the lines intersect the contour;

[0033] 6.2.2.4 Take the value with the largest distance between two coordinates as the maximum diameter R of the tea distribution shape;

[0034] 6.2.3 Using the maximum diameter R of the tea leaf distribution shape as the diagonal, reconstruct a new frame to encompass the entire tea leaf portion;

[0035] 6.2.4 Calculate the side length d of the new box based on the maximum diameter R, and use the value of the side length d as the number of pixels on the side length d;

[0036] 6.2.5 Count the number of pixels in the tea leaf distribution shape by traversing the image of the tea leaf distribution shape, denoted as S1;

[0037] 6.2.6 Judgment, if δ is a threshold, meaning that if the pixels of the tea leaf distribution shape are small enough, it indicates that the tea leaf curling is complete, and the count is M2.

[0038] 6.3 Traverse each tea leaf region in the millstone image and calculate the rolling ratio.

[0039] Preferably, step S7 specifically includes:

[0040] Kneading efficiency X i The calculation formula is as follows:

[0041]

[0042] Where N is the kneading ratio, J i It is the feed rate at the corresponding moment, F i It is the real-time cover pressure, p i It is the real-time damping of the millstone, X n 'r' represents the kneading efficiency of each tea-growing area, and 'r' represents the rotational speed of the grinding disc motor.

[0043] Preferably, for each kneading efficiency X n Normalize it so that its range is [0, 1].

[0044] Preferably, the rolling efficiency X is calculated for each tea leaf region in the image. n The average value is taken as the overall kneading efficiency D of the current image. i .

[0045] Preferably, step S8 specifically includes:

[0046] 8.1 When the kneading efficiency D i If the value is less than the set threshold c1, adjust the speed of the grinding disc motor to balance it.

[0047] 8.1.1 Where r is the original speed of the millstone motor, r0 is the change in the speed of the millstone motor, and D i It is the overall kneading efficiency, and c1 is the set threshold.

[0048] 8.1.2 The speed of the rolling compaction mechanism on the millstone is increased using an incremental method, r i =r i-1 +r0, where r i-1 It is the rotational speed of the previous millstone motor, r i It is the rotational speed of the millstone motor after it has been accelerated;

[0049] 8.2 When the overall kneading efficiency D i If the pressure F exceeds the set threshold c2, a stepper motor controlled by a PID controller is used to adjust the pressure of the cover plate. i This causes the pressure on the cover plate to increase by 0.5%, thereby adjusting the feed rate J. i .

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] 1. To address the lack of quality control in industrial tea rolling machines, a method for effectively evaluating the degree of tea leaf curling was designed by introducing visual image processing. This method is used to statistically analyze the rolling ratio of each tea area on the grinding disc. Through feedback from pressure sensors and the grinding disc motor, the capping pressure F is obtained. iand real-time damping p i Estimate the current feed rate J i Furthermore, the feeding can be controlled by adjusting the stepper motor's control system.

[0052] 2. Based on the rotational speed r of the grinding disc motor, the feed rate J is used. i Cap pressure F i Real-time damping p i By combining the rolling ratio N obtained from image processing, the rolling efficiency X of each tea leaf area can be calculated. i Then, by averaging the values, we can obtain the overall kneading efficiency D of the entire millstone. i .

[0053] 3. Control is achieved through threshold determination; when the overall kneading efficiency D... i When the overall kneading efficiency is low, use an incremental method to adjust the speed of the grinding disc motor to intensify kneading; when the overall kneading efficiency D... i At higher speeds, the overall production speed can be increased by using a stepped method to increase the pressure on the cover plate. Attached Figure Description

[0054] Figure 1 This is a flowchart illustrating the principle of the present invention. Detailed Implementation

[0055] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0056] like Figure 1 As shown, this invention provides a tea-rolling control method based on artificial intelligence, comprising the following steps:

[0057] Step 1: Collect the damping magnitude of the grinding disc of the kneading machine and the pressure of the pressure cover plate on the tea feed, and estimate the current feed rate of the kneading machine based on the grinding disc motor speed information.

[0058] 1. The damping of the motor is detected by testing the built-in feedback current of the grinding disc motor of the kneading machine.

[0059] 1.1 When the drive mill motor is working, its current will fluctuate within the rated range. When the load increases, that is, when the kneading damping is too large, the corresponding phase current will also increase accordingly. When the load decreases, the corresponding phase current will also decrease accordingly.

[0060] 1.2 Current information is converted into voltage information by sampling resistor and collected in real time.

[0061] 1.3 The real-time damping of the millstone is represented by this voltage information, denoted as p. i .

[0062] 1.4 Real-time damping p i Stored in array format according to time series.

[0063] 2. Collect the rotational speed r of the millstone motor.

[0064] 3. The downward pressure F of the pressure cover plate is directly measured using a pressure sensor. i .

[0065] 3.1 The feed inlet applies pressure directly to the tea leaves through a rotating pressure cover, thereby squeezing them into the grinding disc.

[0066] 3.2 Traditional kneading machines require manual cranking of the upper rotating rod to control the pressure of the cover plate, thereby controlling the feeding speed.

[0067] 3.3 In this embodiment, a stepper motor is used to control this process, so that the cover plate pressure F i It becomes controllable and adjustable.

[0068] 3.4 Cover plate pressure F i Stored as an array.

[0069] 4. Estimate the current feed rate J based on the collected information. i .

[0070] 4.1 During operation, tea rolling machines cannot accurately determine the feeding speed and often rely on manual experience. This is achieved by collecting data on the rotational speed (r) of the grinding disc motor and the pressure (F) of the cover plate. i and real-time damping p i This allows us to estimate the current feed rate J. i .

[0071] 4. J i+1 F is the feed rate of the tea rolling machine at time i+1. i+1 This corresponds to the pressure of the cover plate at time i+1, which indicates the degree of compression at the tea inlet. i It is the real-time damping of the millstone, and r is the rotational speed of the millstone motor.

[0072] 4.3 molecules (F) i+1 -F i )*F i+1 Based on the pressure at time i+1, and by comparing it with the pressure F at the previous time... iThe difference is used to represent the feed rate; the larger the difference, the more tea leaves are squeezed out of the outlet per unit time. The denominator uses the real-time damping of the grinding disc as a reference to exponentially amplify the influence of the grinding disc motor speed r.

[0073] 4.4 Further, the feed rate J is collected over a period of time. i+1 Normalize it so that its value range is [0, 1], which facilitates subsequent processing.

[0074] Step 2: Introduce image processing to complete the evaluation and calculation of kneading efficiency, realize intelligent control, and provide decision-making schemes in the control process.

[0075] 5. Collect surface images of the millstone.

[0076] Since the millstone itself doesn't move, the rolling mechanism rotates and rolls from above, creating friction. Therefore, a camera can be fixed in one position to capture images of the millstone's surface. Alternatively, multiple cameras can be used to capture images from different locations. Furthermore, the images are stitched together to create a complete image of the millstone.

[0077] 6. Perform image processing on the millstone image to calculate the kneading ratio N.

[0078] 6.1 Ring Selection: The outer ring of the circular millstone image is selected as the detection area, and the tea distribution area on the millstone is detected. Each tea area is marked with a box of the same size, and the number of boxes is denoted as M1. Specifically, since the tea leaves and the millstone have obvious color differences, the distribution of tea leaves can be detected by color recognition. Here, a camera is used with the pre-built algorithm in OpenCV 3.4 to perform binarization, ellipse fitting, and center point localization on the dark green object to complete the box selection and detect the tea distribution area on the millstone. Each tea area is marked with a box of the same size, and the number of boxes is denoted as M1.

[0079] 6.2 For each tea-producing region, the following treatment shall be performed:

[0080] 6.2.1 Mark boxes to identify the shape of tea leaf distribution within the boxes.

[0081] 6.2.2 Based on the distribution shape of the tea leaves, find its maximum diameter R.

[0082] 6.2.2.1 Specifically, select the shape of the tea leaf distribution for easier processing;

[0083] 6.2.2.2 Contour extraction, identifying the largest contour;

[0084] 6.2.2.3 Traverse all lines that pass through the contour at the center point and record the two coordinates where the lines intersect the contour;

[0085] 6.2.2.4 Take the maximum distance between the two coordinates as the maximum diameter R of the tea distribution shape.

[0086] 6.2.3 Further, using the maximum diameter R of the tea leaf distribution shape as the diagonal, a new frame is reconstructed to encompass the entire tea leaf portion. The orientation of this new frame often changes at this point.

[0087] 6.2.4 Calculate the side length d of the new box based on the maximum diameter R, and use the value of the side length d as the number of pixels on the side length d.

[0088] 6.2.5 Further, the number of pixels in the tea leaf distribution shape can be obtained by traversing the image of the tea leaf distribution shape, denoted as S1.

[0089] 6.2.6 Judgment, if δ is a threshold value, which is set to 0.85 in this embodiment. That is, if the pixels representing the shape of the tea leaves are small enough, it indicates that the tea leaves have finished curling, and the count is M2.

[0090] 6.3 Traverse each tea leaf area in the image and calculate the rolling ratio.

[0091] 7. Calculate the kneading efficiency X i It is used to evaluate the completion speed of the entire kneading process in real time.

[0092] 7.1 Because feeding and discharging occur simultaneously, it is necessary to balance the feeding and kneading speeds during this process. Therefore, a kneading efficiency is set to estimate the completion speed of the kneading process.

[0093] 7.2 N is the kneading ratio, J i It is the feed rate at the corresponding moment, F i It is the real-time cover pressure, p i It is the real-time damping of the millstone, X n 'r' represents the kneading efficiency of each tea-growing area, and 'r' represents the rotational speed of the grinding disc motor.

[0094] 7.3 This evaluation method is based on the rolling ratio N obtained from image acquisition. The larger the N, the higher the overall completion of tea rolling. Meanwhile, arctan(J) i+1 +J i +J i-1 ) is a normalization of multiple feed rates.

[0095] 7.4 per kneading efficiency X n Normalize it so that its range is [0, 1].

[0096] 7.5 Calculate the rolling efficiency X for each tea leaf region in the image. nThe average value is taken as the overall kneading efficiency D of the current image. i .

[0097] 8. The kneading efficiency X was achieved. i The evaluation requires balancing the overall kneading efficiency D. i and feed rate J i This helps avoid stockpiling materials or reducing production efficiency.

[0098] 8.1 When the kneading efficiency is less than the set threshold c1, adjust the speed of the grinding disc motor to balance it.

[0099] 8.1.1 Where r is the original speed of the millstone motor, r0 is the change in the speed of the millstone motor, and D i It is the overall kneading efficiency, and c1 is the set threshold.

[0100] 8.1.2 This completes the calculation of the change r0.

[0101] 8.1.3 Further, using an incremental method through r i =r i-1 +r0 increases the speed of the rolling compaction mechanism on the grinding disc, which facilitates thorough compaction and grinding. Here, r... i-1 It is the rotational speed of the previous millstone motor, r i It is the rotational speed of the millstone motor after it has been accelerated.

[0102] 8.2 When the overall kneading efficiency D i If the pressure F exceeds the set threshold c2, a stepper motor controlled by a PID controller is used to adjust the pressure of the cover plate. i This increases the pressure on the cover plate by 0.5%, thereby adjusting the feeding speed.

[0103] In summary, in view of the lack of quality control in industrial tea rolling machines, visual image processing was introduced to design a method that can effectively evaluate the degree of tea curling, and this method was used to statistically analyze the rolling ratio of each tea area on the grinding disc.

[0104] Furthermore, the pressure cap pressure F is obtained through the self-feedback of the pressure sensor and the grinding disc motor. i and real-time damping p i Estimate the current feed rate J i Furthermore, the feeding can be controlled by adjusting the stepper motor's control system.

[0105] Furthermore, based on the rotational speed r of the grinding disc motor, the feed rate J is used. i Cap pressure F i Real-time damping p i By combining the rolling ratio N obtained from image processing, the rolling efficiency X of each tea leaf area can be calculated.i Then, by averaging the values, we can obtain the overall kneading efficiency D of the entire millstone. i .

[0106] Ultimately, control is achieved through threshold determination; when the overall kneading efficiency D... i When the overall kneading efficiency is low, use an incremental method to adjust the speed of the grinding disc motor to intensify kneading; when the overall kneading efficiency D... i At higher speeds, the overall production speed can be increased by using a stepped method to increase the pressure on the cover plate.

[0107] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. A tea-rolling control method based on artificial intelligence, characterized in that, Includes the following steps: S1. Detect the damping of the motor by detecting the built-in feedback current of the grinding disc motor of the kneading machine. S2. Collect the rotational speed r of the millstone motor; S3. Directly measure the cover plate pressure Fdownwards of the pressure cover plate by means of a pressure sensor i ; S4. Calculate the current feed rate J from the information collected i ; S5. Acquire images of the millstone; S6. Perform image processing on the millstone image and calculate the kneading ratio N; S7, calculate rolling efficiency X i ; S8, the rolling efficiency X calculated i , adjusting the rolling efficiency D i and the feeding speed J i .

2. The tea leaf kneading control method based on artificial intelligence according to claim 1, characterized in that, Step S1 is as follows: When the motor driving the grinding disc is working, its current will fluctuate within the rated range. When the load increases, that is, when the kneading damping is too large, the corresponding phase current will also increase accordingly. When the load decreases, the corresponding phase current will also decrease accordingly. The current information is converted into voltage information by a sampling resistor and collected in real time. The real-time damping of the millstone is represented by this voltage information, denoted as p. i ; Real-time damping p i Stored in array format according to time series.

3. The tea leaf kneading control method based on artificial intelligence according to claim 2, characterized in that, Step S3 is as follows: The feed inlet applies pressure directly to the tea leaves through a rotating pressure cover, forcing them into the grinding disc. This process is controlled by a stepper motor, which maintains the pressure F of the cover. i Controllable and adjustable, cover pressure F i Stored as an array.

4. The tea leaf kneading control method based on artificial intelligence according to claim 3, characterized in that, Step S4 is as follows: The formula for calculating the feed rate is as follows: Among them, J i+1 F is the feed rate of the tea rolling machine at time i+1. i+1 This corresponds to the pressure of the cover plate at time i+1, p i It is the real-time damping of the millstone, and r is the rotational speed of the millstone motor.

5. The tea leaf kneading control method based on artificial intelligence according to claim 4, characterized in that, The feed rate J was collected over a period of time. i+1 Normalize it so that its value range is [0, 1], which facilitates subsequent processing.

6. The tea leaf kneading control method based on artificial intelligence according to claim 5, characterized in that step S6 specifically comprises: 6.1 Ring Selection: The outer ring of the circular millstone image is selected as the detection area, and the tea distribution area on the millstone is detected. Each tea area is marked with a box of the same size, and the number of boxes is denoted as M1. Specifically, since the tea leaves and the millstone have obvious color differences, the distribution of tea leaves is detected by color recognition. The camera and the pre-built algorithm in OpenCV 3.4 are used to perform binarization, ellipse fitting, and center point localization on the dark green object to complete the box selection and detect the tea distribution area on the millstone. Each tea area is marked with a box of the same size, and the number of boxes is denoted as M1. 6.2 For each tea-producing region, the following treatment shall be performed: 6.2.1 Mark boxes to identify the shape of tea leaf distribution within the boxes; 6.2.2 Based on the distribution shape of the tea leaves, find their maximum diameter R; 6.2.2.1 Select the tea leaf distribution shape for easier processing; 6.2.2.2 Contour extraction, identifying the largest contour; 6.2.2.3 Traverse all lines that pass through the contour at the center point and record the two coordinates where the lines intersect the contour; 6.2.2.4 Take the value with the largest distance between two coordinates as the maximum diameter R of the tea distribution shape; 6.2.3 Using the maximum diameter R of the tea leaf distribution shape as the diagonal, reconstruct a new frame to encompass the entire tea leaf portion; 6.2.4 Calculate the side length d of the new box based on the maximum diameter R, and use the value of the side length d as the number of pixels on the side length d; 6.2.5 Count the number of pixels in the tea leaf distribution shape by traversing the image of the tea leaf distribution shape, denoted as S1; 6.2.6 Judgment, if δ is a threshold, meaning that if the pixels of the tea leaf distribution shape are small enough, it indicates that the tea leaf curling is complete, and the count is M2. 6.3 Traverse each tea leaf region in the millstone image and calculate the rolling ratio.

7. The tea leaf kneading control method based on artificial intelligence according to claim 6, characterized in that, Step S7 is as follows: Kneading efficiency X i The calculation formula is as follows: Where N is the kneading ratio, J i It is the feed rate at the corresponding moment, F i It is the real-time cover pressure, p i It is the real-time damping of the millstone, X n 'r' represents the kneading efficiency of each tea-growing area, and 'r' represents the rotational speed of the grinding disc motor.

8. The tea leaf kneading control method based on artificial intelligence according to claim 7, characterized in that, For each kneading efficiency X n Normalize it so that its range is [0, 1].

9. The tea leaf kneading control method based on artificial intelligence according to claim 8, characterized in that, Calculate the rolling efficiency X for each tea leaf region in the image. n The average value is taken as the overall kneading efficiency D of the current image. i .

10. The tea leaf kneading control method based on artificial intelligence according to claim 9, characterized in that, Step S8 is as follows: 8.1 When the kneading efficiency D i If the value is less than the set threshold c1, adjust the speed of the grinding disc motor to balance it. 8.1.1 Where r is the original speed of the millstone motor, r0 is the change in the speed of the millstone motor, and D i It is the overall kneading efficiency, and c1 is the set threshold. 8.1.2 The speed of the rolling compaction mechanism on the millstone is increased using an incremental method, r i =r i-1 +r0, where, r i-1 It is the rotational speed of the previous millstone motor, r i It is the rotational speed of the millstone motor after it has been accelerated; 8.2 When the overall kneading efficiency D i If the pressure F exceeds the set threshold c2, a stepper motor controlled by a PID controller is used to adjust the pressure of the cover plate. i This causes the pressure on the cover plate to increase by 0.5%, thereby adjusting the feed rate J. i .

Images