Automatic control system and method for tobacco leaf baking based on machine vision

Abstract

The present invention discloses a tobacco leaf baking automatic control system and method based on machine vision. A mapping relationship table is constructed according to historical baking data and expert experience. The corresponding mapping relationship table is selected according to the number and type of tobacco leaves to be baked, the meteorological parameters at that time and the moisture content of the tobacco leaves to be baked. During the baking process, according to the determined mapping relationship table, the real-time color value of the tobacco leaves, the calculated real-time value of the color change rate and the real-time value of the temperature change rate in the baking room are obtained according to the collected real-time images, and the corresponding heating tube current, circulating fan current and dehumidification fan current are automatically output; at the same time, the digital twin model also predicts and outputs the color of the baked tobacco leaves at the next moment for display; a comprehensive consideration is given to various factors affecting the baking quality of tobacco leaves, thereby ensuring the safety of the tobacco leaf baking process and the improvement of the tobacco leaf baking quality.

Description

Technical Field

[0001] The present invention relates to a tobacco leaf baking automatic control system and method based on machine vision, and in particular to an automatic baking control method for an airflow rising type baking room. Background Art

[0002] Traditional tobacco curing rooms burn fuel in a furnace to generate hot air, which is then delivered into the curing room via a combustion-supporting fan. The combustion-supporting fan is adjusted and controlled based on the detected chamber temperature, thereby achieving automatic curing control. However, the temperature-based detection method has a lag. For example, when the temperature reaches the required value, even if the fuel addition amount and the combustion-supporting fan are immediately controlled, the combustion of the fuel and the delivery of the hot air require a certain amount of time, resulting in the delivery of excess heat, which in turn affects the curing quality of the tobacco leaves.

[0003] To improve baking quality, traditional baking methods incorporate leaf color as a factor to determine the baking progress. However, leaf color detection relies primarily on manual labor, a subjective judgment method that varies from person to person. Different tobacco workers, under the same conditions, may produce different judgments, resulting in suboptimal baking quality. Although machine vision has gained popularity in recent years and has found initial application in tobacco baking control systems, primarily to determine baking progress based on leaf color and subsequently adjust baking temperature, interference factors can occur during the baking process. For example, abnormal temperature can cause abnormal leaf color changes, leading to visual inspection results that mistakenly indicate the next stage of the baking process, resulting in baking failure.

[0004] In addition, considering only the baking temperature and humidity and the color of the baked tobacco leaves as reference control factors is too simplistic and will also affect the quality of the finished baked tobacco leaves.

[0005] Therefore, in order to solve the above problems in the tobacco baking process, it is necessary to provide a new automatic control method for tobacco baking. Summary of the Invention

[0006] The purpose of the present invention is to overcome the shortcomings of the existing technology and propose an automatic control system and method for tobacco leaf baking based on machine vision, which comprehensively considers various factors affecting the quality of tobacco leaf baking, ensures the safety of the tobacco leaf baking process and improves the tobacco leaf baking quality.

[0007] The present invention provides a tobacco leaf baking automatic control method based on machine vision, which adopts the following steps:

[0008] S1: Based on historical tobacco curing data and expert experience, a mapping table is constructed between meteorological parameters, tobacco leaf quantity, type, moisture content, tobacco leaf color value, color change rate, temperature change rate in the curing room, and heating tube current, circulation fan current, and dehumidification fan current.

[0009] S2: Obtain the quantity and type of tobacco leaves to be cured, the meteorological parameters at the time, and the moisture content of the tobacco leaves to be cured according to the collected images, and determine a corresponding mapping relationship table according to the quantity and type of tobacco leaves to be cured, the meteorological parameters at the time, and the moisture content of the tobacco leaves to be cured;

[0010] S3: During the baking process, according to the determined mapping relationship table, the real-time color value of the tobacco leaves, the calculated real-time color change rate value, and the real-time temperature change rate value in the baking room are obtained based on the real-time image collected, and the corresponding heating tube current, circulation fan current, and dehumidification fan current are automatically output. At the same time, the digital twin model predicts and displays the color of the baked tobacco leaves at the next moment based on the obtained real-time color value, color change rate real-time value, and temperature change rate real-time value.

[0011] S4: After the baking is completed, the mapping relationship table is modified and stored according to the baking result.

[0012] Preferably, the step S1 specifically includes:

[0013] S11: Based on historical baking data and expert experience, empirical control current parameters for heating tube current, circulation fan current, and dehumidification fan current under different meteorological parameters, quantity, type, and moisture content of tobacco leaves to be cured are extracted;

[0014] S12: Calculating the color value, color change rate, and temperature change rate of the flue-cured tobacco leaves in each flue-curing process under the above-mentioned different meteorological parameters, quantity, type, and moisture content of the flue-cured tobacco leaves;

[0015] S13: In each baking process, when the color change rate of the baked tobacco leaves is greater than or equal to a first preset threshold, a mapping output relationship between the color change rate of the baked tobacco leaves and the current of the heating tube, the circulation fan current, and the dehumidification fan current is correspondingly established;

[0016] When the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the temperature change rate in the flue-curing room is greater than or equal to a second threshold value, a mapping output relationship between the temperature change rate in the flue-curing room and the current of the heating tube, the current of the circulating fan, and the current of the dehumidification fan is correspondingly constructed;

[0017] When the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the temperature change rate in the flue-curing room is less than a second threshold value, a mapping output relationship between the tobacco leaf color value and the heating tube current, the circulation fan current, and the dehumidification fan current is correspondingly constructed;

[0018] S14: Integrate the mapping output relationships in each baking process in step S13 to obtain a mapping relationship table for the baking process.

[0019] The calculation of the color change rate of flue-cured tobacco leaves specifically includes:

[0020] S121: Preprocessing the collected real-time image every N frames to extract the RGB three-channel feature values ​​of the tobacco leaf; N is an integer greater than or equal to 1;

[0021] S122: Calculate the change rates of the R channel value, the G channel value, and the B channel value relative to the time interval of collecting N frames of images based on the extracted RGB three-channel feature values;

[0022] S123: Calculate the average of the change rates of the RGB three-channel values ​​as the color change rate value of the flue-cured tobacco leaf.

[0023] The sum of the extracted RGB three-channel feature values ​​is used as the color value of the flue-cured tobacco leaf.

[0024] Preferably, calculating the color change rate value of the flue-cured tobacco leaves specifically includes:

[0025] S1211: Preprocessing the collected real-time image to extract RGB three-channel feature values ​​of the tobacco leaf;

[0026] S1221: Calculate the mean of the RGB three-channel feature values ​​within each T time interval, and calculate the change rate of the R channel mean, the G channel mean, and the B channel mean relative to the two previous time intervals;

[0027] S1231: Calculate the average of the mean change rates of the RGB three channels as the color change rate value of the flue-cured tobacco leaves.

[0028] The mean value of the extracted RGB three-channel feature values ​​is used as the color value of the flue-cured tobacco leaves.

[0029] Preferably, calculating the temperature change rate value in the baking room specifically includes: calculating the temperature change rate in the baking room specifically includes: capturing N frames of images at a time interval of t, obtaining the temperature values ​​at time t-1 and time t respectively, and taking the ratio of the difference between the two temperature values ​​to the acquisition time interval as the temperature change rate value in the baking room.

[0030] Preferably, in step S2, obtaining the moisture content of the tobacco leaves to be cured according to the collected images specifically includes:

[0031] S21: preprocessing the acquired image to extract the color, texture and shape feature values ​​of the tobacco leaves in the image;

[0032] S22: Inputting the color, texture and shape feature values ​​of the tobacco leaves in the extracted image into a neural network prediction model to calculate and output the moisture content of the tobacco leaves to be cured.

[0033] Preferably, in step S3, according to the determined mapping relationship table, the real-time value of the tobacco leaf color, the calculated real-time value of the color change rate, and the real-time value of the temperature change rate in the curing room are obtained according to the real-time image, and the corresponding heating tube current, circulation fan current, and dehumidification fan current are automatically output, specifically including:

[0034] S31: extracting RGB three-channel feature values ​​based on the collected real-time image, and calculating the real-time value of the color of the flue-cured tobacco leaves, the real-time value of the color change rate, and the real-time value of the temperature change rate in the flue-curing room;

[0035] S32: When the real-time value of the flue-cured tobacco leaf color change rate is greater than or equal to a first preset threshold, searching a mapping table for the heating tube current, the circulating fan current, and the dehumidification fan current corresponding to the real-time value of the flue-cured tobacco leaf color change rate;

[0036] When the real-time value of the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the real-time value of the temperature change rate in the flue-curing room is greater than or equal to a second threshold value, the heating tube current, the circulation fan current, and the dehumidification fan current corresponding to the real-time value of the temperature change rate in the flue-curing room are searched in the mapping relationship table;

[0037] When the real-time value of the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the real-time value of the temperature change rate in the flue-curing room is less than a second threshold value, the heating tube current, the circulating fan current, and the dehumidification fan current corresponding to the real-time value of the color of the flue-cured tobacco leaves at this time are searched in the mapping relationship table;

[0038] S33: Outputting corresponding control currents according to the heating pipe current, circulation fan current, and dehumidification fan current obtained by searching to drive the heating pipe, circulation fan, and dehumidification fan to operate respectively.

[0039] Preferably, in step S3, the digital twin model predicts and outputs the color of the flue-cured tobacco leaves at the next moment based on the acquired real-time color value, real-time color change rate value, and real-time temperature change rate value, and displays the predicted color, specifically including:

[0040] An initial digital twin model is built based on the structure of the baking room and the geographical location of the baking room. Then, meteorological parameters, the number, type, and moisture content of tobacco leaves to be cured are input into the initial model for combination to obtain a predictive digital twin model. During the baking process, the real-time color value of the tobacco leaves, the calculated real-time value of the color change rate, and the real-time value of the temperature change rate in the baking room are obtained based on the real-time image collected and input into the predictive digital twin model. The predictive digital twin model predicts and outputs the color of the tobacco leaves to be baked at the next moment according to the rules set by the model and displays it. The next moment can be a moment with a preset time interval between the current moment, and the preset time interval can be adjusted during the baking process. The color predicted and output by the predictive digital twin model can be displayed on the display terminal of the control system, or it can be sent to the remote terminal of the baking worker for display.

[0041] Preferably, in step S4, after baking is completed, the mapping relationship table is modified and stored according to the baking result, specifically including:

[0042] After the baking is completed, the quality of the cured tobacco leaves is tested and evaluated. If the quality of the cured tobacco leaves does not meet the preset standards, the baking process is reviewed to find the reasons for the decline in baking quality, and the corresponding factor values ​​in the mapping relationship table are modified. The modified mapping relationship table is then stored in the database; if the quality of the cured tobacco leaves meets the standards, no correction is made.

[0043] Preferably, in step S3, during the baking process, the temperature and humidity values ​​in the baking room are detected in real time. When the detected temperature and humidity values ​​exceed a preset threshold, an alarm message is output and sent to a remote terminal.

[0044] Preferably, in step S1, the meteorological parameters are real-time temperature and humidity values ​​outside the baking room.

[0045] The present invention also includes a tobacco leaf baking automatic control system based on machine vision, including a controller, a cloud database, an image acquisition module, a data acquisition module and an actuator;

[0046] A mapping relationship table is constructed based on historical baking data and expert experience, which includes meteorological parameters, the number, type, moisture content, color value, color change rate, temperature change rate in the baking room, and heating tube current, circulation fan current, and dehumidification fan current. The mapping relationship table is then modified and stored in the cloud database based on the baking results.

[0047] The controller obtains the input quantity and type of tobacco leaves to be cured, and the current meteorological parameters sent by the data acquisition module, and processes the images collected by the image acquisition module to obtain the moisture content of the tobacco leaves to be cured. The controller then searches the cloud database to determine the corresponding mapping relationship table based on the quantity and type of tobacco leaves to be cured, the current meteorological parameters, and the moisture content of the tobacco leaves to be cured.

[0048] During the baking process, the controller also calculates the real-time value of the tobacco leaf color, the real-time value of the color change rate, and the real-time value of the temperature change rate in the baking room according to the real-time image collected according to the determined mapping relationship table, and automatically outputs the corresponding control signal to the actuator;

[0049] The actuator includes a heating tube, a circulation fan, and a dehumidification fan; the corresponding control signals are the heating tube current, the circulation fan current, and the dehumidification fan current.

[0050] Preferably, the image acquisition module is a defogging camera.

[0051] Preferably, the data acquisition module includes a temperature and humidity sensor inside the baking room and a temperature and humidity sensor outside the baking room.

[0052] The beneficial effects of the present invention are as follows: 1. A heating tube heating method is adopted instead of a traditional fuel heating method. The adjustment of the heating tube is based on the change of the heating tube current, and the circulating fan is coordinated with the control operation, thereby realizing rapid feedback adjustment of the temperature, improving the control lag caused by the traditional fuel furnace heating, and significantly improving the tobacco leaf curing quality;

[0053] 2. Based on the traditional method of judging the baking process based on temperature or tobacco leaf color, a baking process mapping relationship table is proposed based on the calculated real-time value of tobacco leaf color, real-time value of color change rate, and real-time value of temperature change rate in the baking room according to the characteristics of tobacco leaves. Different detection and control parameters are used in different baking stages, making the automatic control of the baking process more precise, eliminating abnormal interference in the baking process, and improving the safety and accuracy of the baking process;

[0054] 3. In addition to considering the temperature and humidity factors of the tobacco leaves to be cured, the quantity and type of tobacco leaves during curing, weather factors and moisture content are also taken into account to improve the curing quality of the tobacco leaves;

[0055] 4. A set of image acquisition modules, namely anti-fog cameras, are used to detect the moisture content and color of tobacco leaves at the same time, saving system construction costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] Figure 1 This is a step diagram of the smoke baking automatic control method based on machine vision;

[0057] Figure 2 This is a step-by-step diagram of the method for calculating the moisture content of tobacco leaves;

[0058] Figure 3 A step-by-step diagram of the method for determining the control parameters of the baking process. DETAILED DESCRIPTION

[0059] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings, but the protection scope of the present invention is not limited to the following.

[0060] The present invention provides a tobacco leaf baking automatic control method based on machine vision, such as Figure 1 As shown, the following steps are taken:

[0061] S1: Based on historical tobacco curing data and expert experience, a mapping table is constructed between meteorological parameters, tobacco leaf quantity, type, moisture content, tobacco leaf color value, color change rate, temperature change rate in the curing room, and heating tube current, circulation fan current, and dehumidification fan current.

[0062] S2: Obtain the quantity and type of tobacco leaves to be cured, the meteorological parameters at the time, and the moisture content of the tobacco leaves to be cured according to the collected images, and determine a corresponding mapping relationship table according to the quantity and type of tobacco leaves to be cured, the meteorological parameters at the time, and the moisture content of the tobacco leaves to be cured;

[0063] S3: During the baking process, according to the determined mapping relationship table, the real-time color value of the tobacco leaves, the calculated real-time color change rate value, and the real-time temperature change rate value in the baking room are obtained based on the real-time image collected, and the corresponding heating tube current, circulation fan current, and dehumidification fan current are automatically output. At the same time, the digital twin model predicts and displays the color of the baked tobacco leaves at the next moment based on the obtained real-time color value, color change rate real-time value, and temperature change rate real-time value.

[0064] S4: After the baking is completed, the mapping relationship table is modified and stored according to the baking result.

[0065] Preferably, the step S1 specifically includes:

[0066] S11: Based on historical baking data and expert experience, empirical control current parameters for heating tube current, circulation fan current, and dehumidification fan current under different meteorological parameters, quantity, type, and moisture content of tobacco leaves to be cured are extracted;

[0067] S12: Calculating the color value, color change rate, and temperature change rate of the flue-cured tobacco leaves in each flue-curing process under the above-mentioned different meteorological parameters, quantity, type, and moisture content of the flue-cured tobacco leaves;

[0068] S13: In each baking process, when the color change rate of the baked tobacco leaves is greater than or equal to a first preset threshold, a mapping output relationship between the color change rate of the baked tobacco leaves and the current of the heating tube, the circulation fan current, and the dehumidification fan current is correspondingly established;

[0069] When the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the temperature change rate in the flue-curing room is greater than or equal to a second threshold value, a mapping output relationship between the temperature change rate in the flue-curing room and the current of the heating tube, the current of the circulating fan, and the current of the dehumidification fan is correspondingly constructed;

[0070] When the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the temperature change rate in the flue-curing room is less than a second threshold value, a mapping output relationship between the tobacco leaf color value and the heating tube current, the circulation fan current, and the dehumidification fan current is correspondingly constructed;

[0071] S14: Integrate the mapping output relationships in each baking process in step S13 to obtain a mapping relationship table for the baking process.

[0072] The calculation of the color change rate of flue-cured tobacco leaves specifically includes:

[0073] S121: Preprocess the collected real-time image every N frames to extract the RGB three-channel feature value R of the tobacco leaf i , G i 、B i ; N is an integer greater than or equal to 1, i is an interval value;

[0074] S122: Based on the extracted RGB three-channel feature values, respectively calculate the change rate of the R channel value, the G channel value, and the B channel value relative to the interval time of collecting N frames of images; specifically, where R i is the R channel eigenvalue extracted at the i-th interval, R i-1 is the R channel feature value extracted at the i-1th interval, i t is the time value between the i-th interval and the i-1-th interval; similarly, G 变 and B 变 value;;

[0075] S123: Calculate the average of the change rates of the RGB three-channel values ​​as the color change rate value of the flue-cured tobacco leaf.

[0076] The sum of the extracted RGB three-channel feature values ​​is used as the color value of the flue-cured tobacco leaf.

[0077] Preferably, calculating the color change rate value of the flue-cured tobacco leaves specifically includes:

[0078] S1211: Preprocessing the collected real-time image to extract RGB three-channel feature values ​​of the tobacco leaf;

[0079] S1221: Calculate the mean of the RGB three-channel feature values ​​in each T time interval, and calculate the change rate of the R channel mean, G channel mean, and B channel mean relative to the two previous time intervals; specifically: Where j is the number of images acquired in the time interval T;

[0080] where R 均x is the mean R value calculated in the xth time interval, R 均x-1 is the R-mean calculated during the previous time interval of the xth time;

[0081] Similarly, get G 均 、B 均 Value and G 均变 、B 均变 value;

[0082] S1231: Calculate the average of the mean change rates of the RGB three channels as the color change rate value of the flue-cured tobacco leaves.

[0083] The mean value of the extracted RGB three-channel feature values ​​is used as the color value of the flue-cured tobacco leaves.

[0084] Preferably, calculating the temperature change rate value in the baking room specifically includes: calculating the temperature change rate in the baking room specifically includes: capturing N frames of images at a time interval of t, obtaining the temperature values ​​at time t-1 and time t respectively, and taking the ratio of the difference between the two temperature values ​​to the acquisition time interval as the temperature change rate value in the baking room.

[0085] Preferably, in step S2, the moisture content of the tobacco leaves to be cured is obtained based on the collected image, such as Figure 2 As shown, specifically including:

[0086] S21: preprocessing the acquired image to extract the color, texture and shape feature values ​​of the tobacco leaves in the image;

[0087] S22: Inputting the color, texture and shape feature values ​​of the tobacco leaves in the extracted image into a neural network prediction model to calculate and output the moisture content of the tobacco leaves to be cured.

[0088] Preferably, in step S3, according to the determined mapping relationship table, the real-time value of the color of the tobacco leaf, the calculated real-time value of the color change rate and the real-time value of the temperature change rate in the baking room are obtained according to the real-time image, and the corresponding heating tube current, circulation fan current and dehumidification fan current are automatically output, such as Figure 3 As shown, specifically including:

[0089] S31: extracting RGB three-channel feature values ​​based on the collected real-time image, and calculating the real-time value of the color of the flue-cured tobacco leaves, the real-time value of the color change rate, and the real-time value of the temperature change rate in the flue-curing room;

[0090] S32: When the real-time value of the flue-cured tobacco leaf color change rate is greater than or equal to a first preset threshold, searching a mapping table for the heating tube current, the circulating fan current, and the dehumidification fan current corresponding to the real-time value of the flue-cured tobacco leaf color change rate;

[0091] When the real-time value of the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the real-time value of the temperature change rate in the flue-curing room is greater than or equal to a second threshold value, the heating tube current, the circulation fan current, and the dehumidification fan current corresponding to the real-time value of the temperature change rate in the flue-curing room are searched in the mapping relationship table;

[0092] When the real-time value of the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the real-time value of the temperature change rate in the flue-curing room is less than a second threshold value, the heating tube current, the circulating fan current, and the dehumidification fan current corresponding to the real-time value of the color of the flue-cured tobacco leaves at this time are searched in the mapping relationship table;

[0093] S33: Outputting corresponding control currents according to the heating pipe current, circulation fan current, and dehumidification fan current obtained by searching to drive the heating pipe, circulation fan, and dehumidification fan to operate respectively.

[0094] Specifically, for the calculated real-time value of tobacco leaf color, real-time value of color change rate and real-time value of temperature change rate in the baking room, when detecting and judging the heating tube current, circulation fan current and dehumidification fan current mapped according to the real-time value of color change rate of the baked tobacco leaves, it is also judged whether the real-time value of tobacco leaf color and the real-time value of temperature change rate in the baking room are within the corresponding preset threshold range of this process in the mapping relationship table. If they are within the preset threshold range, the heating tube current, circulation fan current and dehumidification fan current corresponding to the process are output; if any value of the two does not match the preset threshold range of the corresponding process in the mapping relationship table, an alarm message is output and sent to the remote terminal. Remind workers to intervene in time to avoid baking failure and waste of raw materials; similarly, when detecting and judging the heating tube current, circulation fan current, and dehumidification fan current mapped according to the real-time value of the tobacco leaf color and the real-time value of the temperature change rate in the baking room, it also judges whether the other two parameters are within the corresponding preset threshold range of this process in the mapping relationship table. If they are within the preset threshold range, the heating tube current, circulation fan current, and dehumidification fan current corresponding to the process are output; if any of the two values ​​does not match the preset threshold range of the corresponding process in the mapping relationship table, an alarm message is output and sent to the remote terminal to remind workers to intervene in time to avoid baking failure and waste of raw materials.

[0095] Preferably, in step S3, the digital twin model predicts and outputs the color of the flue-cured tobacco leaves at the next moment based on the acquired real-time color value, real-time color change rate value, and real-time temperature change rate value, and displays the predicted color, specifically including:

[0096] An initial digital twin model is built based on the structure and geographic location of the flue-curing room. Meteorological parameters, the number, type, and moisture content of tobacco leaves to be cured are then input into the initial model and combined to obtain a predictive digital twin model. During the curing process, the real-time color value of the tobacco leaves, the calculated real-time color change rate, and the real-time temperature change rate in the flue-curing room are obtained based on the real-time image captured and input into the predictive digital twin model. The predictive digital twin model predicts and displays the color of the flue-cured tobacco leaves at the next moment based on the rules set by the model. The next moment can be a moment with a preset time interval from the current moment. This preset time interval can be adjusted during the curing process. For example, if a flue-curing worker wants to observe the color of the flue-cured tobacco leaves 0.5 hours and 1 hour after the current curing data moment, the preset time interval can be set to 0.5 hours. After observing the predicted color of the flue-cured tobacco leaves, the preset time interval can be set to 1 hour to observe the predicted color of the flue-cured tobacco leaves. The color predicted by the predictive digital twin model can be displayed on the display terminal of the control system or sent to the flue-curing worker's remote terminal for display.

[0097] Preferably, in step S4, after baking is completed, the mapping relationship table is modified and stored according to the baking result, specifically including:

[0098] After the baking is completed, the quality of the cured tobacco leaves is tested and evaluated. If the quality of the cured tobacco leaves does not meet the preset standards, the baking process is reviewed to find the reasons for the decline in baking quality, and the corresponding factor values ​​in the mapping relationship table are modified. The modified mapping relationship table is then stored in the database; if the quality of the cured tobacco leaves meets the standards, no correction is made.

[0099] Preferably, in step S3, during the baking process, the temperature and humidity values ​​in the baking room are detected in real time. When the detected temperature and humidity values ​​exceed a preset threshold, an alarm message is output and sent to a remote terminal.

[0100] Preferably, in step S1, the meteorological parameters are real-time temperature and humidity values ​​outside the baking room.

[0101] The present invention also includes a tobacco leaf baking automatic control system based on machine vision, including a controller, a cloud database, an image acquisition module, a data acquisition module and an actuator;

[0102] A mapping relationship table is constructed based on historical baking data and expert experience, which includes meteorological parameters, the number, type, moisture content, color value, color change rate, temperature change rate in the baking room, and heating tube current, circulation fan current, and dehumidification fan current. The mapping relationship table is then modified and stored in the cloud database based on the baking results.

[0103] The controller obtains the input quantity and type of tobacco leaves to be cured, and the current meteorological parameters sent by the data acquisition module, and processes the images collected by the image acquisition module to obtain the moisture content of the tobacco leaves to be cured. The controller then searches the cloud database to determine the corresponding mapping relationship table based on the quantity and type of tobacco leaves to be cured, the current meteorological parameters, and the moisture content of the tobacco leaves to be cured.

[0104] During the baking process, the controller also calculates the real-time value of the tobacco leaf color, the real-time value of the color change rate, and the real-time value of the temperature change rate in the baking room according to the real-time image collected according to the determined mapping relationship table, and automatically outputs the corresponding control signal to the actuator;

[0105] The actuator includes a heating tube, a circulation fan, and a dehumidification fan; the corresponding control signals are the heating tube current, the circulation fan current, and the dehumidification fan current.

[0106] Preferably, the image acquisition module is a defogging camera.

[0107] Preferably, the data acquisition module includes a temperature and humidity sensor inside the baking room and a temperature and humidity sensor outside the baking room.

[0108] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a ROM, a RAM, or the like.

[0109] The above disclosure is merely a preferred embodiment of the present invention and certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A tobacco leaf baking automatic control method based on machine vision, characterized in that: Use the following steps: S1: Based on historical tobacco curing data and expert experience, a mapping table is constructed between meteorological parameters, tobacco leaf quantity, type, moisture content, tobacco leaf color value, color change rate, temperature change rate in the curing room, and heating tube current, circulation fan current, and dehumidification fan current. S2: Obtain the quantity and type of tobacco leaves to be cured, the meteorological parameters at the time, and the moisture content of the tobacco leaves to be cured according to the collected images, and determine a corresponding mapping relationship table according to the quantity and type of tobacco leaves to be cured, the meteorological parameters at the time, and the moisture content of the tobacco leaves to be cured; S3: During the baking process, according to the determined mapping relationship table, the real-time color value of the tobacco leaves, the calculated real-time color change rate value, and the real-time temperature change rate value in the baking room are obtained based on the real-time image collected, and the corresponding heating tube current, circulation fan current, and dehumidification fan current are automatically output. At the same time, the digital twin model predicts and displays the color of the baked tobacco leaves at the next moment based on the obtained real-time color value, color change rate real-time value, and temperature change rate real-time value. S4: After the baking is completed, the mapping relationship table is modified and stored according to the baking result; The step S1 specifically includes: S11: Based on historical baking data and expert experience, empirical control current parameters for heating tube current, circulation fan current, and dehumidification fan current under different meteorological parameters, quantity, type, and moisture content of tobacco leaves to be cured are extracted; S12: Calculating the color value, color change rate, and temperature change rate of the flue-cured tobacco leaves in each flue-curing process under the above-mentioned different meteorological parameters, quantity, type, and moisture content of the flue-cured tobacco leaves; S13: In each baking process, when the color change rate of the baked tobacco leaves is greater than or equal to a first preset threshold, a mapping output relationship between the color change rate of the baked tobacco leaves and the current of the heating tube, the circulation fan current, and the dehumidification fan current is correspondingly established; When the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the temperature change rate in the flue-curing room is greater than or equal to a second threshold value, a mapping output relationship between the temperature change rate in the flue-curing room and the current of the heating tube, the current of the circulating fan, and the current of the dehumidification fan is correspondingly constructed; When the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the temperature change rate in the flue-curing room is less than a second threshold value, a mapping output relationship between the tobacco leaf color value and the heating tube current, the circulation fan current, and the dehumidification fan current is correspondingly constructed; S14: Integrate the mapping output relationships of each baking process in step S13 to obtain a mapping relationship table for the baking process; In step S3, according to the determined mapping relationship table, the real-time color value of the tobacco leaf, the calculated real-time value of the color change rate, and the real-time value of the temperature change rate in the curing room are obtained based on the real-time image, and the corresponding heating tube current, circulation fan current, and dehumidification fan current are automatically output, specifically including: S31: extracting RGB three-channel feature values ​​based on the collected real-time image, and calculating the real-time value of the color of the flue-cured tobacco leaves, the real-time value of the color change rate, and the real-time value of the temperature change rate in the flue-curing room; S32: When the real-time value of the flue-cured tobacco leaf color change rate is greater than or equal to a first preset threshold, searching a mapping table for the heating tube current, the circulating fan current, and the dehumidification fan current corresponding to the real-time value of the flue-cured tobacco leaf color change rate; When the real-time value of the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the real-time value of the temperature change rate in the flue-curing room is greater than or equal to a second threshold value, the heating tube current, the circulation fan current, and the dehumidification fan current corresponding to the real-time value of the temperature change rate in the flue-curing room are searched in the mapping relationship table; When the real-time value of the color change rate of the flue-cured tobacco leaves is less than a first preset threshold value, and the real-time value of the temperature change rate in the flue-curing room is less than a second threshold value, the heating tube current, the circulating fan current, and the dehumidification fan current corresponding to the real-time value of the color of the flue-cured tobacco leaves at this time are searched in the mapping relationship table; S33: Outputting corresponding control currents according to the heating pipe current, circulation fan current, and dehumidification fan current obtained by searching to drive the heating pipe, circulation fan, and dehumidification fan to operate respectively.

2. A tobacco leaf baking automatic control method based on machine vision according to claim 1, characterized in that: Calculation of the color change rate of flue-cured tobacco leaves specifically includes: S121: Preprocessing the collected real-time image every N frames to extract the RGB three-channel feature values ​​of the tobacco leaf; N is an integer greater than or equal to 1; S122: Calculate the change rates of the R channel value, the G channel value, and the B channel value relative to the time interval of collecting N frames of images based on the extracted RGB three-channel feature values; S123: Calculate the average of the change rates of the RGB three-channel values ​​as the color change rate value of the flue-cured tobacco leaf.

3. A tobacco leaf baking automatic control method based on machine vision according to claim 1, characterized in that: In step S2, the moisture content of the tobacco leaves to be cured is obtained according to the collected images, which specifically includes: S21: preprocessing the acquired image to extract the color, texture and shape feature values ​​of the tobacco leaves in the image; S22: Inputting the color, texture and shape feature values ​​of the tobacco leaves in the extracted image into a neural network prediction model to calculate and output the moisture content of the tobacco leaves to be cured.

4. A tobacco leaf baking automatic control method based on machine vision according to claim 1, characterized in that: In step S3, the digital twin model predicts and outputs the color of the flue-cured tobacco leaves at the next moment based on the acquired real-time color value, real-time color change rate value, and real-time temperature change rate value, and displays the predicted color. Specifically, the digital twin model includes: An initial digital twin model is built based on the structure of the baking room and the geographical location of the baking room. Then, meteorological parameters, the number, type, and moisture content of tobacco leaves to be baked are input into the initial model and combined to obtain a predictive digital twin model. During the baking process, the real-time color value of the tobacco leaves, the calculated real-time value of the color change rate, and the real-time value of the temperature change rate in the baking room are obtained based on the collected real-time images and input into the predictive digital twin model. The predictive digital twin model predicts and displays the color of the tobacco leaves to be baked at the next moment according to the rules set by the model.

5. A tobacco leaf baking automatic control method based on machine vision according to claim 1, characterized in that: In step S4, after the baking is completed, the mapping relationship table is modified and stored according to the baking result, which specifically includes: After the baking is completed, the quality of the cured tobacco leaves is tested and evaluated. If the quality of the cured tobacco leaves does not meet the preset standards, the baking process is reviewed to find the reasons for the decline in baking quality, and the corresponding factor values ​​in the mapping relationship table are modified. The modified mapping relationship table is then stored in the database; if the quality of the cured tobacco leaves meets the standards, no correction is made.

6. A tobacco leaf baking automatic control method based on machine vision according to claim 1, characterized in that: In step S1, the meteorological parameters are real-time temperature and humidity values ​​outside the baking room.

7. A tobacco leaf curing automatic control system based on machine vision, implementing the tobacco leaf curing automatic control method based on machine vision according to claim 1, comprising a controller, a cloud database, an image acquisition module, a data acquisition module, and an actuator; A mapping relationship table is constructed based on historical baking data and expert experience, which includes meteorological parameters, the number, type, moisture content, color value, color change rate, temperature change rate in the baking room, and heating tube current, circulation fan current, and dehumidification fan current. The mapping relationship table is then modified and stored in the cloud database based on the baking results. The controller obtains the input quantity and type of tobacco leaves to be cured, and the current meteorological parameters sent by the data acquisition module, and processes the images collected by the image acquisition module to obtain the moisture content of the tobacco leaves to be cured. The controller then searches the cloud database to determine the corresponding mapping relationship table based on the quantity and type of tobacco leaves to be cured, the current meteorological parameters, and the moisture content of the tobacco leaves to be cured. During the baking process, the controller also calculates the real-time value of the tobacco leaf color, the real-time value of the color change rate, and the real-time value of the temperature change rate in the baking room according to the real-time image collected according to the determined mapping relationship table, and automatically outputs the corresponding control signal to the actuator; The actuator includes a heating tube, a circulation fan, and a dehumidification fan; the corresponding control signals are the heating tube current, the circulation fan current, and the dehumidification fan current.

8. A tobacco leaf baking automatic control system based on machine vision according to claim 7, characterized in that: The data acquisition module includes a temperature and humidity sensor inside the baking room and a temperature and humidity sensor outside the baking room.

Images