A tobacco growth status monitoring system based on image processing
Through the tobacco growth status monitoring system based on image processing, real-time monitoring and evaluation of tobacco growth status is solved, and the problem of lack of real-time monitoring and relying on manual detection in tobacco cultivation is achieved, achieving more efficient resource management and production efficiency.
Patent Information
- Application Number
- CN202410054245.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-15
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-01-15
AI Technical Summary
The lack of real-time monitoring mechanisms in tobacco cultivation leads to a lag understanding of the growth status of tobacco and affects production efficiency. In addition, traditional detection methods rely on manual experience and are prone to errors due to subjective factors. There is a lack of early warning mechanisms, which increases labor costs and resource waste.
Design a tobacco growth status monitoring system based on image processing, including tobacco collection module, feature extraction module, data processing module, model establishment module, evaluation module and decision-making module, and establish through image acquisition, feature extraction, data processing and deep learning models. , calculate the tobacco growth index to achieve real-time monitoring and evaluation of tobacco growth status.
Real-time and accurate monitoring of tobacco growth status is achieved, reducing the burden of manual patrols, improving production efficiency, timely discovering potential problems, avoiding resource waste, and providing more accurate data support and management strategies.
Smart Images

Figure CN117854006B_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of image processing, in particular to a tobacco growth state monitoring system based on image processing. Background Art
[0002] One of the challenges in the current tobacco planting process is the lack of a real-time monitoring mechanism. Traditional agricultural management methods mainly rely on manual patrols and regular inspections, which can easily lead to a delayed understanding of the growth status of tobacco plants. Due to environmental factors, changes in growth status and rapid changes in pests and diseases, delayed monitoring may lead to a slow response to problems, affecting the production efficiency of the farm. Farm managers lack immediate and accurate support in the decision-making process and find it difficult to obtain information about the growth status of tobacco in a timely manner, making it difficult for farm managers to deal with it in a timely manner, such as appropriate irrigation, fertilization and prevention measures. This time difference leads to a lack of flexibility in the tobacco growth process, which may lead to waste of resources and unstable yields.
[0003] In current tobacco production, it is difficult to detect diseases and pests at an early stage. Traditional detection methods often rely on manual experience and are prone to errors due to subjective factors. The lack of an early warning mechanism may result in diseases and pests causing significant losses when they are discovered. Traditional agricultural management usually requires a lot of manual labor, including inspections, data collection, analysis, and decision-making. This not only increases labor costs, but is also easily subject to the limitations of human subjective judgment. The development of an automated tobacco growth status monitoring system can reduce the workload of managers, improve work efficiency, and enable them to focus more on higher-level agricultural decision-making and planning. Summary of the invention
[0004] 1. Technical issues to be solved
[0005] In view of the deficiencies of the prior art, the present invention provides a tobacco growth status monitoring system based on image processing, which solves the problems mentioned in the background technology.
[0006] (II) Technical solution
[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: a tobacco growth status monitoring system based on image processing, including a tobacco collection module, a tobacco feature extraction module, a data processing module, a tobacco model building module, a tobacco evaluation module and a decision module;
[0008] The tobacco collection module collects tobacco images through an image collection device, and cuts the images so that several tobacco plants become an independent plant image block to form a tobacco image set;
[0009] The tobacco feature extraction module extracts features from the tobacco image set, including leaf quantity information, leaf size information, leaf angle information, and leaf surface information;
[0010] The data processing module processes and integrates the blade quantity information and the blade size information to form a first data set, and processes and integrates the blade angle information and the blade surface information to form a second data set;
[0011] The tobacco model building module builds a tobacco growth state training model for the first data set and the second data set through deep learning technology, determines the growth state mark of the current tobacco, and performs overlapping training with a preset tobacco growth state matching model to obtain: tobacco growth index Szzs;
[0012] The tobacco growth index Szzs is obtained by the following calculation formula:
[0013]
[0014] Wherein, Ztxs represents the tobacco stem and leaf growth state coefficient, Ztxs represents the tobacco leaf coefficient, A and B represent the proportional coefficients of the tobacco stem and leaf growth state coefficient Ztxs and the tobacco leaf surface and the tilt state coefficient Zyxs, respectively, and D represents the first correction constant;
[0015] Among them, 0.37≤A≤0.49, 0.39≤B≤0.51, and A+B≤1.0;
[0016] The tobacco stem and leaf growth state coefficient Ztxs is obtained by calculating the first data set, and matched with a preset tobacco growth stem and leaf state threshold G to obtain a tobacco stem and leaf growth evaluation scheme;
[0017] The tobacco leaf surface and tilt state coefficient Zyxs is obtained by calculating the second data set, and matched with the preset tobacco growth leaf threshold Y to obtain a tobacco leaf evaluation scheme;
[0018] The tobacco evaluation module matches the preset tobacco growth state threshold S with the tobacco growth index Szzs to obtain a tobacco growth state level evaluation strategy scheme;
[0019] The decision-making module is specifically implemented by evaluating the content of the strategy plan through the tobacco growth status grade.
[0020] Preferably, the tobacco collection module includes a shooting unit and a cutting unit;
[0021] The shooting unit shoots tobacco plants at a fixed period through a plurality of high-resolution image acquisition devices, and covers the entire tobacco plant by adjusting the positions and angles of the plurality of image acquisition devices, so as to obtain a complete and clear image of the tobacco plant;
[0022] The cutting unit uses an edge detection algorithm on the photographed tobacco plant image to cut the edge between the tobacco plant image and the background, thereby determining the outline of the tobacco plant and forming a plurality of independent and complete tobacco plant image blocks to form a tobacco image set.
[0023] Preferably, the tobacco feature extraction module comprises an extraction unit;
[0024] The extraction unit extracts features from the tobacco image set to obtain leaf quantity information, leaf size information, leaf angle information, and leaf surface information;
[0025] The leaf quantity information is obtained by performing morphological operations on the tobacco image set using image processing technology to extract features;
[0026] Leaf size information was obtained by feature extraction of tobacco image sets using object measurement techniques;
[0027] The leaf angle information is obtained by extracting features from tobacco image sets using polar coordinate conversion technology and ellipse fitting technology;
[0028] The leaf surface information is obtained by extracting features from tobacco image sets using texture analysis technology.
[0029] Preferably, the data processing module includes a first processing unit and a second processing unit;
[0030] The first processing unit performs dimensionless processing on the leaf quantity information and the leaf size information, and then integrates them to form a first data set;
[0031] The first data set includes: a blade quantity value Zyz, a blade perimeter value Zcz, a blade area value Mjz, a blade maximum width value Kmax, and a blade maximum length value Cmax;
[0032] The second processing unit performs dimensionless processing on the blade angle information and the blade surface information, and then integrates them to form a second data set;
[0033] The second data set includes: the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge inflection points Zdsl.
[0034] Preferably, the tobacco model building module includes a modeling unit and an overlapping unit;
[0035] The modeling unit establishes a tobacco growth state training model for the first data set and the second data set by using deep learning technology, and matches the current height and width of the tobacco with the preset tobacco growth state interval value to obtain the tobacco growth state mark corresponding to the current tobacco growth state interval;
[0036] Among them, tobacco growth states include: germination state, seedling state, growth stage state, flowering stage state and maturity stage state;
[0037] The overlapping unit establishes a tobacco growth state matching model through a preset tobacco growth state interval average value, marks the interval tobacco growth state mark, and performs overlapping training with the tobacco growth state training model with the same mark established by the modeling unit, and performs a first calculation to obtain: tobacco stem and leaf growth state coefficient Ztxs and tobacco leaf surface and tilt state coefficient Zyxs, and then performs a second calculation to obtain: tobacco growth index Szzs.
[0038] Preferably, the tobacco stem and leaf growth state coefficient Ztxs is obtained by the following formula:
[0039]
[0040] Wherein, f, g, h, j and k represent the proportional coefficients of the leaf quantity value Zyz, the leaf perimeter value Zcz, the leaf area value Mjz, the leaf maximum width value Kmax and the leaf maximum length value Cmax respectively. The difference between the current tobacco stem growth state and the preset tobacco stem growth state is obtained by calculating the set proportional coefficient and the leaf quantity value Zyz, the leaf perimeter value Zcz, the leaf area value Mjz, the leaf maximum width value Kmax and the leaf maximum length value Cmax for the stem state under different tobacco growth states: tobacco stem and leaf growth state coefficient Ztxs, L represents the second correction constant;
[0041] Among them, 0.13≤f≤0.26, 0.12≤g≤0.21, 0.11≤h≤0.21, 0.11≤j≤0.16, 0.11≤k≤0.16, and f+g+h+j+k≤1.0;
[0042] And match it with the preset tobacco stem and leaf growth state threshold G to obtain the tobacco stem and leaf growth evaluation scheme:
[0043] Tobacco stem and leaf growth state coefficient Ztxs>tobacco stem and leaf growth state threshold G, obtain the qualified evaluation of tobacco stem growth state;
[0044] The tobacco stem and leaf growth status coefficient Ztxs≤the tobacco growth stem and leaf status threshold G, and an unqualified evaluation of the tobacco stem growth status is obtained. The unqualified items include the number of leaves, leaf circumference, leaf area, maximum leaf width and maximum leaf length. When the tobacco stem and leaf growth status coefficient Ztxs<50% of the tobacco growth stem and leaf status threshold G, an abnormal evaluation of the current tobacco plant stem growth status is obtained, and feedback is notified to notify the staff to come and check the current tobacco growth status.
[0045] Preferably, the tobacco leaf surface and tilt state coefficient Zyxs is obtained by the following formula:
[0046]
[0047] Wherein, m, n, p, r and t represent the proportional coefficients of the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge break points Zdsl, respectively. By calculating the set proportional coefficients and the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge break points Zdsl for leaves under different tobacco growth states, the difference between the current tobacco leaf growth state and the preset tobacco leaf growth state is obtained: the tobacco leaf surface and the tilt state coefficient Zyxs, and U represents the third correction constant;
[0048] Among them, 0.13≤m≤19, 0.15≤n≤0.22, 0.15≤p≤0.22, 0.14≤r≤0.21, 0.12≤t≤0.16, and m+n+p+r+t≤1.0;
[0049] And match it with the preset tobacco growth leaf threshold Y to obtain the tobacco leaf evaluation scheme:
[0050] The tobacco leaf surface and tilt state coefficient Zyxs is less than the tobacco growth leaf threshold Y, and the tobacco leaf growth state qualification evaluation is obtained;
[0051] The tobacco leaf surface and tilt state coefficient Zyxs ≥ the tobacco growth leaf threshold Y, and the tobacco leaf growth state unqualified evaluation is obtained. The unqualified items include: average leaf angle, maximum leaf angle, minimum leaf angle, average leaf surface color difference and the number of leaf surface edge inflection points. When the tobacco leaf surface and tilt state coefficient Zyxs ≥ twice the tobacco growth leaf threshold Y, the current tobacco leaf growth state abnormal evaluation is obtained, and feedback is notified to notify the staff to check the current tobacco growth state.
[0052] Preferably, the tobacco evaluation module includes a storage unit and a matching unit;
[0053] The storage unit is used to store tobacco stem and leaf growth state threshold G, tobacco leaf growth threshold Y, tobacco growth state threshold S, tobacco stem and leaf growth evaluation scheme, tobacco leaf evaluation scheme, tobacco growth state level evaluation strategy scheme, tobacco growth state and relevant staff information and notification methods;
[0054] The matching unit matches the preset relevant information with the required comparison value, including matching the preset tobacco growth state threshold S with the tobacco growth index Szzs, to obtain the tobacco growth state level evaluation strategy scheme.
[0055] Preferably, the tobacco growth status grade assessment strategy is obtained by the following matching method:
[0056] Tobacco growth index Szzs<tobacco growth state threshold S, obtain the first-level evaluation of tobacco growth state, and the current tobacco growth state is normal;
[0057] When the tobacco growth index Szzs ≥ the tobacco growth status threshold S, the secondary evaluation of the tobacco growth status is obtained. The current tobacco growth status is abnormal, including abnormal tobacco stalks and abnormal tobacco leaves. The relevant inspectors are reminded to include it in the inspection task. When the tobacco growth index Szzs ≥ twice the tobacco growth status threshold S, the tertiary evaluation of the tobacco growth status is obtained. The current tobacco status is abnormal, including broken stalks, yellowing leaves, foreign objects on the leaves, and abnormal diseases and insects on the leaves. Real-time feedback is provided to the relevant inspectors for investigation, and the tobacco stem and leaf growth evaluation plan and the tobacco leaf evaluation plan are integrated for feedback.
[0058] Preferably, the decision module includes a notification unit;
[0059] The notification unit notifies the inspectors who are on duty in the real-time shift schedule according to the content of the tobacco growth status grade assessment strategy program, and the notification includes SMS notification, broadcast notification and internal software communication prompt.
[0060] (III) Beneficial effects
[0061] The present invention provides a tobacco growth status monitoring system based on image processing, which has the following beneficial effects:
[0062] (1) When the system is running, tobacco images are collected and cut to form a tobacco image set, and feature extraction is performed, including stem quantity information, leaf size information, leaf angle information and leaf surface information. The images are processed and integrated by the data processing module to form a first data set and a second data set. A tobacco growth state training model is established, and the current tobacco growth state mark is determined. The images are overlapped with the preset tobacco growth state matching model to obtain: tobacco growth index Szzs. The tobacco growth state threshold S preset by the tobacco evaluation module is matched to obtain a tobacco growth state level evaluation strategy. Finally, the decision module is used for specific execution, so as to achieve real-time monitoring of tobacco growth state. By using real-time data, inspectors can quickly understand the current status of tobacco, take timely measures to prevent potential problems, and obtain comprehensive tobacco characteristics including stem quantity, leaf size, leaf angle and leaf surface information, so as to provide a more accurate basis for the evaluation of tobacco growth state, thereby avoiding frequent manual inspections of tobacco growth state and reducing manual burden, avoiding waste of resources, and enabling them to focus more on dealing with abnormal situations that really need attention.
[0063] (2) Through the preset tobacco growth states of germination, seedling, growth period, flowering and maturity, and the tobacco growth state interval value is set according to the interval value of the minimum and maximum tobacco height and width in the specific state, a detailed division of each growth state is achieved. Such a detailed division helps to understand the development stage of tobacco more accurately, improve the grasp and monitoring of tobacco growth status, and realize comprehensive, detailed and real-time monitoring and evaluation of tobacco growth status, providing decision makers with more accurate data support and monitoring feedback.
[0064] (3) The tobacco stem and leaf growth state coefficient Ztxs and the tobacco leaf surface and tilt state coefficient Zyxs are matched with the tobacco growth stem and leaf state threshold G and the tobacco growth leaf threshold Y to obtain the tobacco stem and leaf growth evaluation scheme and the tobacco leaf evaluation scheme, and the tobacco growth state threshold S is matched with the tobacco growth index Szzs to obtain the tobacco growth state grade evaluation strategy scheme, which can effectively deal with various situations of tobacco under different growth states, and then comprehensively evaluate the growth state of tobacco, which is helpful to formulate a more comprehensive and effective tobacco management strategy, and at the same time provide specific abnormality troubleshooting suggestions, which helps decision makers better understand the tobacco growth state and take corresponding management and treatment measures. BRIEF DESCRIPTION OF THE DRAWINGS
[0065] Figure 1 The present invention is a schematic diagram of a tobacco growth status monitoring system based on image processing. DETAILED DESCRIPTION
[0066] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.
[0067] One of the challenges in the current tobacco planting process is the lack of a real-time monitoring mechanism. Traditional agricultural management methods mainly rely on manual patrols and regular inspections, which can easily lead to a delayed understanding of the growth status of tobacco plants. Due to environmental factors, changes in growth status and rapid changes in pests and diseases, delayed monitoring may lead to a slow response to problems, affecting the production efficiency of the farm. Farm managers lack immediate and accurate support in the decision-making process and find it difficult to obtain information about the growth status of tobacco in a timely manner, making it difficult for farm managers to deal with it in a timely manner, such as appropriate irrigation, fertilization and prevention measures. This time difference leads to a lack of flexibility in the tobacco growth process, which may lead to waste of resources and unstable yields.
[0068] In current tobacco production, it is difficult to detect diseases and pests at an early stage. Traditional detection methods often rely on manual experience and are prone to errors due to subjective factors. The lack of an early warning mechanism may result in diseases and pests causing significant losses when they are discovered. Traditional agricultural management usually requires a lot of manual labor, including inspections, data collection, analysis, and decision-making. This not only increases labor costs, but is also easily subject to the limitations of human subjective judgment. The development of an automated tobacco growth status monitoring system can reduce the workload of managers, improve work efficiency, and enable them to focus more on higher-level agricultural decision-making and planning.
[0069] Example 1
[0070] The present invention provides a tobacco growth status monitoring system based on image processing, please refer to Figure 1 , including tobacco collection module, tobacco feature extraction module, data processing module, tobacco model building module, tobacco evaluation module and decision-making module;
[0071] The tobacco collection module collects tobacco images through an image collection device, and cuts the images so that several tobacco plants become an independent plant image block to form a tobacco image set;
[0072] The tobacco feature extraction module extracts features from the tobacco image set, including leaf quantity information, leaf size information, leaf angle information, and leaf surface information;
[0073] The data processing module processes and integrates the blade quantity information and the blade size information to form a first data set, and processes and integrates the blade angle information and the blade surface information to form a second data set;
[0074] The tobacco model building module builds a tobacco growth state training model for the first data set and the second data set through deep learning technology, determines the growth state mark of the current tobacco, and performs overlapping training with a preset tobacco growth state matching model to obtain: tobacco growth index Szzs;
[0075] The tobacco growth index Szzs is obtained by the following calculation formula:
[0076]
[0077] Wherein, Ztxs represents the tobacco stem and leaf growth state coefficient, Ztxs represents the tobacco leaf coefficient, A and B represent the proportional coefficients of the tobacco stem and leaf growth state coefficient Ztxs and the tobacco leaf surface and the tilt state coefficient Zyxs, respectively, and D represents the first correction constant;
[0078] Among them, 0.37≤A≤0.49, 0.39≤B≤0.51, and A+B≤1.0;
[0079] The tobacco stem and leaf growth state coefficient Ztxs is obtained by calculating the first data set, and matched with a preset tobacco growth stem and leaf state threshold G to obtain a tobacco stem and leaf growth evaluation scheme;
[0080] The tobacco leaf surface and tilt state coefficient Zyxs is obtained by calculating the second data set, and matched with the preset tobacco growth leaf threshold Y to obtain a tobacco leaf evaluation scheme;
[0081] The tobacco evaluation module matches the preset tobacco growth state threshold S with the tobacco growth index Szzs to obtain a tobacco growth state level evaluation strategy scheme;
[0082] The decision-making module is specifically implemented by evaluating the content of the strategy plan through the tobacco growth status grade.
[0083] In this embodiment, tobacco images are collected by the tobacco collection module, and the images are cut into independent plant image blocks to form a tobacco image set. The tobacco feature extraction module extracts features of the tobacco image set, including leaf size information, leaf angle information and leaf surface information. The data processing module processes and integrates them to form a first data set and a second data set. The tobacco model establishment module establishes a tobacco growth state training model for the first data set and the second data set, determines the current tobacco growth state mark, and performs overlapping training with the preset tobacco growth state matching model to obtain: tobacco growth index Szzs, and the tobacco evaluation model is used to establish a tobacco growth state training model for the first data set and the second data set. The tobacco growth status threshold S preset by the block is matched to obtain the tobacco growth status grade assessment strategy plan, and finally it is specifically executed through the decision-making module, so as to achieve real-time monitoring of the tobacco growth status. By using real-time data, inspectors can quickly understand the current status of tobacco, take timely measures to prevent potential problems, and obtain comprehensive characteristics of tobacco including stem number, leaf size, leaf angle and leaf surface information, so as to provide a more accurate basis for the assessment of tobacco growth status, thereby avoiding frequent manual inspections of tobacco growth status, reducing manual burden, avoiding waste of resources, and enabling them to focus more on dealing with abnormal situations that really need attention.
[0084] Example 2
[0085] This embodiment is explained in Example 1, please refer to Figure 1 ,Specifically: the tobacco collection module includes a shooting unit and a cutting unit;
[0086] The shooting unit shoots tobacco plants at a fixed period through a plurality of high-resolution image acquisition devices, and covers the entire tobacco plant by adjusting the positions and angles of the plurality of image acquisition devices, so as to obtain a complete and clear image of the tobacco plant;
[0087] The cutting unit uses an edge detection algorithm on the photographed tobacco plant image to cut the edge between the tobacco plant image and the background, thereby determining the outline of the tobacco plant and forming a plurality of independent and complete tobacco plant image blocks to form a tobacco image set.
[0088] The tobacco feature extraction module includes an extraction unit;
[0089] The extraction unit extracts features from the tobacco image set to obtain leaf quantity information, leaf size information, leaf angle information, and leaf surface information;
[0090] The leaf quantity information is obtained by performing morphological operations on the tobacco image set using image processing technology to extract features;
[0091] Leaf size information was obtained by feature extraction of tobacco image sets using object measurement techniques;
[0092] The leaf angle information is obtained by extracting features from tobacco image sets using polar coordinate conversion technology and ellipse fitting technology;
[0093] The leaf surface information is obtained by extracting features from tobacco image sets using texture analysis technology.
[0094] Example 3
[0095] This embodiment is explained in Example 1, please refer to Figure 1 , specifically: the data processing module includes a first processing unit and a second processing unit;
[0096] The first processing unit performs dimensionless processing on the leaf quantity information and the leaf size information, and then integrates them to form a first data set;
[0097] The first data set includes: a blade quantity value Zyz, a blade perimeter value Zcz, a blade area value Mjz, a blade maximum width value Kmax, and a blade maximum length value Cmax;
[0098] The second processing unit performs dimensionless processing on the blade angle information and the blade surface information, and then integrates them to form a second data set;
[0099] The second data set includes: the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge inflection points Zdsl.
[0100] The tobacco model building module includes a modeling unit and an overlapping unit;
[0101] The modeling unit establishes a tobacco growth state training model for the first data set and the second data set by using deep learning technology, and matches the current height and width of the tobacco with the preset tobacco growth state interval value to obtain the tobacco growth state mark corresponding to the current tobacco growth state interval;
[0102] Among them, tobacco growth states include: germination state, seedling state, growth stage state, flowering stage state and maturity stage state;
[0103] Tobacco growth state interval value: the interval value of the minimum and maximum tobacco height and width in the specific state of tobacco growth, specifically: the minimum height value and the minimum width value of the germination state are used as the lower limit of the germination state interval value in the tobacco growth state, and the maximum height value and the maximum width value of the germination state are used as the upper limit of the germination state interval value in the tobacco growth state, and the tobacco growth state interval value of the germination state is composed of the lower limit and the upper limit;
[0104] The overlapping unit establishes a tobacco growth state matching model through a preset tobacco growth state interval average value, marks the interval tobacco growth state mark, and performs overlapping training with the tobacco growth state training model with the same mark established by the modeling unit, and performs a first calculation to obtain: tobacco stem and leaf growth state coefficient Ztxs and tobacco leaf surface and tilt state coefficient Zyxs, and then performs a second calculation to obtain: tobacco growth index Szzs.
[0105] In this embodiment, by pre-setting the tobacco growth states of germination state, seedling state, growth stage state, flowering stage state and maturity stage state, and setting the tobacco growth state interval value according to the interval value of the minimum and maximum tobacco height and width in the specific state, a detailed division of each growth state is achieved. Such a detailed division helps to understand the development stage of tobacco more accurately, improve the grasp and monitoring of the tobacco growth state, and realize comprehensive, detailed and real-time monitoring and evaluation of the tobacco growth state, providing decision makers with more accurate data support and monitoring feedback.
[0106] Example 4
[0107] This embodiment is explained in Example 1, please refer to Figure 1 Specifically: the tobacco stem and leaf growth state coefficient Ztxs is obtained by the following formula:
[0108]
[0109] Wherein, f, g, h, j and k represent the proportional coefficients of the leaf quantity value Zyz, the leaf perimeter value Zcz, the leaf area value Mjz, the leaf maximum width value Kmax and the leaf maximum length value Cmax respectively. The difference between the current tobacco stem growth state and the preset tobacco stem growth state is obtained by calculating the set proportional coefficient and the leaf quantity value Zyz, the leaf perimeter value Zcz, the leaf area value Mjz, the leaf maximum width value Kmax and the leaf maximum length value Cmax for the stem state under different tobacco growth states: tobacco stem and leaf growth state coefficient Ztxs, L represents the second correction constant;
[0110] Among them, 0.13≤f≤0.26, 0.12≤g≤0.21, 0.11≤h≤0.21, 0.11≤j≤0.16, 0.11≤k≤0.16, and f+g+h+j+k≤1.0;
[0111] And match it with the preset tobacco stem and leaf growth state threshold G to obtain the tobacco stem and leaf growth evaluation scheme:
[0112] Tobacco stem and leaf growth state coefficient Ztxs>tobacco stem and leaf growth state threshold G, obtain the qualified evaluation of tobacco stem growth state;
[0113] The tobacco stem and leaf growth status coefficient Ztxs≤the tobacco growth stem and leaf status threshold G, and an unqualified evaluation of the tobacco stem growth status is obtained. The unqualified items include the number of leaves, leaf circumference, leaf area, maximum leaf width and maximum leaf length. When the tobacco stem and leaf growth status coefficient Ztxs<50% of the tobacco growth stem and leaf status threshold G, an abnormal evaluation of the current tobacco plant stem growth status is obtained, and feedback is notified to notify the staff to come and check the current tobacco growth status.
[0114] The tobacco leaf surface and tilt state coefficient Zyxs is obtained by the following formula:
[0115]
[0116] Wherein, m, n, p, r and t represent the proportional coefficients of the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge break points Zdsl, respectively. By calculating the set proportional coefficients and the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge break points Zdsl for leaves under different tobacco growth states, the difference between the current tobacco leaf growth state and the preset tobacco leaf growth state is obtained: the tobacco leaf surface and the tilt state coefficient Zyxs, and U represents the third correction constant;
[0117] Among them, 0.13≤m≤19, 0.15≤n≤0.22, 0.15≤p≤0.22, 0.14≤r≤0.21, 0.12≤t≤0.16, and m+n+p+r+t≤1.0;
[0118] And match it with the preset tobacco growth leaf threshold Y to obtain the tobacco leaf evaluation scheme:
[0119] The tobacco leaf surface and tilt state coefficient Zyxs is less than the tobacco growth leaf threshold Y, and the tobacco leaf growth state qualification evaluation is obtained;
[0120] The tobacco leaf surface and tilt state coefficient Zyxs ≥ the tobacco growth leaf threshold Y, and the tobacco leaf growth state unqualified evaluation is obtained. The unqualified items include: average leaf angle, maximum leaf angle, minimum leaf angle, average leaf surface color difference and the number of leaf surface edge inflection points. When the tobacco leaf surface and tilt state coefficient Zyxs ≥ twice the tobacco growth leaf threshold Y, the current tobacco leaf growth state abnormal evaluation is obtained, and feedback is notified to notify the staff to check the current tobacco growth state.
[0121] The tobacco evaluation module includes a storage unit and a matching unit;
[0122] The storage unit is used to store tobacco stem and leaf growth state threshold G, tobacco leaf growth threshold Y, tobacco growth state threshold S, tobacco stem and leaf growth evaluation scheme, tobacco leaf evaluation scheme, tobacco growth state level evaluation strategy scheme, tobacco growth state and relevant staff information and notification methods;
[0123] The matching unit matches the preset relevant information with the required comparison value, including matching the preset tobacco growth state threshold S with the tobacco growth index Szzs, to obtain the tobacco growth state level evaluation strategy scheme.
[0124] The tobacco growth status grade evaluation strategy scheme is obtained by the following matching method:
[0125] Tobacco growth index Szzs<tobacco growth state threshold S, obtain the first-level evaluation of tobacco growth state, and the current tobacco growth state is normal;
[0126] When the tobacco growth index Szzs ≥ the tobacco growth status threshold S, the secondary evaluation of the tobacco growth status is obtained. The current tobacco growth status is abnormal, including abnormal tobacco stalks and abnormal tobacco leaves. The relevant inspectors are reminded to include it in the inspection task. When the tobacco growth index Szzs ≥ twice the tobacco growth status threshold S, the tertiary evaluation of the tobacco growth status is obtained. The current tobacco status is abnormal, including broken stalks, yellowing leaves, foreign objects on the leaves, and abnormal diseases and insects on the leaves. Real-time feedback is provided to the relevant inspectors for investigation, and the tobacco stem and leaf growth evaluation plan and the tobacco leaf evaluation plan are integrated for feedback.
[0127] The decision module includes a notification unit;
[0128] The notification unit notifies the inspectors who are on duty in the real-time shift schedule according to the content of the tobacco growth status grade assessment strategy program, and the notification includes SMS notification, broadcast notification and internal software communication prompt.
[0129] In this embodiment, the tobacco stem and leaf growth state coefficient Ztxs and the tobacco leaf surface and inclination state coefficient Zyxs are matched with the tobacco growth stem and leaf state threshold G and the tobacco growth leaf threshold Y to obtain the tobacco stem and leaf growth evaluation scheme and the tobacco leaf evaluation scheme, and the tobacco growth state threshold S is matched with the tobacco growth index Szzs to obtain the tobacco growth state level evaluation strategy scheme, which can effectively respond to various situation judgments of tobacco under different growth states, and then comprehensively evaluate the growth state of tobacco, which is helpful to formulate a more comprehensive and effective tobacco management strategy, and at the same time provide specific abnormality troubleshooting suggestions, which helps decision makers better understand the tobacco growth state and take corresponding management and treatment measures.
[0130] Specific example: A tobacco growth status monitoring system based on image processing used in a certain planting base.
[0131] Assume the following parameter values:
[0132] The first data set includes: blade number value Zyz: 12, blade circumference value Zcz: 30, blade area value Mjz: 60, blade maximum width value Kmax: 5, and blade maximum length value Cmax: 12;
[0133] The second data set includes: average leaf angle value Pjjd: 35, maximum leaf angle value Jmax: 55, minimum leaf angle value Jmin: 25, average leaf surface color difference value Spjz: 10, and number of leaf surface edge inflection points Zdsl: 8;
[0134] The second correction constant L: 0.09, proportional coefficients: f: 0.20, g: 0.17, h: 0.17, j: 0.13 and k: 0.13;
[0135] Obtained according to the calculation formula of tobacco stem and leaf growth state coefficient Ztxs:
[0136] Ztxs=[(0.20*12)+(0.17*30)+(0.17*60)+(0.13*5)+(0.13*12)]+0.09=20;
[0137] The tobacco stem and leaf growth state threshold G is set to 9, and matched with the tobacco stem and leaf growth state coefficient Ztxs, to obtain: tobacco stem and leaf growth state coefficient Ztxs>tobacco stem and leaf growth state threshold G, to obtain the tobacco stem growth state qualified evaluation;
[0138] The third correction constant U: 0.8, proportional coefficients: m: 0.18, n: 0.21, p: 0.21, r: 0.19 and t: 0.15;
[0139] Obtained according to the calculation formula of tobacco leaf surface and tilt state coefficient Zyxs:
[0140] Ztxs=[(0.18*35)+(0.21*55)+(0.21*25)+(0.19*10)+(0.15*8)]+0.8=27;
[0141] The tobacco leaf growth threshold value Y is set to 49, and matched with the tobacco leaf surface and tilt state coefficient Zyxs, to obtain: the tobacco leaf surface and tilt state coefficient Zyxs < tobacco leaf growth threshold value Y, and obtain the tobacco leaf growth state qualified evaluation;
[0142] First correction constant D: 0.17, proportional coefficients: A: 0.48 and B: 0.49;
[0143] Obtained according to the calculation formula of tobacco growth index Szzs:
[0144] Szzs=[(0.48*20)+(0.49*27)]+0.81=23;
[0145] The tobacco growth state threshold S is set to 49, and matched with the tobacco growth index Szzs to obtain: tobacco growth index Szzs<tobacco growth state threshold S, and the first-level evaluation of tobacco growth state is obtained. There is no abnormality in the current tobacco growth state.
[0146] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A tobacco growth status monitoring system based on image processing, Features: It includes tobacco collection module, tobacco feature extraction module, data processing module, tobacco model building module, tobacco evaluation module and decision-making module; The tobacco collection module collects tobacco images through an image collection device, and cuts the images so that several tobacco plants become an independent plant image block to form a tobacco image set; The tobacco feature extraction module extracts features from the tobacco image set, including leaf quantity information, leaf size information, leaf angle information, and leaf surface information; The data processing module processes and integrates the blade quantity information and the blade size information to form a first data set, and processes and integrates the blade angle information and the blade surface information to form a second data set; The tobacco model building module builds a tobacco growth state training model for the first data set and the second data set through deep learning technology, determines the growth state mark of the current tobacco, and performs overlapping training with a preset tobacco growth state matching model to obtain: tobacco growth index Szzs; The tobacco growth index Szzs is obtained by the following calculation formula: Wherein, Ztxs represents the tobacco stem and leaf growth state coefficient, Ztxs represents the tobacco leaf coefficient, A and B represent the proportional coefficients of the tobacco stem and leaf growth state coefficient Ztxs and the tobacco leaf surface and the tilt state coefficient Zyxs, respectively, and D represents the first correction constant; The tobacco stem and leaf growth state coefficient Ztxs is obtained by calculating the first data set, and matched with a preset tobacco growth stem and leaf state threshold G to obtain a tobacco stem and leaf growth evaluation scheme; The tobacco stem and leaf growth state coefficient Ztxs is obtained by the following formula: Wherein, f, g, h, j and k represent the proportional coefficients of the leaf quantity value Zyz, the leaf perimeter value Zcz, the leaf area value Mjz, the leaf maximum width value Kmax and the leaf maximum length value Cmax respectively. The difference between the current tobacco stem growth state and the preset tobacco stem growth state is obtained by calculating the set proportional coefficient and the leaf quantity value Zyz, the leaf perimeter value Zcz, the leaf area value Mjz, the leaf maximum width value Kmax and the leaf maximum length value Cmax for the stem state under different tobacco growth states: tobacco stem and leaf growth state coefficient Ztxs, L represents the second correction constant; Among them, 0.13≤f≤0.26, 0.12≤g≤0.21, 0.11≤h≤0.21, 0.11≤j≤0.16, 0.11≤k≤0.16, and f+g+h+j+k≤1.0; And match it with the preset tobacco stem and leaf growth state threshold G to obtain the tobacco stem and leaf growth evaluation scheme: Tobacco stem and leaf growth state coefficient Ztxs>tobacco stem and leaf growth state threshold G, obtain the qualified evaluation of tobacco stem growth state; The tobacco stem and leaf growth state coefficient Ztxs ≤ the tobacco growth stem and leaf state threshold G, and the unqualified evaluation of the tobacco stem growth state is obtained. The unqualified items include the number of leaves, the circumference of the leaves, the area of the leaves, the maximum width of the leaves, and the maximum length of the leaves. When the tobacco stem and leaf growth state coefficient Ztxs is less than 50% of the tobacco growth stem and leaf state threshold G, the abnormal evaluation of the current tobacco plant stem growth state is obtained, and feedback is notified to notify the staff to come to check the current tobacco growth state; The tobacco leaf surface and tilt state coefficient Zyxs is obtained by calculating the second data set, and matched with a preset tobacco growth leaf threshold Y to obtain a tobacco leaf evaluation scheme; The tobacco leaf surface and tilt state coefficient Zyxs is obtained by the following formula: Wherein, m, n, p, r and t represent the proportional coefficients of the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge break points Zdsl, respectively. By calculating the set proportional coefficients and the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge break points Zdsl for leaves under different tobacco growth states, the difference between the current tobacco leaf growth state and the preset tobacco leaf growth state is obtained: the tobacco leaf surface and the tilt state coefficient Zyxs, and U represents the third correction constant; Among them, 0.13≤m≤19, 0.15≤n≤0.22, 0.15≤p≤0.22, 0.14≤r≤0.21, 0.12≤t≤0.16, and m+n+p+r+t≤1.0; And match it with the preset tobacco growth leaf threshold Y to obtain the tobacco leaf evaluation scheme: The tobacco leaf surface and tilt state coefficient Zyxs is less than the tobacco growth leaf threshold Y, and the tobacco leaf growth state qualification evaluation is obtained; The tobacco leaf surface and tilt state coefficient Zyxs ≥ tobacco growth leaf threshold Y, obtain the unqualified evaluation of tobacco leaf growth state, the unqualified items include: average leaf angle, maximum leaf angle, minimum leaf angle, average leaf surface color difference and number of leaf surface edge inflection points. When the tobacco leaf surface and tilt state coefficient Zyxs ≥ twice the tobacco growth leaf threshold Y, obtain the abnormal evaluation of the current tobacco leaf growth state, and feedback to notify the staff to check the current tobacco growth state; The tobacco evaluation module matches the preset tobacco growth state threshold S with the tobacco growth index Szzs to obtain a tobacco growth state level evaluation strategy scheme; The decision-making module is specifically implemented by evaluating the content of the strategy plan through the tobacco growth status grade.
2. A tobacco growth status monitoring system based on image processing according to claim 1, Features: The tobacco collection module includes a shooting unit and a cutting unit; The shooting unit shoots tobacco plants at a fixed period through a plurality of high-resolution image acquisition devices, and covers the entire tobacco plant by adjusting the positions and angles of the plurality of image acquisition devices, so as to obtain a complete and clear image of the tobacco plant; The cutting unit uses an edge detection algorithm on the photographed tobacco plant image to cut the edge between the tobacco plant image and the background, thereby determining the outline of the tobacco plant and forming a plurality of independent and complete tobacco plant image blocks to form a tobacco image set.
3. A tobacco growth status monitoring system based on image processing according to claim 1, Features: The tobacco feature extraction module includes an extraction unit; The extraction unit extracts features from the tobacco image set to obtain leaf quantity information, leaf size information, leaf angle information, and leaf surface information; The leaf quantity information is obtained by performing morphological operations on the tobacco image set using image processing technology to extract features; Leaf size information was obtained by feature extraction of tobacco image sets using object measurement techniques; The leaf angle information is obtained by extracting features from tobacco image sets using polar coordinate conversion technology and ellipse fitting technology; The leaf surface information is obtained by extracting features from tobacco image sets using texture analysis technology.
4. The tobacco growth status monitoring system based on image processing according to claim 1, Features: The data processing module includes a first processing unit and a second processing unit; The first processing unit performs dimensionless processing on the leaf quantity information and the leaf size information, and then integrates them to form a first data set; The first data set includes: a blade quantity value Zyz, a blade perimeter value Zcz, a blade area value Mjz, a blade maximum width value Kmax, and a blade maximum length value Cmax; The second processing unit performs dimensionless processing on the blade angle information and the blade surface information, and then integrates them to form a second data set; The second data set includes: the average leaf angle value Pjjd, the maximum leaf angle value Jmax, the minimum leaf angle value Jmin, the average leaf surface color difference value Spjz and the number of leaf surface edge inflection points Zdsl.
5. The tobacco growth status monitoring system based on image processing according to claim 1, Features: The tobacco model building module includes a modeling unit and an overlapping unit; The modeling unit establishes a tobacco growth state training model for the first data set and the second data set by using deep learning technology, and matches the current height and width of the tobacco with the preset tobacco growth state interval value to obtain the tobacco growth state mark corresponding to the current tobacco growth state interval; Among them, tobacco growth states include: germination state, seedling state, growth stage state, flowering stage state and maturity stage state; The overlapping unit establishes a tobacco growth state matching model through a preset tobacco growth state interval average value, marks the interval tobacco growth state mark, and performs overlapping training with the tobacco growth state training model with the same mark established by the modeling unit, and performs a first calculation to obtain: tobacco stem and leaf growth state coefficient Ztxs and tobacco leaf surface and tilt state coefficient Zyxs, and then performs a second calculation to obtain: tobacco growth index Szzs.
6. A tobacco growth status monitoring system based on image processing according to claim 1, Features: The tobacco evaluation module includes a storage unit and a matching unit; The storage unit is used to store tobacco stem and leaf growth state threshold G, tobacco leaf growth threshold Y, tobacco growth state threshold S, tobacco stem and leaf growth evaluation scheme, tobacco leaf evaluation scheme, tobacco growth state level evaluation strategy scheme, tobacco growth state and relevant staff information and notification methods; The matching unit matches the preset relevant information with the required comparison value, including matching the preset tobacco growth state threshold S with the tobacco growth index Szzs, to obtain the tobacco growth state level evaluation strategy scheme.
7. A tobacco growth status monitoring system based on image processing according to claim 6, Features: The tobacco growth status grade evaluation strategy scheme is obtained by the following matching method: Tobacco growth index Szzs<tobacco growth state threshold S, obtain the first-level evaluation of tobacco growth state, and the current tobacco growth state is normal; When the tobacco growth index Szzs ≥ the tobacco growth status threshold S, the secondary evaluation of the tobacco growth status is obtained. The current tobacco growth status is abnormal, including abnormal tobacco stalks and abnormal tobacco leaves. The relevant inspectors are reminded to include it in the inspection task. When the tobacco growth index Szzs ≥ twice the tobacco growth status threshold S, the tertiary evaluation of the tobacco growth status is obtained. The current tobacco status is abnormal, including broken stalks, yellowing leaves, foreign objects on the leaves, and abnormal diseases and insects on the leaves. Real-time feedback is provided to the relevant inspectors for investigation, and the tobacco stem and leaf growth evaluation plan and the tobacco leaf evaluation plan are integrated for feedback.
8. A tobacco growth status monitoring system based on image processing according to claim 7, Features: The decision module includes a notification unit; The notification unit notifies the inspectors who are on duty in the real-time shift schedule according to the content of the tobacco growth status grade assessment strategy program, and the notification includes SMS notification, broadcast notification and internal software communication prompt.
