A method, medium and system for optimizing tobacco growth period configuration to improve quality

Abstract

The present invention relates to the technical field of tobacco planting and management, and specifically discloses a method, medium, and system for optimizing tobacco growth period configuration to improve quality. The method comprises obtaining meteorological data containing time tags of tobacco samples during their growth process using sensors installed in the planting environment to obtain a meteorological data matrix; obtaining tobacco images containing time tags of the tobacco samples; registering the tobacco images based on the time tags, comparing and clustering the registered tobacco images to obtain tobacco images of different classes; and for tobacco images of the same class, intercepting corresponding meteorological data matrices based on the time tags, comparing the meteorological data matrices, and determining the importance of each parameter. The present invention obtains historical environmental data and tobacco images of the tobacco samples, clusters the tobacco images, and then clusters the historical environmental data. Comparing similar historical environmental data, the importance of each parameter is determined, providing a reference for the resource allocation process.

Description

Technical Field

[0001] The present invention relates to the technical field of tobacco planting and management, and in particular to a method, medium and system for optimizing tobacco growth period configuration for improving quality. Background Art

[0002] The tobacco growth period includes the sowing period (seed germination period), seedling period, growth period, flower bud differentiation period, flowering period, maturity period and harvest period. During the tobacco growth process, different environmental parameters may form different growth states. In the existing planting scenarios, the environmental parameters are adjustable. Adjusting the environmental parameters involves the issue of resource allocation. How to optimize the resource allocation process is the technical problem that the technical solution of the present invention aims to solve. Summary of the Invention

[0003] The object of the present invention is to provide a method, medium and system for optimizing the configuration of tobacco growth periods to improve quality, so as to solve the problems raised in the above background technology.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] A method for optimizing tobacco growth period configuration to improve quality, the method comprising:

[0006] Receiving the target growth period, obtaining tobacco samples within the target growth period;

[0007] The meteorological data containing time tags of tobacco samples during the growth process are obtained by sensors installed in the planting environment to obtain a meteorological data matrix;

[0008] Acquire tobacco images containing time tags of tobacco samples;

[0009] Tobacco images are registered based on time tags, and the registered tobacco images are compared and clustered to obtain tobacco images of different classes;

[0010] For tobacco images of the same type, the corresponding meteorological data matrix is ​​intercepted according to the time tag, and the meteorological data matrix is ​​compared to determine the importance of each parameter; the importance represents the degree of influence of each parameter in the process of tobacco growing into the growth state corresponding to the tobacco image of the same type;

[0011] When receiving the target image input by the user, the class corresponding to the image is determined based on the same image comparison scheme, and the importance of each parameter corresponding to the tobacco image of this class is queried;

[0012] Environmental control resources are configured based on the importance.

[0013] As a further solution of the present invention, the step of obtaining meteorological data containing time tags of tobacco samples during the growth process according to sensors installed in the planting environment to obtain a meteorological data matrix includes:

[0014] Acquiring meteorological data containing time tags of tobacco samples during their growth process based on sensors installed in the planting environment; the meteorological data includes at least air temperature, light intensity, soil temperature, and soil moisture;

[0015] Arrange various meteorological data with the same time tag in order according to the preset indicators to obtain row data;

[0016] Arrange the row data in chronological order according to the time labels to obtain the meteorological data matrix.

[0017] As a further solution of the present invention, the steps of registering tobacco images based on time tags and comparing and clustering the registered tobacco images to obtain tobacco images of different classes include:

[0018] Register tobacco images based on time tags;

[0019] For tobacco images containing the same time label, feature points are marked in the tobacco images;

[0020] Compare the labeling results of different tobacco images and calculate the image distance;

[0021] The tobacco images are clustered according to the image distance to obtain tobacco images of different categories.

[0022] As a further solution of the present invention: for tobacco images containing the same time tag, the step of marking feature points in the tobacco image includes:

[0023] Extract the layers of the tobacco image under different channels, traverse the pixels in each layer, record the pixel values, and construct an array; the traversal order of the pixels in all layers is the same;

[0024] For each array, calculate the independent value of each value in sequence; the independent value is used to represent the degree of difference between each value and other values ​​in the array;

[0025] Accumulate independent values. When the independent value reaches a preset value, query the position corresponding to the value and mark it as a feature point.

[0026] The calculation process of the independent value is: Where, is an independent value, For the preset parameters, is the larger value of the distance between the current element and the two ends of the array, The distance from the current element in the array is The difference between the value of the element that is being retrieved and the value of the current element.

[0027] As a further solution of the present invention, the step of intercepting corresponding meteorological data matrices according to time tags for similar tobacco images, comparing the meteorological data matrices, and determining the importance of each parameter includes:

[0028] For any tobacco image of the same type, locate the row data in the meteorological data matrix according to the time label;

[0029] Get all row data before the located row data as the intercepted meteorological data matrix;

[0030] Compare each intercepted meteorological data matrix to determine the importance of each parameter;

[0031] The calculation process of the importance of each parameter is as follows:

[0032] For any parameter, extract its corresponding column in each intercepted meteorological data matrix;

[0033] Calculate the cosine similarity between each pair in the extracted columns, and then calculate the mean of the cosine similarity;

[0034] The importance is determined according to the positive ratio of the mean; the relationship between the importance and the mean is a linear function with a positive coefficient.

[0035] As a further embodiment of the present invention, the method further comprises:

[0036] Receive the tobacco to be optimized selected by the user, record the meteorological data in real time, and obtain the real-time meteorological matrix;

[0037] Obtain the meteorological data matrix with the time label as the end moment as the reference matrix;

[0038] The real-time meteorological matrix is ​​compared with the reference matrix in real time, the similarity is calculated, and the reference matrix whose similarity reaches the preset threshold is selected to obtain its corresponding tobacco image as the predicted image.

[0039] The technical solution of the present invention also provides a tobacco growth period optimization configuration system for improving quality, the system comprising:

[0040] A tobacco sample acquisition module is used to receive a target growth period and obtain tobacco samples within the target growth period;

[0041] A meteorological data statistics module is used to obtain meteorological data with time tags of tobacco samples during the growth process based on sensors installed in the planting environment to obtain a meteorological data matrix;

[0042] A tobacco image acquisition module is used to acquire tobacco images containing time tags of tobacco samples;

[0043] The tobacco image clustering module is used to register tobacco images based on time tags, compare and cluster the registered tobacco images, and obtain tobacco images of different classes;

[0044] A parameter analysis module is used to intercept the corresponding meteorological data matrix according to the time tag for the tobacco image of the same type, compare the meteorological data matrix, and determine the importance of each parameter; the importance indicates the degree of influence of each parameter in the process of tobacco growing into the growth state corresponding to the tobacco image of the same type;

[0045] The importance query module is used to determine the class of the image based on the same image comparison scheme when receiving the target image input by the user, and query the importance of each parameter corresponding to the tobacco image of this class;

[0046] A resource configuration module is used to configure environmental control resources based on the importance.

[0047] As a further solution of the present invention: the meteorological data statistics module includes:

[0048] A sensor monitoring unit, configured to obtain time-tagged meteorological data of tobacco samples during their growth process based on sensors installed in the growing environment; the meteorological data includes at least air temperature, light intensity, soil temperature, and soil moisture;

[0049] A row data generating unit is used to arrange various meteorological data with the same time tag in a sequence according to a preset index to obtain row data;

[0050] The column data generating unit is used to arrange the row data according to the time sequence of the time tags to obtain the meteorological data matrix.

[0051] As a further solution of the present invention: the tobacco image clustering module includes:

[0052] an image registration unit, for registering tobacco images based on time tags;

[0053] A feature extraction unit, for marking feature points in tobacco images containing the same time tag;

[0054] A distance calculation unit, used to compare the labeling results of different tobacco images and calculate the image distance;

[0055] The clustering execution unit is used to cluster the tobacco images according to the image distance to obtain tobacco images of different categories.

[0056] As a further solution of the present invention: the parameter analysis module includes:

[0057] a row data locating unit, for locating row data of any tobacco image of the same type in the meteorological data matrix according to a time tag;

[0058] A data interception unit is used to obtain all row data before the located row data as the intercepted meteorological data matrix;

[0059] Matrix comparison unit, used to compare each intercepted meteorological data matrix and determine the importance of each parameter;

[0060] The calculation process of the importance of each parameter is as follows:

[0061] For any parameter, extract its corresponding column in each intercepted meteorological data matrix;

[0062] Calculate the cosine similarity between each pair in the extracted columns, and then calculate the mean of the cosine similarity;

[0063] The importance is determined according to the positive ratio of the mean; the relationship between the importance and the mean is a linear function with a positive coefficient.

[0064] Compared with the existing technology, the beneficial effects of the present invention are: the present invention obtains historical environmental data and tobacco images of tobacco samples, clusters the tobacco images, and then clusters the historical environmental data, compares similar historical environmental data, determines the importance of each parameter, and provides a reference for the resource allocation process. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention.

[0066] Figure 1 Flow chart of the method for optimizing the configuration of tobacco growth period to improve quality.

[0067] Figure 2 Block diagram of a tobacco growing period optimization configuration system for improving quality. DETAILED DESCRIPTION

[0068] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0069] Figure 1A flowchart of a method for optimizing tobacco growth period configuration to improve quality is provided. In an embodiment of the present invention, a method for optimizing tobacco growth period configuration to improve quality comprises:

[0070] Step S100: receiving a target growth period and obtaining a tobacco sample within the target growth period;

[0071] The tobacco growth period includes the sowing period (seed germination period), seedling period, growth period, flower bud differentiation period, flowering period, maturity period and harvest period. Each growth period needs to be analyzed separately. The technical solution of the present invention only analyzes tobacco within the same growth period. Accordingly, for the four growth periods, the same steps can be applied four times.

[0072] Specifically, the execution body of this method first receives the growth period selected by the user, which is called the target growth period. The target growth period is one of the sowing period (seed germination period), seedling period, growth period, flower bud differentiation period, flowering period, maturity period and harvest period. After receiving the target growth period, tobacco samples within the target growth period are selected. The selection process is very simple, because tobacco is planted in different regions and stages during planting. One plot corresponds to one growth period. The tobacco planted in the plot itself is registered data and can be read directly.

[0073] Step S200: obtaining meteorological data containing time tags of tobacco samples during their growth process using sensors installed in the planting environment to obtain a meteorological data matrix;

[0074] In existing tobacco growing environments, environmental detectors are pre-installed, that is, collection sensors, which acquire meteorological data and obtain a meteorological data matrix. Each column of the meteorological data matrix corresponds to a type of meteorological data, and each row corresponds to meteorological data at a collection moment. It should be noted that the collection cycles of all sensors must be the same, so that different types of data collected can be used as data in the same row. In addition, in actual scenarios, although the collection cycles are the same, the time tags of the acquired data will still have some discrepancies. For example, if the collection cycle is half an hour, the acquired data may be affected by the transmission process, and there will be errors of several seconds. These errors need to be ignored. Due to the difference in magnitude between the errors and cycles, it is not difficult to identify them.

[0075] Step S300: obtaining a tobacco image containing a time tag of a tobacco sample;

[0076] Each planted plant is a tobacco sample. With the help of a camera installed in the planting environment, tobacco images of the tobacco samples can be obtained. When obtaining the tobacco image, the acquisition time needs to be recorded, which is called a time tag.

[0077] Step S400: registering tobacco images based on time tags, and performing comparison and clustering on the registered tobacco images to obtain tobacco images of different classes;

[0078] For each tobacco sample, a batch of tobacco images with time labels can be obtained, indicating the growth status of the tobacco sample at different times. All tobacco images at the same time are counted and analyzed, and the analysis results obtained are also the analysis results of that time. Its practical significance lies in that, for example, if the time label is half of the development cycle, all tobacco images within the time label are obtained. All the tobacco images obtained represent the status of all tobacco samples after half of the development cycle. The statistically obtained tobacco images are compared and then clustered, so that tobacco samples with similar status after half of the development cycle are classified into one category.

[0079] Step S500: For tobacco images of the same type, extract the corresponding meteorological data matrix according to the time tag, compare the meteorological data matrix, and determine the importance of each parameter; the importance represents the degree of influence of each parameter in the process of tobacco growing into the growth state corresponding to the tobacco image of the same type;

[0080] For similar tobacco images, clustering is performed based on time tags, indicating that after the same growth time, similar growth states are obtained. This is the output in the mapping relationship. Furthermore, the independent variable of the growth state is meteorological data, and the meteorological data matrix before the time tag is intercepted as the input in the mapping relationship. The importance of each parameter in the intercepted meteorological data matrix can be evaluated. The importance indicates the degree of influence of each parameter in the process of tobacco growing into the growth state corresponding to the tobacco image of this type.

[0081] Step S600: upon receiving a target image input by a user, determining the class to which the image corresponds based on the same image comparison scheme, and querying the importance of each parameter corresponding to the tobacco image of that class;

[0082] The target image input by the user indicates what the user wants the tobacco to grow like. When the target image input by the user is received, the target image is classified based on the same scheme as step S400 to obtain the class corresponding to the target image. Then, the importance of each parameter corresponding to the tobacco image of this class is queried.

[0083] It should be noted that in the technical solution of the present invention, the target image input by the user may also contain a time tag (if not entered, the default is the final moment), and then the aligned tobacco image is first queried according to the time tag, and then the class is determined; it is worth mentioning that in the process of first querying the aligned tobacco image according to the time tag, the time range will generally be expanded. For example, when the time tag is half a growth cycle, one-tenth of the growth cycle can be floated based on the time tag.

[0084] Step S700: configuring environmental control resources based on the importance;

[0085] After obtaining the importance of each parameter, the limited environmental control resources can be configured based on the importance. The specific configuration method is very flexible and is not the focus of the technical solution of the present invention. It is mainly determined by the growers themselves. The actual function of the technical solution of the present invention is that the user inputs a target image, the executive body of the present invention determines the importance of each parameter, and feeds back to the user, and the user configures the resources according to the importance. The technical solution of the present invention will not generate a definite configuration plan. In fact, in the mature tobacco planting field, the best planting ratio has long been determined and is common knowledge. However, even under the best planting ratio conditions, the growth of different plants is still unpredictable. The "best planting ratio" is actually just a "better" seed value ratio. In the technical solution of the present invention, the importance of various parameters is determined only based on the most recent planting samples, thereby assisting users in resource configuration. It only assists users in planting in a general direction. The actual growth process still retains extremely high diversity, which can easily produce better planting effects.

[0086] It should be noted that the importance of various parameters is determined based on the most recent planting samples, which changes over time, which also ensures the timeliness of the importance assessment results.

[0087] Regarding step S200, the step of obtaining meteorological data containing time tags of tobacco samples during the growth process according to sensors installed in the planting environment to obtain a meteorological data matrix includes:

[0088] Acquiring meteorological data containing time tags of tobacco samples during their growth process based on sensors installed in the planting environment; the meteorological data includes at least air temperature, light intensity, soil temperature, and soil moisture;

[0089] Arrange various meteorological data with the same time tag in order according to the preset indicators to obtain row data;

[0090] Arrange the row data in chronological order according to the time labels to obtain the meteorological data matrix.

[0091] The process of obtaining meteorological data is very simple. The meteorological data with time tags of tobacco samples during the growth process will be obtained based on the sensors installed in the planting environment. Rows are created in chronological order, and columns are created according to the data type of the meteorological data to obtain the meteorological data matrix.

[0092] Regarding step S400, the steps of registering tobacco images based on time tags, comparing and clustering the registered tobacco images, and obtaining tobacco images of different classes include:

[0093] Register tobacco images based on time tags;

[0094] For tobacco images containing the same time label, feature points are marked in the tobacco images;

[0095] Compare the labeling results of different tobacco images and calculate the image distance;

[0096] The tobacco images are clustered according to the image distance to obtain tobacco images of different categories.

[0097] The above content explains the clustering process. First, tobacco images at the same moment (with sufficiently close time tags) are obtained. This process is called registration. Then, the features of the tobacco images (the positions of the marked feature points) are extracted. The difference in features is used as the image distance to reflect the difference in the images. Finally, after the image distance is determined, the tobacco images can be clustered using a distance-based clustering algorithm.

[0098] The above process is actually two classifications. The first classification is to register tobacco images based on time tags, and the second classification is to cluster images based on image distances.

[0099] Specifically, for tobacco images containing the same time tag, the step of marking feature points in the tobacco images includes:

[0100] Extract the layers of the tobacco image under different channels, traverse the pixels in each layer, record the pixel values, and construct an array; the traversal order of the pixels in all layers is the same;

[0101] For each array, calculate the independent value of each value in sequence; the independent value is used to represent the degree of difference between each value and other values ​​in the array;

[0102] Accumulate independent values. When the independent value reaches a preset value, query the position corresponding to the value and mark it as a feature point.

[0103] The calculation process of the independent value is: Where, is an independent value, For the preset parameters, is the larger value of the distance between the current element and the two ends of the array, The distance from the current element in the array is The difference between the value of the element that is being retrieved and the value of the current element.

[0104] The above calculation process is explained as follows:

[0105] According to the preset order, the pixels in the layer are converted into an array. For example, the value of each pixel is read from left to right and from top to bottom. This process is a dimensionality reduction process, and the information in the layer will be converted into an array. For each array, the difference between each value and other values ​​is calculated in turn. If the difference is small, it can be considered that it forms a region with the surrounding pixels. At this time, the corresponding position is marked as a feature point. It is essentially a region recognition scheme used to extract tobacco areas in the image.

[0106] In addition, regarding the calculation process of the independent value, for each element, the difference between its value and the values ​​of other elements is calculated and then accumulated. Taking into account the distance factor, this application introduces a weight coefficient so that the elements closer to the current element have a greater influence on the current element. Generally speaking, the larger the difference, the greater the difference between the current element and the other elements, the higher the independence and the larger the independent value. Therefore, in the above calculation process, the independent value is inversely proportional to the sum of the differences.

[0107] It is worth mentioning that the above summation process involves all elements in the array. In fact, the staff can set a range, such as 20% of the array elements, to determine the independent value of a certain element relative to a limited number of surrounding elements.

[0108] Regarding step S600, the steps of intercepting the corresponding meteorological data matrix according to the time tag for the tobacco images of the same type, comparing the meteorological data matrix, and determining the importance of each parameter include:

[0109] For any tobacco image of the same type, locate the row data in the meteorological data matrix according to the time label;

[0110] Get all row data before the located row data as the intercepted meteorological data matrix;

[0111] Compare the intercepted meteorological data matrices to determine the importance of each parameter.

[0112] For tobacco images of the same type, each tobacco image corresponds to a meteorological data matrix. The meteorological data matrix is ​​intercepted according to the time label of the tobacco image. At this time, the intercepted meteorological data matrix is ​​the independent variable and the tobacco image is the dependent variable. Finally, a meteorological data matrix set can be obtained for tobacco images of the same type. By comparing the various meteorological data matrices in the meteorological data matrix set, the importance of each parameter can be determined.

[0113] The calculation process of the importance of each parameter is as follows:

[0114] For any parameter, extract its corresponding column in each intercepted meteorological data matrix;

[0115] Calculate the cosine similarity between each pair in the extracted columns, and then calculate the mean of the cosine similarity;

[0116] The importance is determined based on the direct ratio of the mean.

[0117] The calculation principle of importance is explained as follows:

[0118] For any parameter, it corresponds to a column of data in each intercepted meteorological data matrix. All columns of data are extracted to obtain multiple arrays. For arrays, cosine similarity is calculated between each other. The more similar, the closer the cosine similarity is to one. If the cosine similarities are relatively large (the mean is large), it means that the corresponding values ​​of the parameter in each matrix are similar. Accordingly, it can be used as the main factor to determine the importance according to its proportionality; on the contrary, if the mean of the cosine similarity indicates that the same parameter has different corresponding values ​​in different matrices, but their growth states are still similar, it means that the influence of the parameter is not very important and the importance is low.

[0119] Furthermore, regarding the relationship between importance and mean, a linear function with a positive coefficient can achieve a proportional relationship.

[0120] As a preferred embodiment of the technical solution of the present invention, the method further includes:

[0121] Receive the tobacco to be optimized selected by the user, record the meteorological data in real time, and obtain the real-time meteorological matrix;

[0122] Obtain the meteorological data matrix with the time label as the end moment as the reference matrix;

[0123] The real-time meteorological matrix is ​​compared with the reference matrix in real time, the similarity is calculated, and the reference matrix whose similarity reaches the preset threshold is selected to obtain its corresponding tobacco image as the predicted image.

[0124] In an example of the technical solution of the present invention, a prediction function is provided. The existing function of this application is that no matter what stage of the target image the user uploads, the importance of each parameter can be obtained, thereby assisting in resource allocation; on this basis, the user can also select a tobacco, called the tobacco to be optimized, and obtain a real-time meteorological matrix based on the meteorological data recorded by the sensor (the number of columns continues to increase). The real-time meteorological matrix is ​​compared with the meteorological data matrix of the existing tobacco sample at the final moment (the number of columns reaches the maximum), and a sufficiently similar meteorological data matrix can be matched, which is called a reference matrix; the function of this process is to query tobacco samples with similar growth conditions to the tobacco to be optimized, obtain their tobacco images, and use them as prediction images, thereby providing a prediction function. The prediction image can also be used as a reference for resource allocation.

[0125] It is worth mentioning that during the prediction process, the more columns the real-time meteorological matrix has, the more difficult it is to match a sufficiently similar meteorological data matrix, and the fewer the number of predicted images. Therefore, during the prediction process, the number of predicted images decreases as the growth time goes by.

[0126] Figure 2 The structural block diagram of a tobacco growth period optimization configuration system for improving quality is provided. In an embodiment of the present invention, a tobacco growth period optimization configuration system for improving quality is provided. The system 10 includes:

[0127] The tobacco sample acquisition module 11 is used to receive a target growth period and acquire tobacco samples within the target growth period;

[0128] The meteorological data statistics module 12 is used to obtain meteorological data with time tags of tobacco samples during the growth process based on sensors installed in the planting environment to obtain a meteorological data matrix;

[0129] The tobacco image acquisition module 13 is used to acquire tobacco images containing time tags of tobacco samples;

[0130] The tobacco image clustering module 14 is used to register tobacco images based on time tags, compare and cluster the registered tobacco images, and obtain tobacco images of different classes;

[0131] The parameter analysis module 15 is used to intercept the corresponding meteorological data matrix according to the time tag for the tobacco image of the same type, compare the meteorological data matrix, and determine the importance of each parameter; the importance indicates the degree of influence of each parameter in the process of tobacco growing into the growth state corresponding to the tobacco image of the same type;

[0132] The importance query module 16 is used to determine the class corresponding to the image based on the same image comparison scheme when receiving the target image input by the user, and query the importance of each parameter corresponding to the tobacco image of this class;

[0133] The resource configuration module 17 is configured to configure the environment control resources based on the importance.

[0134] Furthermore, the meteorological data statistics module 12 includes:

[0135] A sensor monitoring unit, configured to obtain time-tagged meteorological data of tobacco samples during their growth process based on sensors installed in the growing environment; the meteorological data includes at least air temperature, light intensity, soil temperature, and soil moisture;

[0136] A row data generating unit is used to arrange various meteorological data with the same time tag in a sequence according to a preset index to obtain row data;

[0137] The column data generating unit is used to arrange the row data according to the time sequence of the time tags to obtain the meteorological data matrix.

[0138] Specifically, the tobacco image clustering module 14 includes:

[0139] an image registration unit, for registering tobacco images based on time tags;

[0140] A feature extraction unit, for marking feature points in tobacco images containing the same time tag;

[0141] A distance calculation unit, used to compare the labeling results of different tobacco images and calculate the image distance;

[0142] The clustering execution unit is used to cluster the tobacco images according to the image distance to obtain tobacco images of different categories.

[0143] Furthermore, the parameter analysis module 15 includes:

[0144] a row data locating unit, for locating row data of any tobacco image of the same type in the meteorological data matrix according to a time tag;

[0145] A data interception unit is used to obtain all row data before the located row data as the intercepted meteorological data matrix;

[0146] Matrix comparison unit, used to compare each intercepted meteorological data matrix and determine the importance of each parameter;

[0147] The calculation process of the importance of each parameter is as follows:

[0148] For any parameter, extract its corresponding column in each intercepted meteorological data matrix;

[0149] Calculate the cosine similarity between each pair in the extracted columns, and then calculate the mean of the cosine similarity;

[0150] The importance is determined according to the positive ratio of the mean; the relationship between the importance and the mean is a linear function with a positive coefficient.

[0151] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for optimizing tobacco growth period configuration to improve quality, characterized in that: The method comprises: Receiving the target growth period, obtaining tobacco samples within the target growth period; The sensors installed in the growing environment acquire meteorological data with time tags during the growth process of tobacco samples to obtain a meteorological data matrix. The acquisition cycle of all sensors is the same. Acquire tobacco images containing time tags of tobacco samples; Tobacco images are registered based on time tags, and the registered tobacco images are compared and clustered to obtain tobacco images of different classes; For tobacco images of the same type, the corresponding meteorological data matrix is ​​intercepted according to the time tag, and the meteorological data matrix is ​​compared to determine the importance of each parameter; the importance represents the degree of influence of each parameter in the process of tobacco growing into the growth state corresponding to the tobacco image of the same type; The steps to determine the importance of each parameter include: For any tobacco image of the same type, locate the row data in the meteorological data matrix according to the time label; Get all row data before the located row data as the intercepted meteorological data matrix; Compare each intercepted meteorological data matrix to determine the importance of each parameter; The calculation process of the importance of each parameter is as follows: For any parameter, extract its corresponding column in each intercepted meteorological data matrix; Calculate the cosine similarity between each pair in the extracted columns, and then calculate the mean of the cosine similarity; Determining the importance according to the positive ratio of the mean; the relationship between the importance and the mean is a linear function with a positive coefficient; When receiving the target image input by the user, the class corresponding to the image is determined based on the same image comparison scheme, and the importance of each parameter corresponding to the tobacco image of this class is queried; Environmental control resources are configured based on the importance.

2. The method for optimizing tobacco growth period configuration for improving quality according to claim 1, characterized in that: The step of obtaining meteorological data containing time tags of tobacco samples during the growth process according to sensors installed in the planting environment to obtain a meteorological data matrix includes: Acquiring meteorological data containing time tags of tobacco samples during their growth process based on sensors installed in the planting environment; the meteorological data includes at least air temperature, light intensity, soil temperature, and soil moisture; Arrange various meteorological data with the same time tag in order according to the preset indicators to obtain row data; Arrange the row data in chronological order according to the time labels to obtain the meteorological data matrix.

3. The method for optimizing tobacco growth period configuration for improving quality according to claim 1, characterized in that: The steps of registering tobacco images based on time tags, comparing and clustering the registered tobacco images, and obtaining tobacco images of different classes include: Register tobacco images based on time tags; For tobacco images containing the same time label, feature points are marked in the tobacco images; Compare the labeling results of different tobacco images and calculate the image distance; The tobacco images are clustered according to the image distance to obtain tobacco images of different categories.

4. The method for optimizing tobacco growth period configuration for improving quality according to claim 3, characterized in that: For tobacco images containing the same time tag, the step of marking feature points in the tobacco images includes: Extract the layers of the tobacco image under different channels, traverse the pixels in each layer, record the pixel values, and construct an array; the traversal order of the pixels in all layers is the same; For each array, calculate the independent value of each value in sequence; the independent value is used to represent the degree of difference between each value and other values ​​in the array; Accumulate independent values. When the independent value reaches a preset value, query the position corresponding to the value and mark it as a feature point. The calculation process of independent value is: Where is an independent value, For the preset parameters, The maximum value of the distance between the current element and the two ends of the array The distance from the current element in the array is The difference between the value of the element that is being retrieved and the value of the current element.

5. The method for optimizing tobacco growth period configuration for improving quality according to claim 1, characterized in that: The method further comprises: Receive the tobacco to be optimized selected by the user, record the meteorological data in real time, and obtain the real-time meteorological matrix; Obtain the meteorological data matrix with the time label as the end moment as the reference matrix; The real-time meteorological matrix is ​​compared with the reference matrix in real time, the similarity is calculated, and the reference matrix whose similarity reaches the preset threshold is selected to obtain its corresponding tobacco image as the predicted image.

6. A tobacco growth period optimization configuration system for improving quality, characterized in that: The system comprises: A tobacco sample acquisition module is used to receive a target growth period and obtain tobacco samples within the target growth period; The meteorological data statistics module is used to obtain meteorological data with time tags of tobacco samples during the growth process based on sensors installed in the planting environment to obtain a meteorological data matrix; the collection cycle of all sensors is the same; A tobacco image acquisition module is used to acquire tobacco images containing time tags of tobacco samples; The tobacco image clustering module is used to register tobacco images based on time tags, compare and cluster the registered tobacco images, and obtain tobacco images of different classes; The parameter analysis module is used to intercept the corresponding meteorological data matrix according to the time label for the same tobacco image. Comparing the meteorological data matrix, determining the importance of each parameter; the importance represents the degree of influence of each parameter in the process of tobacco growing into the growth state corresponding to the tobacco image; The steps to determine the importance of each parameter include: For any tobacco image of the same type, locate the row data in the meteorological data matrix according to the time label; Get all row data before the located row data as the intercepted meteorological data matrix; Compare each intercepted meteorological data matrix to determine the importance of each parameter; The calculation process of the importance of each parameter is as follows: For any parameter, extract its corresponding column in each intercepted meteorological data matrix; Calculate the cosine similarity between each pair in the extracted columns, and then calculate the mean of the cosine similarity; Determining the importance according to the positive ratio of the mean; the relationship between the importance and the mean is a linear function with a positive coefficient; The importance query module is used to determine the class of the image based on the same image comparison scheme when receiving the target image input by the user, and query the importance of each parameter corresponding to the tobacco image of this class; A resource configuration module is used to configure environmental control resources based on the importance.

7. The tobacco growth period optimization configuration system for improving quality according to claim 6, characterized in that: The meteorological data statistics module includes: A sensor monitoring unit, configured to obtain time-tagged meteorological data of tobacco samples during their growth process based on sensors installed in the growing environment; the meteorological data includes at least air temperature, light intensity, soil temperature, and soil moisture; A row data generating unit is used to arrange various meteorological data with the same time tag in a sequence according to a preset index to obtain row data; The column data generating unit is used to arrange the row data according to the time sequence of the time tags to obtain the meteorological data matrix.

8. The tobacco growth period optimization configuration system for improving quality according to claim 6, characterized in that: The tobacco image clustering module includes: an image registration unit, for registering tobacco images based on time tags; A feature extraction unit, for marking feature points in tobacco images containing the same time tag; A distance calculation unit, used to compare the labeling results of different tobacco images and calculate the image distance; The clustering execution unit is used to cluster the tobacco images according to the image distance to obtain tobacco images of different categories.

9. The tobacco growth period optimization configuration system for improving quality according to claim 6, characterized in that: The parameter analysis module includes: a row data locating unit, for locating row data of any tobacco image of the same type in the meteorological data matrix according to a time tag; A data interception unit is used to obtain all row data before the located row data as the intercepted meteorological data matrix; Matrix comparison unit, used to compare each intercepted meteorological data matrix and determine the importance of each parameter; The calculation process of the importance of each parameter is as follows: For any parameter, extract its corresponding column in each intercepted meteorological data matrix; Calculate the cosine similarity between each pair in the extracted columns, and then calculate the mean of the cosine similarity; The importance is determined according to the positive ratio of the mean; the relationship between the importance and the mean is a linear function with a positive coefficient.

Images