A carbon fixation plant combination design method and system based on reinforcement learning

Abstract

The present invention discloses a carbon fixation plant combination design method and system based on reinforcement learning, which relates to the field of data processing technology and is used to solve the problem that when sudden rainfall occurs in a garden, plant combinations at certain positions will absorb water to accelerate carbon fixation; the root systems of plant combinations at certain positions may be washed away, causing trees to fall down, thereby losing the original carbon fixation effect. The method comprises the following steps: selecting a plant combination with a well-developed root system or a plant combination with a high efficiency in carbon fixation according to the carbon fixation demand of each position in the garden, collecting planting feedback data of each position in the garden and emergency environment data of the corresponding position; judging whether the plant combination at the current position needs to be adjusted based on the planting feedback data of each position in the garden and the emergency environment data of the corresponding position; analyzing the risk degree of each position according to the planting feedback data of each position and the emergency environment data of each position by collecting the planting feedback data of each position and the emergency environment data of each position, and further judging whether the plant combination at each position needs to be replaced.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and more specifically, to a carbon fixation plant combination design method and system based on reinforcement learning. Background Art

[0002] Reinforcement learning is a machine learning method that aims to learn how to make decisions through interaction with the environment to obtain the maximum cumulative reward. The application of reinforcement learning technology in combination with carbon-fixing plants can screen and update carbon-fixing plants in different environments to improve the carbon fixation efficiency in different environments and thus improve the overall carbon fixation efficiency.

[0003] The prior art has the following deficiencies:

[0004] Traditional garden ecological design mostly only considers the beauty and greening effect of tree species matching, ignoring the connection between the various locations in the garden and the characteristics of the plant combination in some unexpected situations. When there is sudden rainfall in the garden, the plant combination in some locations will absorb water to accelerate growth and carbon fixation; the root system of the plant combination in some locations may be washed away, causing the trees to fall, thus losing the original carbon fixation and landscape effects.

[0005] In view of the above problems, the present invention proposes a solution. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a carbon fixation plant combination design method and system based on reinforcement learning to solve the problems raised in the above-mentioned background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A carbon fixation plant combination design method based on reinforcement learning comprises the following steps:

[0009] Step S1, collecting tree species data and garden data;

[0010] Step S2, classifying the plant combination into a plant combination with a well-developed root system and a plant combination with high carbon fixation efficiency according to the tree species data, dividing the garden according to the garden data and analyzing the carbon fixation demand of each position in the garden;

[0011] Step S3, selecting a combination of plants with well-developed root systems or a combination of plants with high carbon fixation efficiency according to the carbon fixation demand of each location in the garden, and collecting planting feedback data of each location in the garden and emergency environmental data of the corresponding location;

[0012] Step S4, comprehensively considering the planting feedback data of each location in the garden and the emergency environment data of the corresponding location to determine whether the plant combination at the current location needs to be adjusted.

[0013] In a preferred embodiment, in step S1, the tree species data includes the tree species life cycle and the tree species gibberellin content; the garden data includes the garden terrain data and the garden carbon dioxide content.

[0014] In a preferred embodiment, in step S2, a common carbon-fixing garden plant combination with high carbon-fixing efficiency is obtained by reviewing the historical garden planting records, and the life cycle of the common carbon-fixing garden plant combination is obtained by accessing the historical garden planting records; the average life cycle of several garden tree species is calculated as the life cycle of the carbon-fixing garden plant combination; and the life cycles of all common carbon-fixing garden plant combinations are merged into a life data set;

[0015] The gibberellin content of common carbon-fixing garden plant combinations was obtained by high performance liquid chromatography (HPLC). The average value of gibberellin of garden tree species in the common carbon-fixing garden plant combinations was taken as the gibberellin content of the common carbon-fixing garden plant combinations. The gibberellin content of all common carbon-fixing garden plant combinations was merged into a gibberellin content data set.

[0016] In a preferred embodiment, in step S2, the common carbon fixation garden plant combinations are classified by using reinforcement learning technology, and the specific steps are as follows:

[0017] Calculate the average value of the life data set; randomly select a data in the life data set, and compare the size of the selected data with the average value of the life data set; set the bonus score to be initialized to 0. If the selected data is greater than the average value of the life data set, increase the bonus score by one and record the selected data and put it back into the life data set; if the selected data is less than the average value of the life data set, reduce the bonus score by one and record the selected data and put it back into the data set; if the selected data is equal to the average value of the life data set, put it back directly; repeat the operation N times to obtain N cumulative scores and arrange the cumulative scores in ascending order, and randomly select the selected data with several cumulative score records in the middle as the analysis sample; take the average value in the analysis sample as the life cycle threshold;

[0018] Obtain the gibberellin content threshold. If the life cycle of the current carbon-fixing plant combination is greater than or equal to the life cycle threshold and the gibberellin content of the current carbon-fixing plant combination is greater than or equal to the gibberellin content threshold, then the current carbon-fixing plant combination is judged to be a carbon-fixing plant combination with a well-developed root system, otherwise it is a carbon-fixing plant combination with an efficient carbon fixation;

[0019] Obtaining terrain data of each location in the garden, the terrain data is the altitude of each location in the garden; after obtaining the altitude of each location in the garden, normalization processing can be performed; dividing the altitude of each location in the garden by the highest altitude of the garden to obtain the altitude coefficient of each location in the garden, which is recorded as a1, a2, a3, and a4;

[0020] Then, the carbon dioxide content at the corresponding position is obtained through an infrared carbon dioxide sensor; a certain time period at night can be selected for detection, and the same detection time interval can be selected to use the infrared carbon dioxide sensor to perform real-time detection at various positions in the garden, and the detection results are recorded, and the carbon dioxide content at various positions in the garden is normalized, and the normalized value of the carbon dioxide concentration at various positions in the garden is used as the carbon dioxide content coefficient at various positions in the garden and recorded as b1, b2, b3, and b4;

[0021] The carbon fixation demand coefficient of each location in the garden is calculated by the altitude coefficient of each location in the garden and the carbon dioxide content coefficient of each location in the garden. The carbon fixation demand of each location in the garden is analyzed according to the carbon fixation demand coefficient of each location in the garden. The formula is X i =a i -b i , where X i is the carbon fixation demand coefficient of each location in the garden, and i is the serial number of each location in the garden;

[0022] By setting weights for the carbon sequestration coefficients at each location in the garden, the carbon sequestration demand of the entire garden is divided into regions, the carbon sequestration coefficients at each location in the garden are sorted from small to large, and the weights are set using the hierarchical analysis method to determine the proportion of carbon sequestration demand at each location in the garden;

[0023] After obtaining the carbon sequestration demand weights of each location in the garden, the regional carbon sequestration threshold is set for the total carbon sequestration in the garden according to the carbon sequestration demand weights of each location in the garden; the regional carbon sequestration threshold is obtained by multiplying the total carbon sequestration in the garden by the carbon sequestration demand weights of the corresponding locations in the garden.

[0024] In a preferred embodiment, in step S3, the scientific research report in the garden protection website is accessed to obtain the total carbon sequestration data of the garden and the carbon sequestration data of each position in the garden within a fixed time period; by calculating the carbon sequestration demand weight of each position in the garden, the total carbon sequestration of the garden is set according to the carbon sequestration demand weight of each position in the garden to set the regional carbon sequestration threshold; then the carbon sequestration data of each position in the garden is compared with the regional carbon sequestration threshold;

[0025] If the carbon fixation data of a certain location in the garden is greater than or equal to the carbon fixation threshold of the corresponding area, a combination of carbon fixation plants with well-developed root systems will be selected; if the carbon fixation data of a certain location in the garden is less than the carbon fixation threshold of the corresponding area, a combination of carbon fixation plants with efficient carbon fixation will be selected;

[0026] After selecting a good carbon-fixing plant combination for planting, collect planting feedback data at each location in the garden and emergency environmental data at each location in the garden; the planting feedback data at each location in the garden include changes in oxygen concentration at each location in the garden and changes in soil organic carbon content at each location in the garden;

[0027] After selecting the carbon fixation plant combination for planting at each location in the garden, the oxygen concentration at each location in the garden and the soil organic carbon content at each location in the garden were detected in the same time period as before the carbon fixation plant combination was planted; the average oxygen concentration at each location in the garden before and after the carbon fixation plant combination was planted and the average soil organic carbon content at each location in the garden were calculated and recorded as in, and The average oxygen concentration at each location in the garden before planting the carbon-fixing plant combination and the average soil organic carbon content at each location in the garden; and The average oxygen concentration at each location in the garden and the average soil organic carbon content at each location in the garden after planting the carbon fixation plant combination; the coefficient of change of oxygen concentration at each location in the garden and the coefficient of change of soil organic carbon content at each location in the garden are calculated, and the formula is: Among them, i is the coefficient of variation of oxygen concentration at each location in the garden, H i is the coefficient of variation of soil organic carbon content at each location in the garden;

[0028] After planting the carbon-fixing plant combination at various locations in the garden, the number of plants at the corresponding location is recorded. When an emergency occurs at a certain location, the number of plants is detected by lidar. The survival rate of the plant combination at the corresponding location is obtained by dividing the number of plants detected by the lidar by the number of plants at the corresponding location, which is recorded as Z.

[0029] In a preferred embodiment, in step S4, the coefficient of variation of oxygen concentration at each position in the garden is calculated. i , the coefficient of variation of soil organic carbon content at each location in the garden H i and the survival rate of plant combinations is denoted as Z, which is defined as the input variable and divided into different fuzzy sets;

[0030] Define the risk level as the output variable and divide it into fuzzy sets;

[0031] Formulate a set of fuzzy rules to describe the impact of different input variables on the output variables;

[0032] Perform fuzzy reasoning based on fuzzy rules to determine the risk level of the corresponding location.

[0033] A carbon fixation plant combination design system based on reinforcement learning, used to implement the above-mentioned carbon fixation plant combination design method based on reinforcement learning, characterized in that: it includes a data acquisition module, a data analysis module and a data storage module, and the modules are signal connected;

[0034] The data acquisition module is used to acquire tree species data and garden data and send them to the data analysis module;

[0035] The data analysis module determines the carbon sequestration plant combination and its risk level at each location based on tree species data and garden data;

[0036] The data storage module is used to store all data in the data processing process of the carbon fixation plant combination invention system based on reinforcement learning.

[0037] Technical effects and advantages of the carbon fixation plant combination design method and system based on reinforcement learning of the present invention:

[0038] The present invention collects garden data and tree species data, classifies each garden location, collects classified garden environment data for each garden location, and classifies plant combinations according to the tree species data. The garden environment data for each garden location and the tree species data are integrated to analyze the degree of adaptation, select plant combinations according to the degree of adaptation, collect planting feedback data for each location and emergency environment data for each location, analyze the risk level of each location according to the planting feedback data for each location and the emergency environment data for each location, and judge whether the plant combination of each location needs to be replaced according to the risk level of each location. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A first logic schematic diagram of a carbon fixation plant combination design system based on reinforcement learning according to the present invention;

[0040] Figure 2 A second logical schematic diagram of a carbon fixation plant combination design system based on reinforcement learning according to the present invention;

[0041] Figure 3 The present invention is a schematic flow chart of a carbon fixation plant combination design method based on reinforcement learning. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] The present invention collects garden data and tree species data, classifies each garden location, collects classified garden environment data for each garden location, and classifies plant combinations according to the tree species data. The garden environment data for each garden location and the tree species data are integrated to analyze the degree of adaptation, select plant combinations according to the degree of adaptation, collect planting feedback data for each location and emergency environment data for each location, analyze the risk level of each location according to the planting feedback data for each location and the emergency environment data for each location, and judge whether the plant combination of each location needs to be replaced according to the risk level of each location.

[0044] Example 1, the present invention discloses a method for designing a carbon fixation plant combination based on reinforcement learning, such as Figure 1-Figure 3 As shown, the following steps are included:

[0045] Step S1, collecting tree species data and garden data.

[0046] Step S2, classifying the plant combinations into plant combinations with well-developed root systems and plant combinations with efficient carbon fixation according to the tree species data, dividing the garden according to the garden data and analyzing the carbon fixation requirements of each location in the garden.

[0047] Step S3, selecting a combination of plants with well-developed root systems or a combination of plants with high carbon fixation efficiency according to the carbon fixation demand of each location in the garden, and collecting planting feedback data of each location in the garden and emergency environmental data of the corresponding location.

[0048] Step S4, comprehensively considering the planting feedback data of each location in the garden and the emergency environment data of the corresponding location to determine whether the plant combination at the current location needs to be adjusted.

[0049] The specific implementation is as follows:

[0050] In step S1, the tree species data includes the tree species life cycle and the tree species gibberellin content. The longer a tree species survives, the better its root system develops. Gibberellin is a hormone that promotes the development of plant roots. The higher the gibberellin content in a tree species, the better its root system develops. The garden tree species life cycle can be obtained by accessing the garden's historical planting records. Receive the tree species gibberellin data detected by a high performance liquid chromatography instrument.

[0051] Garden data includes garden topography data and garden carbon dioxide content. Visit the garden protection website to obtain garden topography data and garden carbon sequestration data at various locations. The garden carbon dioxide content can be obtained through an infrared carbon dioxide sensor. Gardens have high terrain, rapid water loss, poor carbon sequestration capacity, and high carbon sequestration requirements. Carbon dioxide is the main raw material for carbon sequestration, and the carbon dioxide content is low, the carbon sequestration capacity is poor, and the carbon sequestration demand is high.

[0052] It should be noted that historical garden planting records refer to records of past vegetation restoration, ecological protection or agroforestry planting activities in garden areas. These records usually include planted tree species, planting time, location, survival rate, etc. The life cycle of garden tree species and common carbon-fixing garden plant combinations can be obtained by accessing historical garden planting records. High-performance liquid chromatography instruments are instruments that efficiently separate and quantify chemical substances. They can detect and measure the content of some hormones and display the content data on the receiving device. They can be used to obtain the gibberellin content of garden tree species. Garden protection websites usually provide information and services on garden ecological environment, gardening prevention and control, vegetation restoration, etc. Garden terrain data and carbon fixation data at various locations in the garden can be obtained through the garden protection website. Infrared carbon dioxide sensor is a common sensor used to detect carbon dioxide concentration in the air. It uses infrared technology to detect carbon dioxide concentration.

[0053] In step S2, several common carbon-fixing garden plant combinations with high carbon fixation efficiency can be obtained by reviewing the historical garden planting records, such as garden poplars and garden willows, cacti and agaves, as well as acacias and locusts. The life cycle of common carbon-fixing garden plant combinations is obtained by accessing the historical garden planting records. Specifically, if there are two garden tree species in the common carbon-fixing garden plant combination, the average life cycle of the two garden tree species is calculated as the life cycle of the carbon-fixing garden plant combination. The life cycles of all common carbon-fixing garden plant combinations are merged into a life data set.

[0054] The gibberellin content of common carbon-fixing garden plant combinations was obtained by high performance liquid chromatography. Similarly, the average value of gibberellin of garden tree species in common carbon-fixing garden plant combinations was taken as the gibberellin content of common carbon-fixing garden plant combinations. The gibberellin content of all common carbon-fixing garden plant combinations was merged into a gibberellin content data set.

[0055] Common carbon-fixing garden plant combinations can be classified using reinforcement learning technology. The specific steps are as follows:

[0056] Calculate the average value of the life data set. Then randomly select a data in the life data set and compare the size of the selected data with the average value of the life data set. Set the bonus score to be initialized to 0. If the selected data is greater than the average value of the life data set, add one to the bonus score and record the selected data and put it back into the life data set; if the selected data is less than the average value of the life data set, subtract one from the bonus score and record the selected data and put it back into the data set; if the selected data is equal to the average value of the life data set, put it back directly. Set the number of selections according to the actual situation. You can select 10 data to compare with the average value of the life data set. The cumulative score is the value of the bonus score after the set number of selections, and the cumulative score is calculated. Repeat the operation N times to obtain N cumulative scores and arrange the cumulative scores in ascending order. Take the selected data with the cumulative scores in the middle as the analysis sample. Take the average value in the analysis sample as the life cycle threshold. For example, set N to 5 times, get 5 cumulative scores and arrange the cumulative scores in ascending order. Take the three cumulative scores in the middle as the analysis sample. If the number of selections is 10, there are 30 selected data in the analysis sample, and the average value is calculated as the life cycle threshold.

[0057] Similarly, the gibberellin content threshold is obtained. If the life cycle of the current carbon-fixing plant combination is greater than or equal to the life cycle threshold and the gibberellin content of the current carbon-fixing plant combination is greater than or equal to the gibberellin content threshold, then the current carbon-fixing plant combination is judged to be a carbon-fixing plant combination with a well-developed root system, otherwise it is a carbon-fixing plant combination with a high carbon fixation efficiency. It should be explained that the carbon-fixing plant combination with a well-developed root system needs to release more carbon dioxide during growth, and its carbon fixation effect is lower than that of the carbon-fixing plant combination with a high carbon fixation efficiency. However, because the carbon-fixing plant combination with a well-developed root system has stronger vitality, it is more adaptable to extreme environments than the carbon-fixing plant combination with a high carbon fixation efficiency.

[0058] According to historical experience, carbon sequestration in gardens is mainly carried out in the following areas: the edge of the garden, the center of the garden, the wetland of the garden, and the river valley of the garden. Different locations in the garden also have different carbon sequestration requirements.

[0059] The carbon fixation demand of a garden is related to the terrain conditions and carbon dioxide concentration. The higher the terrain at a certain location in the garden, the faster the water loss, the less water the carbon fixation plant combination needs to fix carbon, the lower the carbon fixation efficiency, and the higher the carbon fixation demand. Carbon dioxide is the main raw material for carbon fixation, and the lower its concentration, the lower the carbon fixation efficiency of the carbon fixation plant combination, and the higher the carbon fixation demand.

[0060] Visit the garden protection website to obtain the terrain data of each location in the garden. Further, the terrain data can be the altitude of each location in the garden. After obtaining the altitude of each location in the garden, normalization processing can be performed. Divide the altitude of each location in the garden by the highest altitude of the garden to obtain the altitude coefficient of each location in the garden, which is recorded as a1, a2, a3, and a4. Among them, the highest altitude of the garden can also be obtained by visiting the garden protection website.

[0061] Then use the infrared carbon dioxide sensor to obtain the carbon dioxide content at the corresponding location. You can choose a certain time period at night for detection. Since there is no sunlight at night, plants cannot perform photosynthesis. During this period, carbon dioxide changes less due to environmental factors, and the detection is more accurate. Select the same detection time interval and use the infrared carbon dioxide sensor to perform real-time detection at various locations in the garden, record the detection results, and normalize the carbon dioxide content at various locations in the garden. The normalization formula can be in, is the average carbon dioxide concentration within the detection time interval, C min is the minimum carbon dioxide concentration within the detection time interval, C max is the maximum carbon dioxide concentration within the detection time interval, C normal is the normalized value of the carbon dioxide concentration at the corresponding position. The normalized value of the carbon dioxide concentration at each position in the garden is used as the carbon dioxide content coefficient at each position in the garden and recorded as b1, b2, b3, and b4.

[0062] The carbon fixation demand coefficient of each location in the garden is calculated by the altitude coefficient of each location in the garden and the carbon dioxide content coefficient of each location in the garden. The carbon fixation demand of each location in the garden is analyzed according to the carbon fixation demand coefficient of each location in the garden. The formula can be X i =a i -b i , where X i is the carbon fixation demand coefficient of each location in the garden, and i is the serial number of each location in the garden.

[0063] By setting weights for the carbon sequestration coefficients at each location in the garden, the carbon sequestration demand of the entire garden is divided into regions, and the carbon sequestration demand coefficients at each location in the garden are sorted from small to large. The analytic hierarchy process can be used to set weights and determine the proportion of carbon sequestration demand at each location in the garden. The specific steps are as follows:

[0064] Construct a judgment matrix: First construct a judgment matrix to compare the relative importance of each position. Use a 4-point scale from 1 to 4, where 1 indicates equal importance, 2 indicates slight importance difference, 3 indicates moderate importance difference, and 4 indicates very important difference.

[0065]

[0066] Calculate the eigenvector: Calculate the eigenvector of each factor. The eigenvector is the weighted average of each column of the judgment matrix. First, normalize each column of the judgment matrix so that the sum of each column is equal to 1. Then for each column, calculate the weighted average as the element value of the eigenvector. For the above judgment matrix, the normalized matrix is ​​as follows:

[0067]

[0068] Then, calculate the weighted average for each factor:

[0069] A position importance = (0.1000 + 0.0770 + 0.0870 + 0.1200) / 4 = 0.0960;

[0070] B position importance = (0.2000 + 0.1538 + 0.1305 + 0.1600) / 4 = 0.1611;

[0071] C position importance = (0.3000 + 0.3077 + 0.2608 + 0.2400) / 4 = 0.2771;

[0072] D position importance = (0.4000 + 0.4615 + 0.5217 + 0.4800) / 4 = 0.4658.

[0073] Among them, areas A, B, C, and D correspond one by one to the four garden locations in this example. After sorting the carbon fixation demand coefficients of each garden location, they correspond to areas A, B, C, and D in ascending order.

[0074] After obtaining the carbon sequestration demand weights of each location in the garden, the regional carbon sequestration threshold is set for the total carbon sequestration in the garden according to the carbon sequestration demand weights of each location in the garden. The regional carbon sequestration threshold can be obtained by multiplying the total carbon sequestration in the garden by the carbon sequestration demand weights of the corresponding location in the garden. For example, if location A is a garden sand area, the regional carbon sequestration threshold of the garden sand dune location should be the total carbon sequestration in the garden multiplied by 0.0960.

[0075] It should be noted that reinforcement learning technology is a machine learning method that aims to achieve a certain goal through interactive learning with the environment. The carbon sequestration garden plant combination can be classified by setting a reward signal to obtain the maximum cumulative reward. The formulas involved in the above steps and the method of setting the regional carbon sequestration threshold are not unique.

[0076] In step S3, the scientific research report in the garden protection website is accessed to obtain the total carbon sequestration data of the garden and the carbon sequestration data of each location in the garden within a fixed time period. By calculating the carbon sequestration demand weight of each location in the garden, the total carbon sequestration of the garden is set according to the carbon sequestration demand weight of each location in the garden. Then, the carbon sequestration data of each location in the garden is compared with the regional carbon sequestration threshold.

[0077] If the carbon fixation data at a certain location in the garden is greater than or equal to the carbon fixation threshold of the corresponding area, it means that the carbon fixation amount at the corresponding location meets the carbon fixation requirements, and a carbon fixation plant combination with well-developed root systems is selected; if the carbon fixation data at a certain location in the garden is less than the carbon fixation threshold of the corresponding area, it means that the carbon fixation amount at the corresponding location does not meet the carbon fixation requirements, and a carbon fixation plant combination with high-efficiency carbon fixation is selected.

[0078] After selecting the carbon-fixing plant combination for planting, collect the planting feedback data and emergency environmental data of each location in the garden. The planting feedback data of each location in the garden include the changes in oxygen concentration at each location in the garden and the changes in soil organic carbon content at each location in the garden. The higher the oxygen concentration rises or the more the organic carbon content increases, the better the carbon fixation effect at the corresponding location. The emergency environmental data of the corresponding location is the survival rate of the plant combination at the corresponding location after an emergency occurs at a certain location in the garden. There are many emergencies, such as sandstorms, heavy rains, etc. The monitoring technology is existing technology and will not be repeated here.

[0079] Oxygen sensors can be used to detect oxygen concentrations at various locations in the garden, and the organic carbon content in the soil at various locations in the garden can be provided by data detected by a TOC analyzer. Before planting the carbon-fixing plant combination, a period of time is selected as the detection time to detect the oxygen concentration at various locations in the garden and the organic carbon content in the soil at various locations in the garden in real time. After selecting the carbon-fixing plant combination to be planted at various locations in the garden, the same time period as the detection time before planting the carbon-fixing plant combination is selected to detect the oxygen concentration at various locations in the garden and the organic carbon content in the soil at various locations in the garden. Calculate the average oxygen concentration at various locations in the garden before and after planting the carbon-fixing plant combination and the average organic carbon content in the soil at various locations in the garden, and record them as in, and The average oxygen concentration at each location in the garden before planting the carbon-fixing plant combination and the average soil organic carbon content at each location in the garden; and The average oxygen concentration at each location in the garden and the average soil organic carbon content at each location in the garden after planting the carbon fixation plant combination. The coefficient of change of oxygen concentration at each location in the garden and the coefficient of change of soil organic carbon content at each location in the garden can be calculated using the formula: Among them, i is the coefficient of variation of oxygen concentration at each location in the garden, H i is the coefficient of variation of soil organic carbon content at each location in the garden.

[0080] After planting carbon-fixing plant combinations at various locations in the garden, the number of plants at the corresponding location is recorded. When an emergency occurs at a certain location, the number of plants is detected using a lidar. The survival rate of the plant combination at the corresponding location is obtained by dividing the number of plants detected by the lidar by the number of plants at the corresponding location, which is recorded as Z.

[0081] It should be noted that scientific research reports are documents that record, summarize and share research results and findings. The total amount of carbon sequestered in the garden and the amount of carbon sequestered in the garden area within a fixed time period can be obtained from it. An oxygen sensor is a sensor used to detect oxygen concentration. Common types include electrochemical oxygen sensors and optical oxygen sensors, which can be used to detect oxygen concentrations at various locations in the garden. A TOC analyzer is an instrument used to measure the total organic carbon content in samples such as water bodies, wastewater, and soil. LiDAR is a remote sensing technology that measures the distance, position, and shape of a target object by emitting a laser beam. When calculating the survival rate of a plant community, LiDAR can provide high-precision terrain and vegetation structure information.

[0082] In step S4, the oxygen concentration variation coefficient O of each position in the garden is calculated. i , the coefficient of variation of soil organic carbon content at each location in the garden H i and the survival rate of plant combinations is denoted as Z, which is defined as the input variable and divided into different fuzzy sets.

[0083] It should be explained that the garden is divided into four locations in the example. i and H i They represent the oxygen concentration variation coefficient and soil organic carbon content variation coefficient of the four locations in the garden respectively. If an emergency occurs at a certain location in the garden, the oxygen concentration variation coefficient, soil organic carbon content variation coefficient and plant combination survival rate of the corresponding location will be used as input variables.

[0084] For example, “High”, “Low” for the coefficient of variation of oxygen concentration, “Big”, “Small” for the coefficient of variation of soil organic carbon content, and “Good”, “Bad” for the survival rate of plant assemblages.

[0085] Define the risk degree as the output variable and divide it into fuzzy sets, for example, “YES”, “NO” for the risk degree.

[0086] Formulate a set of fuzzy rules to describe the impact of different input variables on output variables. The definition of rules can be based on professional knowledge or obtained through data analysis and experiments. For example:

[0087] The coefficient of variation of oxygen concentration is marked as O, the coefficient of variation of soil organic carbon content is marked as H, the survival rate of plant combination is marked as Z, and the risk level is marked as F.

[0088] Rule 1:IF(O is High)AND(H is Big)AND(Z is Good)THEN(F is NO)

[0089] Rule 2:IF(O is Low)AND(H is Small)AND(Z is Bad)THEN(F is YES) ...

[0090] Fuzzy reasoning is performed according to fuzzy rules. When the output result is NO, the risk level of the corresponding location is judged to be low and there is no need to adjust the plant combination. When the output result is YES, the risk level of the corresponding location is judged to be high and the plant combination needs to be adjusted.

[0091] It should be noted that the division of fuzzy sets can be adjusted according to actual conditions. For example, this example takes two fuzzy sets as examples. In fact, the oxygen concentration variation coefficient, soil organic carbon content variation coefficient, plant combination survival rate and risk level can be divided into three or more sets to facilitate better precise adjustment according to different conditions.

[0092] In addition, for the oxygen concentration variation coefficient, the soil organic carbon content variation coefficient, and the plant combination survival rate, thresholds can be set according to actual conditions. For example, when the oxygen concentration variation coefficient exceeds 1%, it is marked as "High"; when the soil organic matter content variation coefficient exceeds 10%, it is marked as "Big"; when the plant combination survival rate exceeds 90%, it is marked as "Good", etc., which will not be elaborated here.

[0093] It should be noted that there is not only one way to adjust the plant combination. The present invention provides a method. If the original plant combination is a high-efficiency carbon fixation plant combination, and the survival rate is low after an emergency occurs, then a plant combination with a well-developed root system is planted instead; if the original plant combination is a plant combination with a well-developed root system, and the survival rate is high but the carbon fixation amount is low after an emergency occurs, then a high-efficiency carbon fixation plant combination is planted instead. The specific adjustment is implemented according to the fine adjustment of the fuzzy set by professionals. Furthermore, common high-efficiency carbon fixation plant combinations can consider planting poplars and willows. These woody plants have a high growth rate and biomass, and can absorb a large amount of carbon dioxide and convert it into organic carbon. Common root system well-developed plant combinations can consider planting cacti and agaves. The root systems of cacti and agaves are usually very developed, and most of the roots are in the surface layer, forming a dense root pile structure. No in-depth analysis is made here.

[0094] Example 2

[0095] The present invention also discloses a carbon fixation plant combination design system based on reinforcement learning, including a data acquisition module, a data analysis module and a data storage module, and the modules are signal connected;

[0096] The data acquisition module is used to acquire tree species data and garden data and send them to the data analysis module;

[0097] The data analysis module determines the carbon sequestration plant combination and its risk level at each location based on tree species data and garden data;

[0098] The data storage module is used to store all data in the data processing process of the carbon fixation plant combination invention system based on reinforcement learning.

[0099] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0100] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0101] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0102] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0103] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0104] Finally: 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, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A carbon fixation plant combination design method based on reinforcement learning, characterized in that: Step S1, collecting tree species data and garden data; Step S2, classifying the plant combination into a plant combination with a well-developed root system and a plant combination with high carbon fixation efficiency according to the tree species data, dividing the garden according to the garden data and analyzing the carbon fixation demand of each position in the garden; Step S3, selecting a combination of plants with well-developed root systems or a combination of plants with high carbon fixation efficiency according to the carbon fixation demand of each location in the garden, and collecting planting feedback data of each location in the garden and emergency environmental data of the corresponding location; Step S4, comprehensively analyzing the planting feedback data of each location in the garden and the emergency environment data of the corresponding location to determine whether the current location needs to adjust the plant combination; In step S2, a common carbon-fixing garden plant combination with high carbon-fixing efficiency is obtained by reviewing the historical garden planting records, and the life cycle of the common carbon-fixing garden plant combination is obtained by accessing the historical garden planting records; the average life cycle of several garden tree species is calculated as the life cycle of the carbon-fixing garden plant combination; the life cycles of all common carbon-fixing garden plant combinations are merged into a life data set; The gibberellin content of common carbon-fixing garden plant combinations was obtained by high performance liquid chromatography; the average value of the gibberellin of garden tree species in the common carbon-fixing garden plant combinations was taken as the gibberellin content of the common carbon-fixing garden plant combinations; the gibberellin content of all common carbon-fixing garden plant combinations was combined into a gibberellin content data set; In step S2, common carbon-fixing garden plant combinations are classified by using reinforcement learning technology, and the specific steps are as follows: Calculate the average value of the life data set; randomly select a data in the life data set, and compare the size of the selected data with the average value of the life data set; set the bonus score to be initialized to 0. If the selected data is greater than the average value of the life data set, increase the bonus score by one and record the selected data and put it back into the life data set; if the selected data is less than the average value of the life data set, reduce the bonus score by one and record the selected data and put it back into the data set; if the selected data is equal to the average value of the life data set, put it back directly; repeat the operation N times to obtain N cumulative scores and arrange the cumulative scores in ascending order, and randomly select the selected data with several cumulative score records in the middle as the analysis sample; take the average value in the analysis sample as the life cycle threshold; Obtain the gibberellin content threshold. If the life cycle of the current carbon-fixing plant combination is greater than or equal to the life cycle threshold and the gibberellin content of the current carbon-fixing plant combination is greater than or equal to the gibberellin content threshold, then the current carbon-fixing plant combination is judged to be a carbon-fixing plant combination with a well-developed root system, otherwise it is a carbon-fixing plant combination with an efficient carbon fixation; Obtaining terrain data of each location in the garden, the terrain data is the altitude of each location in the garden; performing normalization after obtaining the altitude of each location in the garden; dividing the altitude of each location in the garden by the highest altitude of the garden to obtain the altitude coefficient of each location in the garden, which is recorded as a1, a2, a3, and a4; Then, the carbon dioxide content at the corresponding position is obtained by using an infrared carbon dioxide sensor; a certain time period at night is selected for detection, and the same detection time interval is selected to use the infrared carbon dioxide sensor to perform real-time detection at various positions in the garden, and the detection results are recorded, and the carbon dioxide content at various positions in the garden is normalized, and the normalized value of the carbon dioxide concentration at various positions in the garden is used as the carbon dioxide content coefficient at various positions in the garden and recorded as b1, b2, b3, and b4; The carbon fixation demand coefficient of each location in the garden is calculated by the altitude coefficient of each location in the garden and the carbon dioxide content coefficient of each location in the garden. The carbon fixation demand of each location in the garden is analyzed according to the carbon fixation demand coefficient of each location in the garden. The formula is: , where X i is the carbon fixation demand coefficient of each location in the garden, and i is the serial number of each location in the garden; By setting weights for the carbon sequestration demand coefficients at each location in the garden, the carbon sequestration demand of the entire garden is divided into regions, the carbon sequestration demand coefficients at each location in the garden are sorted from small to large, and the weights are set using the hierarchical analysis method to determine the carbon sequestration demand weights at each location in the garden; After obtaining the carbon sequestration demand weights of each location in the garden, set the regional carbon sequestration threshold for the total carbon sequestration in the garden according to the carbon sequestration demand weights of each location in the garden; multiply the total carbon sequestration in the garden by the carbon sequestration demand weights of the corresponding locations in the garden to obtain the regional carbon sequestration threshold; In step S3, the total amount of carbon sequestration in the garden and the carbon sequestration data at each location in the garden within a fixed time period are obtained; then the carbon sequestration data at each location in the garden are compared with the regional carbon sequestration threshold; If the carbon fixation data of a certain location in the garden is greater than or equal to the carbon fixation threshold of the corresponding area, a combination of carbon fixation plants with well-developed root systems will be selected; if the carbon fixation data of a certain location in the garden is less than the carbon fixation threshold of the corresponding area, a combination of carbon fixation plants with efficient carbon fixation will be selected; After selecting a good carbon-fixing plant combination for planting, collect planting feedback data at each location in the garden and emergency environmental data at each location in the garden; the planting feedback data at each location in the garden include changes in oxygen concentration at each location in the garden and changes in soil organic carbon content at each location in the garden; After selecting the carbon fixation plant combination for planting at each location in the garden, the oxygen concentration at each location in the garden and the soil organic carbon content at each location in the garden were detected in the same time period as before the carbon fixation plant combination was planted; the average oxygen concentration at each location in the garden before and after the carbon fixation plant combination was planted and the average soil organic carbon content at each location in the garden were calculated and recorded as ;in, and The average oxygen concentration at each location in the garden before planting the carbon-fixing plant combination and the average soil organic carbon content at each location in the garden; and The average oxygen concentration at each location in the garden and the average soil organic carbon content at each location in the garden after planting the carbon fixation plant combination; the coefficient of change of oxygen concentration at each location in the garden and the coefficient of change of soil organic carbon content at each location in the garden are calculated, and the formula is: , ; Among them O i is the coefficient of variation of oxygen concentration at each location in the garden, H i is the coefficient of variation of soil organic carbon content at each location in the garden; After planting carbon-fixing plant combinations at various locations in the garden, the number of plants at the corresponding location is recorded. When an emergency occurs at a certain location, the number of plants is detected using a laser radar. The survival rate of the plant combination at the corresponding location is obtained by dividing the number of plants detected by the laser radar by the number of plants at the corresponding location, which is recorded as Z. In step S4, the oxygen concentration variation coefficient O of each position in the garden is calculated. i , the coefficient of variation of soil organic carbon content at each location in the garden H i and the survival rate of plant combinations is denoted as Z, which is defined as the input variable and divided into different fuzzy sets; Define the risk level as the output variable and divide it into fuzzy sets; Formulate a set of fuzzy rules to describe the impact of different input variables on the output variables; Perform fuzzy reasoning based on fuzzy rules to determine the risk level of the corresponding location.

2. The method for designing a carbon-fixing plant combination based on reinforcement learning according to claim 1, characterized in that: In step S1, the tree species data includes the tree species life cycle and the tree species gibberellin content; the garden data includes the garden terrain data and the garden carbon dioxide content.

3. A carbon fixation plant combination design system based on reinforcement learning, used to implement a carbon fixation plant combination design method based on reinforcement learning as described in any one of claims 1-2, characterized in that : It includes data acquisition module, data analysis module and data storage module, and the signal connection between each module; The data acquisition module is used to obtain tree species data and garden data and send them to the data analysis module; The data analysis module determines the carbon sequestration plant combination and its risk level at each location based on tree species data and garden data: The data storage module is used to store all data in the data processing process of the carbon fixation plant combination design system based on reinforcement learning.

Images