An intelligent adjustment system for agricultural greenhouses

Abstract

The present invention provides an intelligent regulation system for agricultural planting greenhouses, and relates to the technical field of intelligent regulation of greenhouses. The system includes: an environmental parameter module, in the i-th regulation cycle, determining the environmental parameter operation according to the environmental parameter model of the i-1th regulation cycle; a sample plant module, sampling to obtain sample plants and sample leaves, obtaining plant height, leaf number, leaf width and natural distance; a growth coefficient module, determining plant growth vectors; a judgment training module, judging whether the environmental regulation model of the i-1th regulation cycle needs training; a loss function module, determining a loss function according to the i-th regulation cycle and the plant growth vector of the i-1th regulation cycle; a training module, training according to the loss function, obtaining the environmental regulation model of the i-th regulation cycle; and a regulation module, determining the environmental parameters of the i+1th regulation cycle. According to the present invention, the rationality of greenhouse parameter regulation can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of greenhouse intelligent regulation, and in particular to an agricultural planting greenhouse intelligent regulation system. Background Art

[0002] In the related technology, CN115812485A discloses an intelligent adjustment system for an agricultural planting greenhouse, which relates to the field of greenhouse technology, including a main control unit, an air temperature detection unit, an air humidity detection unit, a carbon dioxide detection unit, a light detection unit, a data storage unit, a temperature regulator, a humidity adjustment mechanism, ventilation equipment and an intelligent monitoring terminal, and the humidity adjustment mechanism is composed of water supply equipment, a shunt pipe, a spray pipe, a turntable and a spray assembly; the scheme facilitates real-time monitoring of the temperature and humidity inside the greenhouse through the cooperation of the main control unit, the temperature regulator and the humidity adjustment mechanism, and can automatically perform rapid humidification and temperature control operations on the inside of the greenhouse according to the set humidity threshold, effectively improving the work efficiency when performing humidification operations on the humidity inside the greenhouse, and at the same time can perform uniform spraying operations on the crops inside the greenhouse; greatly improving the convenience of regulating the temperature and humidity inside the greenhouse.

[0003] CN113133364B discloses a greenhouse temperature and humidity intelligent control method and system, the method includes: comparing the current indoor and outdoor temperature difference of the target greenhouse with the preset indoor temperature difference, obtaining the comparison result, uploading the comparison result to the preset server terminal, detecting the indoor humidity of the target greenhouse, calculating the indoor dryness of the target greenhouse according to the indoor humidity and the indoor temperature, feeding back the indoor dryness to the preset server terminal, using the preset server terminal to confirm whether it is necessary to adjust the external roller shutter or vent and whether it is necessary to humidify the air according to the comparison result and the indoor dryness, and when it is confirmed that the external roller shutter or vent needs to be adjusted and the air needs to be humidified, the control device of the external roller shutter or vent and the indoor air humidification device of the target greenhouse are started to intelligently adjust the temperature and humidity. The temperature and humidity in the target greenhouse are intelligently controlled without manual adjustment, saving labor costs.

[0004] Based on the above related technologies, the technical problem of unreasonable greenhouse adjustment can be solved. However, the related technologies do not take into account the different environmental requirements of different plant growth conditions, that is, it is impossible to accurately adjust the greenhouse parameters according to the real-time growth conditions of the plants.

[0005] The information disclosed in the background technology section of this application is only intended to deepen the understanding of the general background technology of this application, and should not be regarded as an admission or any form of suggestion that the information constitutes the prior art already known to those skilled in the art. Summary of the invention

[0006] The present invention provides an intelligent regulating system for an agricultural planting greenhouse, which can solve the technical problem that related technologies cannot accurately regulate greenhouse parameters according to the real-time growth conditions of plants.

[0007] According to a first aspect of an embodiment of the present invention, there is provided an intelligent adjustment system for an agricultural greenhouse, comprising:

[0008] Environmental parameter module, in the i-th adjustment cycle, the greenhouse operates according to the environmental parameters determined by the environmental parameter model of the i-1-th adjustment cycle, i is a positive integer greater than or equal to 1, and when i=1, the greenhouse operates according to the preset environmental parameters;

[0009] The sample plant module samples all plants to obtain sample plants, samples leaves of the sample plants to obtain sample leaves, obtains the plant height and the number of leaves of the sample plants in the i-th adjustment cycle, and obtains the leaf width and the natural distance of the leaf edge of the sample leaves in the i-th adjustment cycle;

[0010] A growth coefficient module, determining a plant growth vector according to the plant height, the number of leaves, the leaf width and the natural distance;

[0011] A judgment training module, judging whether the environmental regulation model of the i-1th regulation period needs to be trained according to the plant growth vector of the sample plant of the i-th regulation period and the plant growth vector of the sample plant of the i-1th regulation period;

[0012] A loss function module, when the environmental regulation model of the i-1th regulation period needs to be trained, determines the loss function of the environmental regulation model of the i-1th regulation period according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period;

[0013] A training module, which trains the environmental adjustment model of the i-1th adjustment period according to the loss function of the environmental adjustment model of the i-1th adjustment period to obtain the environmental adjustment model of the i-th adjustment period;

[0014] The adjustment module determines the environmental parameters of the (i+1)th adjustment period through the environmental adjustment model of the (i)th adjustment period.

[0015] According to a second aspect of an embodiment of the present invention, there is provided an intelligent adjustment method for an agricultural greenhouse, comprising:

[0016] In the i-th adjustment cycle, the greenhouse operates according to the environmental parameters determined by the environmental parameter model of the i-1-th adjustment cycle, where i is a positive integer greater than or equal to 1. When i=1, the greenhouse operates according to the preset environmental parameters;

[0017] Sampling all plants to obtain sample plants, sampling the sample plants to obtain sample leaves, obtaining the plant height and the number of leaves of the sample plants in the i-th adjustment cycle, and obtaining the leaf width and the natural distance of the leaf edge of the sample leaves in the i-th adjustment cycle;

[0018] determining a plant growth vector according to the plant height, the number of leaves, the leaf width and the natural distance;

[0019] According to the plant growth vector of the sample plant in the i-th adjustment period and the plant growth vector of the sample plant in the i-1th adjustment period, determining whether the environmental adjustment model of the i-1th adjustment period needs to be trained;

[0020] When the environmental regulation model of the i-1th regulation period needs to be trained, determining the loss function of the environmental regulation model of the i-1th regulation period according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period;

[0021] According to the loss function of the environmental adjustment model of the i-1th adjustment period, training the environmental adjustment model of the i-1th adjustment period to obtain the environmental adjustment model of the i-th adjustment period;

[0022] The environmental parameters of the (i+1)th adjustment period are determined by using the environmental adjustment model of the i-th adjustment period.

[0023] According to the third aspect of the present invention, there is provided an intelligent adjustment device for an agricultural greenhouse, comprising: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to call the instructions stored in the memory to execute the intelligent adjustment method for the agricultural greenhouse.

[0024] According to a fourth aspect of the present invention, there is provided a computer-readable storage medium having computer program instructions stored thereon, wherein the computer program instructions, when executed by a processor, implement the intelligent adjustment method for the agricultural greenhouse.

[0025] Technical effect: According to the present invention, the growth status of plants can be accurately analyzed during the growth of plants, the loss function can be determined based on the growth status of plants, and the environmental regulation model can be trained based on the loss function, so that the growth status of plants can be optimized when the environmental parameters in the greenhouse are adjusted. When determining the plant height growth coefficient, the plant height growth coefficient can be determined based on the plant height and the plant height change rate. During the calculation process, the plant height growth status is determined from two aspects: the change in plant height growth rate and the plant height, thereby improving the objectivity, accuracy and scientificity of the plant height growth coefficient. When determining the number of plant leaves growing, the plant leaf growth coefficient can be determined based on the number of leaves, leaf width and natural distance. During the calculation process, the plant leaf growth status can be determined based on the number of leaves and leaf development status, thereby improving the scientificity, accuracy and comprehensiveness of the plant leaf growth coefficient. When determining whether the environmental regulation model needs to be trained, it can be determined whether the environmental regulation model of the i-1 regulation period needs to be trained according to the plant growth vector of the sample plant of the i-1 regulation period and the plant growth vector of the sample plant of the i-1 regulation period. When determining whether the environmental regulation model needs to be trained, the relationship between the plant growth vectors of the sample plants of two adjacent regulation periods is referred to, so as to determine whether the environmental regulation model has reached a stable state and whether the environmental parameters of the environmental regulation model are suitable for the plant growth of the i-1 regulation period. If the conditions for training are met, the environmental regulation model is continuously optimized, thereby improving the accuracy and effectiveness of the environmental regulation model. When training the environmental regulation model, the loss function of the environmental regulation model of the i-1 regulation period is determined by the conditions satisfied by the plant growth vectors of each sample plant and the plant growth vectors of the adjacent regulation periods, so that the environmental regulation model quickly reaches a stable state, and the overall plant height growth coefficient and plant leaf growth coefficient of the sample plants are maximized, and the overall growth conditions of each sample plant are homogenized, thereby improving the training efficiency and the accuracy of the environmental regulation model.

[0026] It should be understood that the above general description and the following detailed description are exemplary and explanatory only and do not limit the present invention. Other features and aspects of the present invention will become more apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other embodiments can be obtained based on these drawings without creative work.

[0028] Figure 1A schematic diagram of an intelligent adjustment system for an agricultural greenhouse according to an embodiment of the present invention is exemplarily shown;

[0029] Figure 2 The following is a flow chart of an intelligent adjustment method for an agricultural greenhouse according to an embodiment of the present invention. DETAILED DESCRIPTION

[0030] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, 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.

[0031] The technical solution of the present invention is described in detail with specific embodiments below. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described in detail in some embodiments.

[0032] Figure 1 A schematic diagram of an intelligent adjustment system for an agricultural greenhouse according to an embodiment of the present invention is exemplarily shown, and the system includes:

[0033] Environmental parameter module, in the i-th adjustment cycle, the greenhouse operates according to the environmental parameters determined by the environmental parameter model of the i-1-th adjustment cycle, i is a positive integer greater than or equal to 1, and when i=1, the greenhouse operates according to the preset environmental parameters;

[0034] The sample plant module samples all plants to obtain sample plants, samples leaves of the sample plants to obtain sample leaves, obtains the plant height and the number of leaves of the sample plants in the i-th adjustment cycle, and obtains the leaf width and the natural distance of the leaf edge of the sample leaves in the i-th adjustment cycle;

[0035] A growth coefficient module, determining a plant growth vector according to the plant height, the number of leaves, the leaf width and the natural distance;

[0036] A judgment training module, judging whether the environmental regulation model of the i-1th regulation period needs to be trained according to the plant growth vector of the sample plant of the i-th regulation period and the plant growth vector of the sample plant of the i-1th regulation period;

[0037] A loss function module, when the environmental regulation model of the i-1th regulation period needs to be trained, determines the loss function of the environmental regulation model of the i-1th regulation period according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period;

[0038] A training module, which trains the environmental adjustment model of the i-1th adjustment period according to the loss function of the environmental adjustment model of the i-1th adjustment period to obtain the environmental adjustment model of the i-th adjustment period;

[0039] The adjustment module determines the environmental parameters of the (i+1)th adjustment period through the environmental adjustment model of the (i)th adjustment period.

[0040] According to the intelligent adjustment system for agricultural greenhouses according to the embodiments of the present invention, the growth conditions of plants can be accurately analyzed during plant growth, a loss function can be determined based on the growth conditions of the plants, and an environmental adjustment model can be trained based on the loss function, thereby optimizing the growth conditions of plants when adjusting the environmental parameters in the greenhouse.

[0041] According to one embodiment of the present invention, in the environmental parameter module, in the i-th adjustment cycle, the greenhouse operates according to the environmental parameters determined by the environmental parameter model of the i-1-th adjustment cycle, i is a positive integer greater than or equal to 1, and when i=1, the greenhouse operates according to the preset environmental parameters.

[0042] For example, the adjustment cycle is set to three days. In the second adjustment cycle, the greenhouse operates according to the environmental parameters determined by the environmental parameter model of the first adjustment cycle. In the first adjustment cycle, the greenhouse operates according to the preset environmental parameters.

[0043] According to one embodiment of the present invention, sampling is performed among all plants to obtain sample plants, and the sample plants are sampled to obtain sample leaves. The plant height and the number of leaves of the sample plants in the i-th adjustment cycle are obtained, and the leaf width and the natural distance of the leaf edge of the sample leaves in the i-th adjustment cycle are obtained.

[0044] For example, the greenhouse is divided into several areas, and two plants are randomly selected in each area as sample plants, so that the growth status of the sample plants can accurately reflect the growth status of the overall plants in the greenhouse, and the plant height and number of leaves of the sample plants are obtained. In addition, when the number of leaves of the plants planted in the greenhouse is small (such as small potted plants), the number of leaves of the sample plants can be manually counted. When the number of leaves of the plants planted in the greenhouse is large (such as large fruit trees), the number of leaves of the sample plants can be estimated by counting the leaf density. Since the natural distance and leaf width of the leaf edges need to be measured manually, it is inefficient to measure all the leaves of the sample plants. All the leaves of the sample plants are sampled to obtain the leaf width and natural distance of the leaf edges of the sample leaves.

[0045] According to one embodiment of the present invention, in the growth coefficient module, the plant height growth coefficient and the plant leaf growth coefficient of the sample plant in the i-th adjustment period are determined according to the plant height, the number of leaves, the leaf width and the natural distance.

[0046] According to an embodiment of the present invention, determining a plant growth vector according to the plant height, the number of leaves, the leaf width and the natural distance includes:

[0047] Fitting the plant height and the time in the i-th adjustment cycle to obtain a plant height function of the plant height in the i-th adjustment cycle;

[0048] According to the plant height function, obtaining a plant height derivative function;

[0049] Determining the height change rate at multiple moments in the i-th adjustment cycle according to the plant height derivative function;

[0050] Determining a plant height growth coefficient according to the plant height and the height change rate;

[0051] Determining a plant leaf growth coefficient according to the leaf quantity, the leaf width and the natural distance;

[0052] A plant growth vector is determined according to the plant height growth coefficient and the plant leaf growth coefficient.

[0053] For example, the plant height and the moments in the adjustment cycle are fitted to obtain a plant height function that is used to describe the law of plant height changes over time in the adjustment cycle; the plant height function is differentiated to obtain a plant height derivative function; multiple moments in the adjustment cycle are substituted into the plant height derivative function to determine the plant height change rate at multiple moments in the adjustment cycle; based on the plant height change rate and the plant height, the plant height growth coefficient can be determined.

[0054] According to one embodiment of the present invention, determining the plant height growth coefficient according to the plant height and the height change rate includes: determining the plant height growth coefficient HEI of the sth sample plant in the ith adjustment period according to formula (1): p,i,s ,

[0055]

[0056] Among them, α1 and α2 are preset weights, H,1(j) is the height change rate of the sth sample plant at the jth moment of the ith adjustment cycle, and H e is the plant height of the s-th sample plant at the end of the ith adjustment cycle, H T is the preset plant height threshold, n is the number of moments in the adjustment cycle, n is an even number, j≤n, and both j and n are positive integers.

[0057] According to one embodiment of the present invention, Before the adjustment cycle The sum average of the height change rates at each moment, After the adjustment cycle The sum average of the height change rates at each moment, Before the adjustment cycle The sum of the average values ​​of the height change rates at each moment is The smaller the ratio is, the shorter the adjustment period will be. The height change rate at the moment is compared with the previous The greater the difference in height change rate at each moment, the smaller the plant height growth coefficient, that is, the greater the decrease in plant height growth rate. e -H T is the difference between the plant height at the end of the adjustment cycle and the preset plant height threshold, H e -H T The difference between the plant height at the end of the adjustment cycle and the preset plant height threshold, the larger the difference is, the better the growth condition of the sample plant is, and the larger the plant height growth coefficient is.

[0058] In this way, the plant height growth coefficient can be determined based on the plant height and the plant height change rate. During the calculation process, the plant height growth status is determined from two aspects: the change in plant height growth rate and the plant height, which improves the objectivity, accuracy and scientificity of the plant height growth coefficient.

[0059] According to one embodiment of the present invention, determining the plant leaf growth coefficient according to the leaf quantity, the leaf width and the natural distance includes:

[0060] Sampling all leaves of the sample plant to obtain sample leaves;

[0061] According to formula (2), the plant leaf growth coefficient LEA of the sth sample plant in the ith adjustment period is determined as follows: p,i,s ,

[0062]

[0063] Among them, N e,lea is the number of leaves of the sth sample plant at the end of the ith adjustment cycle, lw k,j is the maximum leaf width of the kth sample leaf of the sth sample plant at the jth moment of the ith adjustment cycle, ln k,j The natural distance of the kth sample leaf of the sth sample plant at the jth moment in the ith adjustment cycle, K is the number of sample leaves, k≤K, and both k and K are positive integers.

[0064] According to one embodiment of the present invention, is the ratio of the difference between the maximum leaf width of the kth sample leaf of the sth sample plant at the jth moment in the ith adjustment cycle and the natural distance to the leaf edge, to the maximum leaf width of the kth sample leaf at the jth moment in the adjustment cycle, indicating the leaf curling degree of the kth sample leaf of the sth sample plant at the jth moment in the adjustment cycle. To sum and average the number of moments in the adjustment cycle, the average leaf curling degree of the kth sample leaf at each moment in the adjustment cycle is expressed as: The summation and average are based on the number of sample leaves, which represents the average leaf curling degree of each sample leaf at each moment of the adjustment cycle. The greater the leaf curling degree, the more abnormal the leaf development, and the smaller the plant leaf growth coefficient. To determine the leaf growth coefficient LEA of plants based on the number of leaves and the degree of leaf curling p .

[0065] In this way, the plant leaf growth coefficient can be determined based on the number of leaves, leaf width and natural distance. During the calculation process, the plant leaf growth status can be determined by both the number of leaves and the leaf development status, which improves the scientificity, accuracy and comprehensiveness of the plant leaf growth coefficient.

[0066] According to one embodiment of the present invention, in the judgment training module, based on the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-th adjustment cycle and the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-1th adjustment cycle, it is judged whether the environmental adjustment model of the i-1th adjustment cycle needs to be trained.

[0067] According to an embodiment of the present invention, judging whether the environmental regulation model of the i-1th regulation period needs to be trained according to the plant growth vector of the sample plant of the i-th regulation period and the plant growth vector of the sample plant of the i-1th regulation period includes: obtaining plant height growth condition A, plant leaf growth condition B and balanced growth condition C according to formula (3);

[0068]

[0069] Among them, β1 and β2 are preset weights, HEI p,i,s HEI is the plant height growth coefficient of the sth sample plant in the ith adjustment period, p,i-1,s is the plant height growth coefficient of the s-th sample plant in the i-1th adjustment period, LEA p,i,s is the leaf growth coefficient of the sth sample plant in the ith adjustment period, LEA p,i-1,s HEI is the leaf growth coefficient of the s-th sample plant in the i-1th adjustment cycle, p,i,max is the maximum value of the plant height growth coefficient of multiple sample plants in the i-th adjustment period, HEI p,i,min is the minimum value of the plant height growth coefficient of multiple sample plants in the ith adjustment period, LEA p,i,max is the maximum value of the plant leaf growth coefficient of multiple sample plants in the i-th adjustment period, LEA p,i,min The minimum value of the plant leaf growth coefficient of multiple sample plants in the i-th adjustment period, T P is the growth uniformity coefficient, S is the number of sample plants, s≤S, and both s and S are positive integers;

[0070] When at least one of the first condition A, the second condition B and the third condition C is satisfied, it is determined that the environment adjustment model at the (i-1)th moment needs to be trained.

[0071] According to one embodiment of the present invention, in plant height growth condition A, The average value of the difference between the plant height growth coefficient of the S-th sample plant in the i-th adjustment cycle and the plant height growth coefficient in the i-1-th adjustment cycle is less than or equal to 0, which means that the plant height growth coefficient in the i-th adjustment cycle is less than the plant height growth coefficient in the i-1-th adjustment cycle, and the plant height growth rate slows down, indicating that the environmental parameters of the i-1-th adjustment cycle are set incorrectly, or the environmental parameters of the i-1-th adjustment cycle are not suitable for plant growth in the i-th adjustment cycle, and the environmental adjustment model needs to be trained.

[0072] According to one embodiment of the present invention, in plant leaf growth condition B, The average value of the difference between the plant leaf growth coefficient of the S-th sample plant in the i-th adjustment cycle and the plant leaf growth coefficient in the i-1-th adjustment cycle is less than or equal to 0, which means that the plant leaf growth coefficient in the i-th adjustment cycle is less than the plant leaf growth coefficient in the i-1-th adjustment cycle, and the growth rate of the plant leaves slows down, indicating that the environmental parameters of the i-1-th adjustment cycle are set incorrectly, or the environmental parameters of the i-1-th adjustment cycle are not suitable for plant growth in the i-th adjustment cycle, and the environmental adjustment model needs to be trained.

[0073] According to one embodiment of the present invention, in the balanced growth condition C, is the ratio of the difference between the maximum and minimum values ​​of the plant height growth coefficients of multiple sample plants in the i-th adjustment period to the maximum value of the plant height growth coefficients of multiple sample plants, indicating the degree of uneven height growth of multiple sample plants. is the ratio of the difference between the maximum and minimum values ​​of the plant leaf growth coefficients of multiple sample plants in the i-th adjustment period to the maximum value of the plant leaf growth coefficients of multiple sample plants, indicating the degree of uneven growth of the leaves of multiple sample plants. The weighted summation is based on the height growth differences and leaf growth differences of multiple sample plants, which indicates the overall uneven growth of multiple sample plants. The overall growth unevenness of multiple sample plants is greater than or equal to the growth uniformity coefficient, which means that the growth unevenness of multiple sample plants is too large and uneven, and some sample plants have poor growth and development. It is necessary to train the environmental regulation model.

[0074] According to one embodiment of the present invention, when at least one of the above plant height growth condition A, plant leaf growth condition B and balanced growth condition C is met, it is determined that the environmental adjustment model at the i-1th moment needs to be trained.

[0075] In this way, whether the environmental regulation model of the i-1th regulation cycle needs to be trained can be judged based on the plant growth vectors of the sample plants in the i-1th regulation cycle and the plant growth vectors of the sample plants in the i-1th regulation cycle. When judging whether the environmental regulation model needs to be trained, the relationship between the plant growth vectors of the sample plants in two adjacent regulation cycles is referred to, so as to judge whether the environmental regulation model has reached a stable state and whether the environmental parameters of the environmental regulation model are suitable for the growth of plants in the i-th regulation cycle. If the conditions for training are met, the environmental regulation model is continuously optimized to improve the accuracy and effectiveness of the environmental regulation model.

[0076] According to one embodiment of the present invention, in the loss function module, when the environmental regulation model of the i-1th regulation cycle needs to be trained, the loss function of the environmental regulation model of the i-1th regulation cycle is determined according to the plant growth vector of the sample plant of the i-1th regulation cycle and the plant growth vector of the sample plant of the i-1th regulation cycle.

[0077] According to an embodiment of the present invention, when the environmental regulation model of the i-1th regulation period needs to be trained, the loss function of the environmental regulation model of the i-1th regulation period is determined according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period, including:

[0078] In the plant height growth condition A, the plant leaf growth condition B and the balanced growth condition C, determine the conditions satisfied by the plant height growth coefficient and the plant leaf growth coefficient of the sample plant in the i-th adjustment period and the plant height growth coefficient and the plant leaf growth coefficient of the sample plant in the i-1th adjustment period;

[0079] According to the conditions satisfied by the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-th adjustment cycle and the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-1th adjustment cycle, as well as the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-th adjustment cycle and the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-1th adjustment cycle, the loss function of the environmental adjustment model of the i-1th adjustment cycle is determined.

[0080] According to an embodiment of the present invention, when the environmental regulation model of the i-1th regulation period needs to be trained, the loss function of the environmental regulation model of the i-1th regulation period is determined according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period, including: determining the loss function of the environmental regulation model of the i-1th regulation period according to formula (4),

[0081]

[0082] Among them, θ1 and θ2 are preset parameters, and θ1<θ2, HEI p,i-1,max is the maximum value of the plant height growth coefficient of multiple sample plants in the i-1th adjustment period, HEI p,i-1,min is the minimum value of the plant height growth coefficient of multiple sample plants in the i-1th adjustment period, LEA p,i-1,max is the maximum value of the plant leaf growth coefficient of multiple sample plants in the i-1th adjustment period, LEA p,i-1,minis the minimum value of the plant leaf growth coefficient of multiple sample plants in the i-1th adjustment period, T H is the plant height growth coefficient growth threshold, T L is the growth threshold of plant leaf coefficient, and if is the conditional function.

[0083] According to one embodiment of the present invention, in formula (4), the conditional function Indicates the following two situations in the form of conditional functions. When When , the value of the above conditional function is 0, and the parameters related to plant height growth in the loss function do not need to be adjusted. When , that is, when the average value of the difference between the plant height growth coefficient of the i-th adjustment period and the plant height growth coefficient of the i-1-th adjustment period of the S sample plants is less than or equal to 0, the value of the conditional function is That is, the difference between the plant leaf growth coefficient growth threshold and the average value of the difference between the plant height growth coefficients of two adjacent adjustment cycles is used as the loss function. In the process of training the environmental adjustment model, the difference is narrowed so that the average value of the difference between the plant height growth coefficients of two adjacent adjustment cycles is close to T H , meet the plant height growth standards and make the environmental regulation model reach a stable state as soon as possible.

[0084] According to one embodiment of the present invention, in formula (4), the conditional function The following two situations are expressed in the form of When , the above conditional function takes 0, and the environmental parameters related to plant leaf growth in the greenhouse do not need to be adjusted. When , that is, when the average value of the difference between the plant leaf growth coefficient of the i-th adjustment cycle and the plant leaf growth coefficient of the i-1-th adjustment cycle of the S sample plants is less than or equal to 0, the value of the conditional function is That is, the difference between the plant leaf growth coefficient growth threshold and the average value of the difference between the plant leaf growth coefficients of two adjacent adjustment cycles is used as the loss function. In the process of training the environmental adjustment model, the difference is narrowed so that the average value of the difference between the plant leaf growth coefficients of two adjacent adjustment cycles is close to T L , meet the growth standards of plant leaves and make the environmental regulation model reach a stable state as soon as possible.

[0085] According to one embodiment of the present invention, in formula (4), the conditional function

[0086] The following two situations are expressed in the form of conditional functions: When the overall growth unevenness of each sample plant in the i-th adjustment period is less than the overall growth unevenness of each sample plant in the i-1-th adjustment period, the value of the conditional function is That is, the average value of the difference between the plant height growth coefficient, plant leaf growth coefficient of each sample plant and the average plant height growth coefficient, average plant leaf growth coefficient is weighted summed, and during the training process, the average value of the difference is reduced, thereby improving the average of the overall growth condition of the plant. When the overall growth unevenness of each sample plant in the i-th adjustment period is greater than or equal to the overall growth unevenness of each sample plant in the i-1th adjustment, it means that the environmental parameters adjusted by the environmental adjustment model are wrong, and the conditional function value at this time is Since θ1<θ2, when errors occur in the environmental parameters adjusted by the environmental adjustment model, the training intensity of the model can be increased by using larger preset parameters to quickly correct the adjustment errors of the environmental adjustment model.

[0087] According to an embodiment of the present invention, in formula (4), the above three items may be summed to obtain the loss function of the environment adjustment model in the (i-1)th adjustment cycle.

[0088] In this way, the loss function of the environmental regulation model in the i-1th regulation period is determined by the conditions satisfied by the plant growth vectors of each sample plant and the plant growth vectors of adjacent regulation periods, so that the environmental regulation model can quickly reach a stable state, maximize the overall plant height growth coefficient and plant leaf growth coefficient of the sample plants, and make the overall growth conditions of each sample plant uniform, thereby improving training efficiency and the accuracy of the environmental regulation model.

[0089] According to an embodiment of the present invention, in the training module, the environmental adjustment model of the i-1th adjustment period is trained according to the loss function of the environmental adjustment model of the i-1th adjustment period to obtain the environmental adjustment model of the i-th adjustment period.

[0090] According to an embodiment of the present invention, in the adjustment module, the environmental parameters of the (i+1)th adjustment period are determined by using the environmental adjustment model of the (i)th adjustment period.

[0091] For example, based on the environmental adjustment model of the ith adjustment cycle that has been debugged, the environmental parameters of the (i+1)th adjustment cycle are determined to maximize the plant height growth coefficient and the plant leaf growth coefficient and to make the growth conditions of each plant uniform.

[0092] According to the intelligent regulation system for agricultural greenhouses of the embodiments of the present invention, the growth status of plants can be accurately analyzed during the growth of plants, the loss function can be determined by the growth status of plants, and the environmental regulation model can be trained according to the loss function, so that the growth status of plants can be optimized when the environmental parameters in the greenhouse are adjusted. When determining the plant height growth coefficient, the plant height growth coefficient can be determined according to the plant height and the plant height change rate. In the calculation process, the plant height growth status is determined from the plant height growth rate change and the plant height, respectively, which improves the objectivity, accuracy and scientificity of the plant height growth coefficient. When determining the number of plant leaf growth, the plant leaf growth coefficient can be determined according to the number of leaves, the leaf width and the natural distance. In the calculation process, the plant leaf growth status can be determined by the number of leaves and the leaf development status, which improves the scientificity, accuracy and comprehensiveness of the plant leaf growth coefficient. When determining whether the environmental regulation model needs to be trained, it can be determined whether the environmental regulation model of the i-1 regulation period needs to be trained according to the plant growth vector of the sample plant of the i-1 regulation period and the plant growth vector of the sample plant of the i-1 regulation period. When determining whether the environmental regulation model needs to be trained, the relationship between the plant growth vectors of the sample plants of two adjacent regulation periods is referred to, so as to determine whether the environmental regulation model has reached a stable state and whether the environmental parameters of the environmental regulation model are suitable for the plant growth of the i-1 regulation period. If the conditions for training are met, the environmental regulation model is continuously optimized, thereby improving the accuracy and effectiveness of the environmental regulation model. When training the environmental regulation model, the loss function of the environmental regulation model of the i-1 regulation period is determined by the conditions satisfied by the plant growth vectors of each sample plant and the plant growth vectors of the adjacent regulation periods, so that the environmental regulation model quickly reaches a stable state, and the overall plant height growth coefficient and plant leaf growth coefficient of the sample plants are maximized, and the overall growth conditions of each sample plant are homogenized, thereby improving the training efficiency and the accuracy of the environmental regulation model.

[0093] Figure 2 A flow chart of an intelligent adjustment method for an agricultural greenhouse according to an embodiment of the present invention is exemplarily shown, and the method includes:

[0094] Step S101, in the i-th adjustment cycle, the greenhouse operates according to the environmental parameters determined by the environmental parameter model of the i-1-th adjustment cycle, i is a positive integer greater than or equal to 1, and when i=1, the greenhouse operates according to the preset environmental parameters;

[0095] Step S102, sampling all plants to obtain sample plants, sampling the sample plants to obtain sample leaves, obtaining the plant height and the number of leaves of the sample plants in the i-th adjustment cycle, and obtaining the leaf width and the natural distance of the leaf edge of the sample leaves in the i-th adjustment cycle;

[0096] Step S103, determining a plant growth vector according to the plant height, the number of leaves, the leaf width and the natural distance;

[0097] Step S104, judging whether the environmental regulation model of the i-1th regulation period needs to be trained according to the plant growth vector of the sample plant in the i-1th regulation period and the plant growth vector of the sample plant in the i-1th regulation period;

[0098] Step S105, when the environmental regulation model of the i-1th regulation period needs to be trained, determining the loss function of the environmental regulation model of the i-1th regulation period according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period;

[0099] Step S106, training the environment adjustment model of the i-1th adjustment period according to the loss function of the environment adjustment model of the i-1th adjustment period to obtain the environment adjustment model of the i-th adjustment period;

[0100] Step S107: determining the environmental parameters of the (i+1)th adjustment period through the environmental adjustment model of the i-th adjustment period.

[0101] According to one embodiment of the present invention, there is provided an intelligent adjustment device for an agricultural greenhouse, comprising: a processor; a memory for storing instructions executable by the processor; wherein the processor is configured to call the instructions stored in the memory to execute the intelligent adjustment method for the agricultural greenhouse.

[0102] According to one embodiment of the present invention, there is provided a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the intelligent adjustment method for the agricultural greenhouse is implemented.

[0103] The present invention may be a method, an apparatus, a system and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for executing various aspects of the present invention.

[0104] It should be understood by those skilled in the art that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functional and structural principles of the present invention have been demonstrated and explained in the embodiments, and the embodiments of the present invention may be deformed or modified in any way without departing from the principles.

[0105] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent adjustment system for agricultural greenhouses, characterized in that: include: Environmental parameter module: in the i-th adjustment cycle, the greenhouse operates according to the environmental parameters determined by the environmental parameter model of the i-1-th adjustment cycle, where i is a positive integer greater than or equal to 1. When i=1, the greenhouse operates according to the preset environmental parameters; The sample plant module samples all plants to obtain sample plants, samples leaves of the sample plants to obtain sample leaves, obtains the plant height and the number of leaves of the sample plants in the i-th adjustment cycle, and obtains the leaf width and the natural distance of the leaf edge of the sample leaves in the i-th adjustment cycle; A growth coefficient module, determining a plant growth vector according to the plant height, the number of leaves, the leaf width and the natural distance; A judgment training module, judging whether the environmental regulation model of the i-1th regulation period needs to be trained according to the plant growth vector of the sample plant of the i-th regulation period and the plant growth vector of the sample plant of the i-1th regulation period; A loss function module, when the environmental regulation model of the i-1th regulation period needs to be trained, determines the loss function of the environmental regulation model of the i-1th regulation period according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period; A training module, which trains the environmental adjustment model of the i-1th adjustment period according to the loss function of the environmental adjustment model of the i-1th adjustment period to obtain the environmental adjustment model of the i-th adjustment period; The adjustment module determines the environmental parameters of the (i+1)th adjustment period through the environmental adjustment model of the (i)th adjustment period; Determining a plant growth vector according to the plant height, the number of leaves, the leaf width and the natural distance includes: Fitting the plant height and the time in the i-th adjustment cycle to obtain a plant height function of the plant height in the i-th adjustment cycle; According to the plant height function, obtaining a plant height derivative function; Determining the height change rate at multiple moments in the i-th adjustment cycle according to the plant height derivative function; Determining a plant height growth coefficient according to the plant height and the height change rate; Determining a plant leaf growth coefficient according to the leaf quantity, the leaf width and the natural distance; Determining a plant growth vector according to the plant height growth coefficient and the plant leaf growth coefficient; Determining a plant height growth coefficient according to the plant height and the height change rate includes: According to the formula: , determine the plant height growth coefficient of the sth sample plant in the ith regulation period ,in, and is the preset weight, is the height change rate of the s-th sample plant at the j-th moment in the ith adjustment cycle, is the plant height of the s-th sample plant at the end of the ith adjustment cycle, is the preset plant height threshold, n is the number of moments in the adjustment cycle, n is an even number, j≤n, and both j and n are positive integers; Determining a plant leaf growth coefficient according to the leaf quantity, the leaf width and the natural distance includes: Sampling all leaves of the sample plant to obtain sample leaves; According to the formula: , determine the plant leaf growth coefficient of the sth sample plant in the ith adjustment period ,in, is the number of leaves of the s-th sample plant at the end of the ith adjustment cycle, is the maximum leaf width of the kth sample leaf of the sth sample plant at the jth moment of the ith adjustment cycle, The natural distance of the kth sample leaf of the sth sample plant at the jth moment of the ith adjustment cycle, K is the number of sample leaves, k≤K, and both k and K are positive integers; Judging whether the environmental regulation model of the i-1 regulation period needs to be trained according to the plant growth vector of the sample plant of the i-1 regulation period and the plant growth vector of the sample plant of the i-1 regulation period, comprises: According to the formula: , obtain plant height growth condition A, plant leaf growth condition B and balanced growth condition C, where β1 and β2 are preset weights, is the plant height growth coefficient of the s-th sample plant in the i-th adjustment period, is the plant height growth coefficient of the s-th sample plant in the i-1th adjustment cycle, is the plant leaf growth coefficient of the s-th sample plant in the i-th adjustment cycle, is the plant leaf growth coefficient of the s-th sample plant in the i-1th adjustment cycle, is the maximum value of the plant height growth coefficient of multiple sample plants in the i-th adjustment period, is the minimum value of the plant height growth coefficient of multiple sample plants in the i-th adjustment period, is the maximum value of the plant leaf growth coefficient of multiple sample plants in the i-th adjustment period, The minimum value of the plant leaf growth coefficient of multiple sample plants in the i-th adjustment period, T P is the growth uniformity coefficient, S is the number of sample plants, s≤S, and both s and S are positive integers; When at least one of the conditions A, B and C is satisfied, it is determined that the environment adjustment model at the i-1th moment needs to be trained.

2. The intelligent adjustment system for agricultural greenhouse according to claim 1 is characterized in that: When the environmental regulation model of the i-1th regulation period needs to be trained, the loss function of the environmental regulation model of the i-1th regulation period is determined according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period, including: In the plant height growth condition A, the plant leaf growth condition B and the balanced growth condition C, determine the conditions satisfied by the plant height growth coefficient and the plant leaf growth coefficient of the sample plant in the i-th adjustment period and the plant height growth coefficient and the plant leaf growth coefficient of the sample plant in the i-1th adjustment period; According to the conditions satisfied by the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-th adjustment cycle and the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-1th adjustment cycle, as well as the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-th adjustment cycle and the plant height growth coefficient and the plant leaf growth coefficient of the sample plants in the i-1th adjustment cycle, the loss function of the environmental adjustment model of the i-1th adjustment cycle is determined.

3. The intelligent adjustment system for agricultural greenhouse according to claim 2 is characterized in that: When the environmental regulation model of the i-1th regulation period needs to be trained, the loss function of the environmental regulation model of the i-1th regulation period is determined according to the plant growth vector of the sample plant of the i-1th regulation period and the plant growth vector of the sample plant of the i-1th regulation period, including: According to the formula: , determine the loss function of the environmental regulation model in the i-1th regulation cycle, where, and is the preset parameter, and , is the maximum value of the plant height growth coefficient of multiple sample plants in the i-1th adjustment period, is the minimum value of the plant height growth coefficient of multiple sample plants in the i-1th adjustment period, is the maximum value of the plant leaf growth coefficient of multiple sample plants in the i-1th adjustment period, is the minimum value of the plant leaf growth coefficient of multiple sample plants in the i-1th adjustment period, is the plant height growth coefficient growth threshold, is the growth threshold of plant leaf coefficient, and if is the conditional function.

Images