Method and system for secondary flowering of saffron crocus

Abstract

The invention discloses a method and a system for secondary flowering of saffron crocus, and relates to the technical field of cultivation of saffron crocus. Through the design of the overall technical scheme, the yield can be increased; the flowering times of saffron crocus can be increased in one growth cycle through secondary flowering, so that the flower bud yield per unit area is increased; more flowers can be harvested every year, so that the earnings of producers are remarkably improved, especially under the condition that market demands are vigorous; a secondary flowering technology can prolong the flowering phase of saffron crocus, so that the supply time of saffron crocus on the market is prolonged, and the saffron crocus has higher market competitiveness; by regulating and controlling the environment and managing nutrients, agricultural resources such as water and fertilizers can be utilized more efficiently, and resource waste is reduced; and quality improvement: through refined environment control, the flower quality of the saffron crocus can be improved, and the medicinal value and market quality of the saffron crocus can be ensured.

Description

Technical Field

[0001] The invention relates to the technical field of saffron cultivation, and in particular to a saffron secondary flowering method and a system thereof. Background Art

[0002] Saffron (Crocus sativus) is a high-value plant known for its bright purple flowers and medicinal value, especially its dried stamens, which are widely used in the food, medicine and beauty industries. Saffron has a long history of cultivation, but due to its special requirements for the growing environment, the planting process is relatively complicated and the yield is low. In recent years, with the increase in demand for saffron, researchers and agricultural producers have been constantly exploring ways to increase yields and optimize planting techniques, among which the "secondary flowering" technology has become an important direction for improving saffron yield and benefits.

[0003] Saffron is a plant that blooms in autumn and usually blooms only once a year. Its flowering mainly depends on seasonal changes in temperature and humidity, and it usually starts to bloom in the cooler environment of autumn. However, the flowering cycle of saffron is relatively short, and each plant can only bloom once a year. "Secondary flowering" means that after the normal flowering cycle of saffron, the growth environment is artificially controlled or specific technical means are adopted, so that saffron can bloom twice in one growing season. Secondary flowering not only helps to increase the yield of flower buds, but also enhances the economic value of saffron. Through the secondary flowering technology, farmers can obtain more flowers in a shorter period of time, thereby increasing product output and benefits.

[0004] Traditional cultivation methods are restricted by the natural environment and climatic conditions, and the yield fluctuates greatly. Therefore, a new solution to the above problems is needed. Summary of the invention

[0005] The object of the present invention is to provide a method and system for secondary flowering of saffron to solve the technical problems raised in the background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a method for secondary flowering of saffron, comprising at least the following steps:

[0007] S1: Control the light to ensure that saffron receives enough light, which helps promote photosynthesis and secondary flowering. Pay attention to the uniformity of light intensity to avoid local over- or under-intensity.

[0008] S2: Control humidity. Saffron usually prefers a drier environment, but keeping it moderately moist helps it grow healthily.

[0009] S3: Temperature control, maintaining a constant temperature, which is more in line with the growth needs of saffron;

[0010] S4: using supplements to supplement the elements required by saffron, the supplements including but not limited to fermented cow dung, ecological fertilizer and water-soluble fertilizer;

[0011] S5: Build a traceability system to track the entire process of saffron from planting to harvesting to ensure quality and safety;

[0012] S6: Build a monitoring system, use the monitoring system to check the environment and crop growth conditions of the planting base throughout the process, and adjust the planting conditions in real time.

[0013] Furthermore, the illumination in S1 is 4500lux-5500lux, and the illumination control of S1 at least includes algorithm control.

[0014] Furthermore, the humidity in S2 is controlled to be 65%.

[0015] Furthermore, the temperature in S3 is 10°C-25°C.

[0016] A saffron secondary flowering system, comprising a light control module, a humidity control module, a temperature control module, a traceability system and a monitoring system;

[0017] The lighting control module includes but is not limited to lighting lamps and glass greenhouses, which are used to supplement lighting to ensure the light intensity and are controlled by lighting algorithms;

[0018] The humidity control module uses a humidification device in conjunction with a dehumidification device to control humidity. The humidification device includes but is not limited to a modern digital atomization device. The humidification device and the dehumidification device are controlled by a humidity control algorithm.

[0019] The temperature control module uses a self-learning temperature algorithm to control the corresponding temperature regulating device;

[0020] The traceability system includes a data acquisition and recording module, a traceability system function module and an algorithm control module;

[0021] The monitoring system includes at least a variety of different monitoring sensors, visual monitoring equipment and alarm equipment. The monitoring sensors include but are not limited to temperature sensors, light sensors and temperature sensors. The visual monitoring equipment includes but is not limited to multi-angle monitors and infrared cameras. The alarm equipment includes but is not limited to sound and light alarms.

[0022] Furthermore, the illumination algorithm control comprises at least the following steps:

[0023] Calculate deviation: First calculate the deviation between the current ambient light intensity and the target light intensity;

[0024] ΔL=Ltarget -L current

[0025] in:

[0026] If ΔL>0, it means that the current ambient light is insufficient and the artificial light source needs to be brightened. If ΔL<0, it means that the current ambient light is too strong and the artificial light source needs to be dimmed.

[0027] Calculate the brightening ratio: If L current <L min , that is, the current illumination is lower than the target minimum value, and the light source needs to be brightened. The brightness of the brightened artificial light source is L artifical The calculation formula is:

[0028]

[0029] Indicates the relative deficiency between the current illumination and the target minimum value. If ΔL is large, the brightness of the artificial light source needs to be adjusted to 100%;

[0030] Brightening ratio: If L currebt >L max , that is, the current illumination exceeds the target maximum value, and the light source needs to be dimmed. The brightness of the dimmed artificial light source is L artificial The calculation formula is:

[0031]

[0032] Indicates the relative excess between the current light and the target maximum value. If ΔL is large, it means that the current light is too strong and the artificial light source needs to be adjusted to the lowest brightness;

[0033] To summarize:

[0034] If L current Within the target range, i.e. L min ≤L current ≤L max , then the artificial light source does not need to be adjusted and maintains the original brightness. current <L min , then increase the brightness according to the following formula:

[0035]

[0036] If L current >L max , then reduce the brightness according to the following formula:

[0037]

[0038] Among them, L currentis the current ambient light intensity; L min is the minimum value of the target light intensity range; L max : Maximum value of the target light intensity range; L target is the ideal target light intensity, and L artificial The brightness of the artificial light source ranges from 0% to 100%, where 0% is completely off and 100% is fully bright;

[0039] By establishing a feedback system, the lighting control is not only dependent on the current ambient light value, but also can be adaptively adjusted based on historical lighting data and user feedback factors. In this way, the system can better "learn" and optimize the control strategy over time, and adjust the parameters of the lighting control algorithm based on historical errors and current adjusted feedback, see the following formula:

[0040] Error(t) = L c urrent(t)-L target (t)

[0041] L adjusted (t) = L adjusted (t-1)+α·error(t)+β·trend(t)

[0042] Among them, α and β are the learning rate and trend adjustment factor, L adjusted (t) is the final light intensity adjusted according to historical data.

[0043] Furthermore, the humidity control algorithm comprises at least the following steps:

[0044] First, calculate the error between the current humidity and the target humidity to determine whether the humidity needs to be increased or decreased;

[0045] ΔH=H target -H current

[0046] If ΔH>0, it means that the current humidity is lower than the target humidity and humidification is required. If ΔH<0, it means that the current humidity is higher than the target humidity and dehumidification is required.

[0047] If the current humidity is lower than the target humidity, H current <H min , the humidification equipment needs to be started, and the working intensity calculation formula of the humidification equipment is:

[0048]

[0049] If ΔH is large, you may need to adjust the humidification equipment intensity to 100%. Indicates the relative deficiency between the current humidity and the target minimum humidity;

[0050] If the current humidity is higher than the target humidity, H current >H max , then you need to start the dehumidifier, and the working intensity calculation formula of the dehumidifier is:

[0051]

[0052] If ΔH is large, it means that the current humidity is too high and the dehumidifier intensity needs to be adjusted to 100%. Indicates the relative excess between the current humidity and the target maximum humidity;

[0053] To summarize:

[0054] If H current Within the target range, that is, H min ≤H current ≤H max , then the humidification equipment and dehumidifier do not need to be adjusted and remain in their original state;

[0055] If H current <H min , then increase the humidity according to the following formula:

[0056]

[0057] If H current >H max , then reduce the humidity according to the following formula:

[0058]

[0059] Among them, H current is the current ambient humidity; H min is the minimum value of the target humidity range; H max is the maximum value of the target humidity range; H target is the ideal target humidity, and the calculation is H humidifier It is the working intensity of the humidification equipment, ranging from 0% to 00%; H dehumidifier The working intensity of the dehumidifier ranges from 0% to 100%.

[0060] Furthermore, the temperature control algorithm comprises at least the following steps:

[0061] State space and action space, assuming that the state space S and action space A of the temperature control system are defined as follows:

[0062] The state space S represents the discretization interval of the current temperature, S = {s 1 ,s 2 ,…,s n}, each state s i Corresponding to a specific temperature range;

[0063] The action space A represents the temperature control actions that the system can take, and is set as A = {a 1 ,a 2 ,a 3}, where: a 1 To increase the temperature; a 2 To reduce the temperature, a 3 To maintain the current temperature;

[0064] Using the Q-learning update formula, the Q value represents the state s t Next, take action a t After that, the expected value of the cumulative return is obtained, and the core formula for updating the Q value is:

[0065]

[0066] Where: Q(s t ,a t ) is the current state s t Take action a t Q value; α is the learning rate, which determines the influence of the newly acquired information on the Q value update, ranging from 0≤α≤1; r t At time step t, the state s t Execute action a t The immediate reward obtained after the decision is made; γ is the discount factor, which indicates the influence of future rewards on the current decision, ranging from 0≤γ≤1; max a Q(s t+1 ,a) represents the next state s t+1 The maximum Q value among all possible actions a is used to represent the reward of the best future action;

[0067] The reward function is In the temperature control problem, the reward function r t It is used to measure the closeness between the current system temperature and the target temperature. The goal is to keep the temperature within an ideal range.

[0068] Assume the target temperature is T targetι The current temperature is T curreut , the design reward is as follows:

[0069] r t =-|T current -T target |

[0070] This means that if the current temperature T current The closer to the target temperature T target , the higher the reward value, the smaller the deviation, and the greater the reward;

[0071] The state transfer function is that at each time step, after taking a certain action, the state of the system will change. Assume that the state s t Represents the current temperature, action a t May cause temperature changes, the state transfer function is expressed as:

[0072] s t+1 =f(s t ,a t )

[0073] in:

[0074] s t+1 is the next state, i.e. the new temperature state; s t : Current state, that is, current temperature; a t the action to be taken, i.e. heating, cooling or maintaining temperature;

[0075] If heating (a 1 ), then s t+1 =s t +ΔT heating ;

[0076] If refrigeration (a 2 ), then s t+1 =s t -ΔT cooling ;

[0077] If the temperature is maintained (a 3 ), then s t+1 =s t ;

[0078] Update process and optimal strategy:

[0079] By continuously alternating the process of selecting actions, obtaining rewards, and updating Q values, the Q-learning algorithm can eventually learn the optimal action strategy under different states. The optimal strategy π * (s) is achieved by choosing the action that maximizes the Q value:

[0080]

[0081] This means that, in state s, action a is taken to maximize Q(s,a), thereby obtaining the optimal temperature control strategy.

[0082] Furthermore, the data acquisition and recording module includes equipment and sensors, data storage, and data transmission and management;

[0083] The equipment and sensors include temperature and humidity sensors, light intensity sensors, and soil moisture sensors;

[0084] The temperature and humidity sensor is used to monitor the temperature and humidity changes in the greenhouse;

[0085] The light intensity sensor is used to monitor the light intensity in real time;

[0086] The soil moisture sensor is used to detect soil moisture.

[0087] The data storage includes database and data transmission and management;

[0088] The database uses MySQL or MongoDB to store all planting data, including but not limited to the growth records, environmental data and nutrient addition of each batch of saffron. During each planting and processing operation, the system will record the planting data and assign a unique ID to each batch of saffron;

[0089] The data transmission and management includes IoT platform devices and application program interfaces;

[0090] The IoT platform device is connected to the IoT platform via Wi-Fi or LoRa to upload the collected data in real time;

[0091] Application Programming Interface: Adopt application programming interface so that data can be accessed through third-party applications;

[0092] The traceability system function module is used to generate a QR code or barcode. Each batch of saffron is equipped with a QR code or barcode, and the code contains planting batch information, which includes but is not limited to planting time, environmental records and fertilizers used, so that users can access complete planting information by scanning the QR code.

[0093] The algorithm control module tracing system includes abnormal data detection algorithm and optimization algorithm;

[0094] The abnormal data detection algorithm automatically generates an alarm when the temperature, humidity or light exceeds the normal fluctuation range by using mean and standard deviation analysis, and uses the Z-score detection algorithm to remove invalid or unreasonable data and perform outlier filtering;

[0095] The optimization algorithm adopts a regression analysis algorithm, performs regression analysis based on historical data and growth cycles, predicts the optimal environmental conditions for saffron growth, helps adjust the planting environment, and uses a decision tree algorithm or a random forest to predict the effects of different environmental parameters (such as temperature, humidity, and light) on the secondary flowering of saffron, to assist in determining whether the environment needs to be adjusted.

[0096] Compared with the prior art, the present invention has the following beneficial effects:

[0097] The present invention can improve the yield through the design of the overall technical solution: the secondary flowering can increase the flowering times of saffron in one growth cycle, thereby increasing the flower bud yield per unit area;

[0098] Increased economic benefits: Since more flowers can be harvested each year, producers' profits are significantly improved, especially when market demand is strong;

[0099] Prolong the flowering period: The secondary flowering technology can prolong the flowering period of saffron, thereby increasing its supply time in the market and making it more competitive in the market;

[0100] Optimize resource utilization: By regulating the environment and nutrient management, agricultural resources such as water and fertilizer can be used more efficiently and reduce resource waste;

[0101] Improve quality: Through refined environmental control, the quality of saffron flowers can be improved to ensure its medicinal value and market quality;

[0102] Promote technological development: The implementation of this technology has promoted the application of agricultural automation and intelligent equipment, and improved the technical level and management efficiency of modern agriculture. DETAILED DESCRIPTION

[0103] The technical solutions in the embodiments of the present invention will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0104] Embodiment 1:

[0105] A method for secondary flowering of saffron, comprising at least the following steps:

[0106] S1: Control the light to ensure that saffron receives enough light, which helps promote photosynthesis and secondary flowering. Pay attention to the uniformity of light intensity to avoid local over- or under-intensity.

[0107] S2: Control humidity. Saffron usually prefers a drier environment, but keeping it moderately moist helps it grow healthily.

[0108] S3: Temperature control, maintaining a constant temperature, which is more in line with the growth needs of saffron;

[0109] S4: Use supplements to supplement the elements required by saffron, including but not limited to fermented cow dung, ecological fertilizer and water-soluble fertilizer;

[0110] S5: Build a traceability system to track the entire process of saffron from planting to harvesting to ensure quality and safety;

[0111] S6: Build a monitoring system, use the monitoring system to check the environment and crop growth conditions of the planting base throughout the process, and adjust the planting conditions in real time.

[0112] The illumination in S1 is 4500lux-5500lux, and the illumination control of S1 at least includes algorithm control.

[0113] The humidity in S2 was controlled at 65%.

[0114] The temperature in S3 is 10°C-25°C.

[0115] Embodiment 2:

[0116] Based on the above embodiment, this embodiment proposes a system for secondary flowering of saffron, including a light control module, a humidity control module, a temperature control module, a traceability system and a monitoring system;

[0117] The lighting control module includes but is not limited to lighting lamps and glass greenhouses. Lighting lamps and glass greenhouses are used for supplementary lighting to ensure the light intensity, and lighting algorithms are used for control;

[0118] The humidity control module uses a humidification device in conjunction with a dehumidification device to control humidity. The humidification device includes but is not limited to a modern digital atomization device. The humidification device and the dehumidification device are controlled by a humidity control algorithm.

[0119] The temperature control module uses a self-learning temperature algorithm to control the corresponding temperature regulating equipment;

[0120] The traceability system includes data collection and recording module, traceability system function module and algorithm control module;

[0121] The monitoring system includes at least a variety of different monitoring sensors, visual monitoring equipment and alarm equipment. The monitoring sensors include but are not limited to temperature sensors, light sensors and temperature sensors. The visual monitoring equipment includes but is not limited to multi-angle monitors and infrared cameras. The alarm equipment includes but is not limited to sound and light alarms.

[0122] The lighting algorithm control includes at least the following steps:

[0123] Calculate deviation: First calculate the deviation between the current ambient light intensity and the target light intensity.

[0124] ΔL=L target -L current

[0125] in:

[0126] If ΔL>0, it means that the current ambient light is insufficient and the artificial light source needs to be brightened. If ΔL<0, it means that the current ambient light is too strong and the artificial light source needs to be dimmed.

[0127] Calculate the brightening ratio: If L current <L min , that is, the current illumination is lower than the target minimum value, and the light source needs to be brightened. The brightness of the brightened artificial light source is L artifical The calculation formula is:

[0128]

[0129] Indicates the relative deficiency between the current illumination and the target minimum value. If ΔL is large, the brightness of the artificial light source needs to be adjusted to 100%;

[0130] Brightening ratio: If L current >L max , that is, the current illumination exceeds the target maximum value, and the light source needs to be dimmed. The brightness of the dimmed artificial light source is L artificial The calculation formula is:

[0131]

[0132] Indicates the relative excess between the current light and the target maximum value. If ΔL is large, it means that the current light is too strong and the artificial light source needs to be adjusted to the lowest brightness;

[0133] To summarize:

[0134] If l current Within the target range, i.e. L min ≤L current ≤L max , then the artificial light source does not need to be adjusted and maintains the original brightness. current <L min , then increase the brightness according to the following formula:

[0135]

[0136] If L current >L max , then reduce the brightness according to the following formula:

[0137]

[0138] Among them, L current is the current ambient light intensity; L min is the minimum value of the target light intensity range; L max : Maximum value of the target light intensity range; L target is the ideal target light intensity, and L artificial The brightness of the artificial light source ranges from 0% to 100%, where 0% is completely off and 100% is fully bright;

[0139] By establishing a feedback system, the lighting control is not only dependent on the current ambient light value, but also can be adaptively adjusted based on historical lighting data and user feedback factors. In this way, the system can better "learn" and optimize the control strategy over time, and adjust the parameters of the lighting control algorithm based on historical errors and current adjusted feedback, see the following formula:

[0140] Error(t) = L c urrent(t)-L target (t)

[0141] L adjusted (t) = L adjusted (t-1)+α·error(t)+β·trend(t)

[0142] Among them, α and β are the learning rate and trend adjustment factor, L adjusted (t) is the final light intensity adjusted according to historical data.

[0143] The humidity control algorithm includes at least the following steps:

[0144] First, calculate the error between the current humidity and the target humidity to determine whether the humidity needs to be increased or decreased;

[0145] ΔH=H target -H current

[0146] If ΔH>0, it means that the current humidity is lower than the target humidity and humidification is required. If ΔH<0, it means that the current humidity is higher than the target humidity and dehumidification is required.

[0147] If the current humidity is lower than the target humidity, H current <H min , the humidification equipment needs to be started, and the working intensity calculation formula of the humidification equipment is:

[0148]

[0149] If ΔH is large, you may need to adjust the humidification equipment intensity to 100%. Indicates the relative deficiency between the current humidity and the target minimum humidity;

[0150] If the current humidity is higher than the target humidity, H current >H max , then you need to start the dehumidifier, and the working intensity calculation formula of the dehumidifier is:

[0151]

[0152] If ΔH is large, it means that the current humidity is too high and the dehumidifier intensity needs to be adjusted to 100%. Indicates the relative excess between the current humidity and the target maximum humidity;

[0153] To summarize:

[0154] If H current Within the target range, that is, H min ≤H current ≤H max , then the humidification equipment and dehumidifier do not need to be adjusted and remain in their original state;

[0155] If H current <H min , then increase the humidity according to the following formula:

[0156]

[0157] If H current >H max , then reduce the humidity according to the following formula:

[0158]

[0159] Among them, H current is the current ambient humidity; H min is the minimum value of the target humidity range; H max is the maximum value of the target humidity range; H target is the ideal target humidity, and the calculation is H humidifier It is the working intensity of the humidification equipment, ranging from 0% to 00%; H dehumidifier The working intensity of the dehumidifier ranges from 0% to 100%.

[0160] The temperature control algorithm includes at least the following steps:

[0161] State space and action space, assuming that the state space S and action space A of the temperature control system are defined as follows:

[0162] The state space S represents the discretization interval of the current temperature, S = {s 1 ,s 2 ,…,s n}, each state s i Corresponding to a specific temperature range;

[0163] The action space A represents the temperature control actions that the system can take, and is set as A = {a 1 ,a 2 ,a 3}, where: a 1 To increase the temperature; a2 To reduce the temperature, a 3 To maintain the current temperature;

[0164] Using the Q-learning update formula, the Q value represents the state s t Next, take action a t After that, the expected value of the cumulative return is obtained, and the core formula for updating the Q value is:

[0165]

[0166] Where: Q(s t ,a t ) is the current state s t Take action a t Q value; α is the learning rate, which determines the influence of the newly acquired information on the Q value update, ranging from 0≤α≤1; r t At time step t, the state s t Execute action a t The immediate reward obtained after the decision is made; γ is the discount factor, which indicates the influence of future rewards on the current decision, ranging from 0≤γ≤1; max a Q(s t+1 ,a) represents the next state s t+1 The maximum Q value among all possible actions a is used to represent the reward of the best future action;

[0167] The reward function is In the temperature control problem, the reward function r t It is used to measure the closeness between the current system temperature and the target temperature. The goal is to keep the temperature within an ideal range.

[0168] Assume the target temperature is T targetι The current temperature is T curreut , the design reward is as follows:

[0169] r t =-|T current -T target |

[0170] This means that if the current temperature T current The closer to the target temperature T target , the higher the reward value, the smaller the deviation, and the greater the reward;

[0171] The state transfer function is that at each time step, after taking a certain action, the state of the system will change. Assume that the state s t Represents the current temperature, action a t May cause temperature changes, the state transfer function is expressed as:

[0172] st+1 =f(s t ,a t )

[0173] in:

[0174] s t+1 is the next state, i.e. the new temperature state; s t : Current state, that is, current temperature; a t the action to be taken, i.e. heating, cooling or maintaining temperature;

[0175] If heating (a 1 ), then s t+1 =s t +ΔT heating ;

[0176] If refrigeration (a 2 ), then s t+1 =s t -ΔT cooling ;

[0177] If the temperature is maintained (a 3 ), then s t+1 =s t ;

[0178] Update process and optimal strategy:

[0179] By continuously alternating the process of selecting actions, obtaining rewards, and updating Q values, the Q-learning algorithm can eventually learn the optimal action strategy under different states. The optimal strategy π * (s) is achieved by choosing the action that maximizes the Q value:

[0180]

[0181] This means that, in state s, action a is taken to maximize Q(s,a), thereby obtaining the optimal temperature control strategy.

[0182] The data acquisition and recording module includes equipment and sensors, data storage, and data transmission and management;

[0183] Equipment and sensors include temperature and humidity sensors, light intensity sensors, and soil moisture sensors;

[0184] Temperature and humidity sensors are used to monitor temperature and humidity changes in the greenhouse;

[0185] The light intensity sensor is used to monitor the light intensity in real time;

[0186] Soil moisture sensor is used to detect soil moisture.

[0187] Data storage includes database and data transmission and management;

[0188] The database uses MySQL or MongoDB to store all planting data, including but not limited to the growth records, environmental data and nutrient addition of each batch of saffron. The system will record the planting data and assign a unique ID to each batch of saffron during each planting and processing operation;

[0189] Data transmission and management include IoT platform devices and application programming interfaces;

[0190] IoT platform devices connect to the IoT platform via Wi-Fi or LoRa and upload collected data in real time;

[0191] Application Programming Interface: Adopt application programming interface so that data can be accessed through third-party applications;

[0192] The traceability system function module is used to generate QR codes or barcodes. Each batch of saffron is equipped with a QR code or barcode. The code contains the planting batch information, which includes but is not limited to the planting time, environmental records and fertilizers used. Users can access the complete planting information by scanning the QR code.

[0193] The algorithm control module of the traceability system includes abnormal data detection algorithm and optimization algorithm;

[0194] The abnormal data detection algorithm uses mean and standard deviation analysis to automatically generate an alarm when temperature, humidity or light exceeds the normal fluctuation range, and uses the Z-score detection algorithm to remove invalid or unreasonable data and filter outliers;

[0195] The optimization algorithm uses a regression analysis algorithm to perform regression analysis based on historical data and growth cycles to predict the optimal environmental conditions for saffron growth, help adjust the planting environment, and use a decision tree algorithm or random forest to predict the effects of different environmental parameters (such as temperature, humidity, and light) on the secondary flowering of saffron, to assist in determining whether the environment needs to be adjusted.

[0196] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the present invention can be implemented in other specific forms without departing from the spirit or essential features of the present invention. Therefore, the embodiments should be considered exemplary and non-restrictive in all respects, and the scope of the present invention is defined by the appended claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims be included in the present invention.

Claims

1. A method for secondary flowering of saffron, characterized in that: At least the following steps are included: S1: Control the light to ensure that saffron receives enough light, which helps promote photosynthesis and secondary flowering. Pay attention to the uniformity of light intensity to avoid local over- or under-intensity. S2: Control humidity. Saffron usually prefers a drier environment, but keeping it moderately moist helps it grow healthily. S3: Temperature control, maintaining a constant temperature, which is more in line with the growth needs of saffron; S4: using supplements to supplement the elements required by saffron, the supplements including but not limited to fermented cow dung, ecological fertilizer and water-soluble fertilizer; S5: Build a traceability system to track the entire process of saffron from planting to harvesting to ensure quality and safety; S6: Build a monitoring system, use the monitoring system to check the environment and crop growth conditions of the planting base throughout the process, and adjust the planting conditions in real time.

2. The method for secondary flowering of saffron according to claim 1, characterized in that: The illumination in S1 is 4500lux-5500lux, and the illumination control of S1 at least includes algorithm control.

3. A method for secondary flowering of saffron according to claim 1, characterized in that: The humidity in S2 is controlled to be 65%.

4. A method for secondary flowering of saffron according to claim 1, characterized in that: The temperature in S3 is 10°C-25°C.

5. A system for secondary flowering of saffron, used for the method for secondary flowering of saffron according to any one of claims 1 to 4, characterized in that: Including lighting control module, humidity control module, temperature control module, traceability system and monitoring system; The lighting control module includes but is not limited to lighting lamps and glass greenhouses, which are used to supplement lighting to ensure the light intensity and are controlled by lighting algorithms; The humidity control module uses a humidification device in conjunction with a dehumidification device to control humidity. The humidification device includes but is not limited to a modern digital atomization device. The humidification device and the dehumidification device are controlled by a humidity control algorithm. The temperature control module uses a self-learning temperature algorithm to control the corresponding temperature regulating device; The traceability system includes a data acquisition and recording module, a traceability system function module and an algorithm control module; The monitoring system includes at least a variety of different monitoring sensors, visual monitoring equipment and alarm equipment. The monitoring sensors include but are not limited to temperature sensors, light sensors and temperature sensors. The visual monitoring equipment includes but is not limited to multi-angle monitors and infrared cameras. The alarm equipment includes but is not limited to sound and light alarms.

6. The system for secondary flowering of saffron according to claim 5, characterized in that: The illumination algorithm control comprises at least the following steps: Calculate deviation: First calculate the deviation between the current ambient light intensity and the target light intensity; ΔL=L target -L current in: If ΔL>0, it means that the current ambient light is insufficient and the artificial light source needs to be brightened. If ΔL<0, it means that the current ambient light is too strong and the artificial light source needs to be dimmed. Calculate the brightening ratio: If L current <L min , that is, the current illumination is lower than the target minimum value, and the light source needs to be brightened. The brightness of the brightened artificial light source is L artifical The calculation formula is: Indicates the relative deficiency between the current illumination and the target minimum value. If ΔL is large, the brightness of the artificial light source needs to be adjusted to 100%; Brightening ratio: If L current >L max , that is, the current illumination exceeds the target maximum value, and the light source needs to be dimmed. The brightness of the dimmed artificial light source is L artificial The calculation formula is: Indicates the relative excess between the current light and the target maximum value. If ΔL is large, it means that the current light is too strong and the artificial light source needs to be adjusted to the lowest brightness; To summarize: If L current Within the target range, i.e. L min ≤L current ≤L max , then the artificial light source does not need to be adjusted and maintains the original brightness. current <L min , then increase the brightness according to the following formula: If L current >L max , then reduce the brightness according to the following formula: Among them, L current is the current ambient light intensity; L min is the minimum value of the target light intensity range; L max : Maximum value of the target light intensity range; L target is the ideal target light intensity, and L artificial The brightness of the artificial light source ranges from 0% to 100%, where 0% is completely off and 100% is fully bright; By establishing a feedback system, the lighting control is not only dependent on the current ambient light value, but also can be adaptively adjusted based on historical lighting data and user feedback factors. In this way, the system can better "learn" and optimize the control strategy over time, and adjust the parameters of the lighting control algorithm based on historical errors and current adjusted feedback, see the following formula: Error(t) = L c urrent(t)-L target (t) L adjusted (t) = L adjusted (t-1)+α·error(t)+β·trend(t) Among them, α and β are the learning rate and trend adjustment factor, L adjusted (t) is the final light intensity adjusted according to historical data.

7. The system for secondary flowering of saffron according to claim 5, characterized in that: The humidity control algorithm comprises at least the following steps: First, calculate the error between the current humidity and the target humidity to determine whether the humidity needs to be increased or decreased; ΔH=H target -H current If ΔH>0, it means that the current humidity is lower than the target humidity and humidification is required. If ΔH<0, it means that the current humidity is higher than the target humidity and dehumidification is required. If the current humidity is lower than the target humidity, H current <H min , the humidification equipment needs to be started, and the working intensity calculation formula of the humidification equipment is: If ΔH is large, you may need to adjust the humidification equipment intensity to 100%. Indicates the relative deficiency between the current humidity and the target minimum humidity; If the current humidity is higher than the target humidity, H current >H max , then you need to start the dehumidifier, and the working intensity calculation formula of the dehumidifier is: If ΔH is large, it means that the current humidity is too high and the dehumidifier intensity needs to be adjusted to 100%; Indicates the relative excess between the current humidity and the target maximum humidity; To summarize: If H current Within the target range, that is, H min ≤H current ≤H max , then the humidification equipment and dehumidifier do not need to be adjusted and remain in their original state; If H current <H min , then increase the humidity according to the following formula: If H current >H max , then reduce the humidity according to the following formula: Among them, H current is the current ambient humidity; H min is the minimum value of the target humidity range; H max is the maximum value of the target humidity range; H target is the ideal target humidity, and the calculation is H humidifier It is the working intensity of the humidification equipment, ranging from 0% to 00%; H dehumidifier The working intensity of the dehumidifier ranges from 0% to 100%.

8. The system for secondary flowering of saffron according to claim 5, characterized in that: The temperature control algorithm comprises at least the following steps: State space and action space, assuming that the state space S and action space A of the temperature control system are defined as follows: The state space S represents the discretization interval of the current temperature, S = {s1, s2, ..., s n }, each state s i Corresponding to a specific temperature range; The action space A represents the temperature control actions that the system can take, set as A = {a1, a2, a3}, where: a1 is to increase the temperature; a2 is to reduce the temperature, and a3 is to maintain the current temperature; Using the Q-learning update formula, the Q value represents the state s t Next, take action a t After that, the expected value of the cumulative return is obtained, and the core formula for updating the Q value is: Where: Q(s t ,a t ) is the current state s t Take action a t Q value; α is the learning rate, which determines the influence of the newly acquired information on the Q value update, ranging from 0≤α≤1; r t At time step t, the state s t Execute action a t The immediate reward obtained after the decision is made; γ is the discount factor, which indicates the influence of future rewards on the current decision, ranging from 0≤γ≤1; max a Q(s t+1 ,a) represents the next state s t+1 The maximum Q value among all possible actions a is used to represent the reward of the best future action; The reward function is In the temperature control problem, the reward function r t It is used to measure the closeness between the current system temperature and the target temperature. The goal is to keep the temperature within an ideal range. Assume the target temperature is T targetι The current temperature is T curreut , the design reward is as follows: r t =-|T current -T target | This means that if the current temperature T current The closer to the target temperature T target , the higher the reward value, the smaller the deviation, and the greater the reward; The state transfer function is that at each time step, after taking an action, the state of the system will change. Assume that the state s t Represents the current temperature, action a t May cause temperature changes, the state transfer function is expressed as: s t+1 =f(s t ,a t ) in: s t+1 is the next state, i.e. the new temperature state; s t : Current state, that is, current temperature; a t the action to be taken, i.e. heating, cooling or maintaining temperature; If heating (a1) is adopted, then s t+1 =s t +ΔT heating ; If cooling (a2) is adopted, then s t+1 =s t -ΔT cooling ; If the temperature (a3) ​​is maintained, then s t+1 =s t ; Update process and optimal strategy: By continuously alternating the process of selecting actions, obtaining rewards, and updating Q values, the Q-learning algorithm can eventually learn the optimal action strategy under different states. The optimal strategy π * (s) is achieved by choosing the action that maximizes the Q value: This means that, in state s, action a is taken to maximize Q(s,a), thereby obtaining the optimal temperature control strategy.

9. The method for the secondary flowering of saffron according to claim 5, characterized in that: The data acquisition and recording module includes equipment and sensors, data storage, and data transmission and management; The equipment and sensors include temperature and humidity sensors, light intensity sensors, and soil moisture sensors; The temperature and humidity sensor is used to monitor the temperature and humidity changes in the greenhouse; The light intensity sensor is used to monitor the light intensity in real time; The soil moisture sensor is used to detect soil moisture; The data storage includes database and data transmission and management; The database uses MySQL or MongoDB to store all planting data, including but not limited to the growth records, environmental data and nutrient addition of each batch of saffron. During each planting and processing operation, the system will record the planting data and assign a unique ID to each batch of saffron; The data transmission and management includes IoT platform devices and application program interfaces; The IoT platform device is connected to the IoT platform via Wi-Fi or LoRa to upload the collected data in real time; Application Programming Interface: Adopt application programming interface so that data can be accessed through third-party applications; The traceability system function module is used to generate a QR code or barcode. Each batch of saffron is equipped with a QR code or barcode, and the code contains planting batch information, which includes but is not limited to planting time, environmental records and fertilizers used, so that users can access complete planting information by scanning the QR code. The algorithm control module tracing system includes abnormal data detection algorithm and optimization algorithm; The abnormal data detection algorithm automatically generates an alarm when the temperature, humidity or light exceeds the normal fluctuation range by using mean and standard deviation analysis, and uses the Z-score detection algorithm to remove invalid or unreasonable data and perform outlier filtering; The optimization algorithm adopts a regression analysis algorithm, performs regression analysis based on historical data and growth cycles, predicts the optimal environmental conditions for saffron growth, helps adjust the planting environment, and uses a decision tree algorithm or a random forest to predict the impact of different environmental parameters on the secondary flowering of saffron, to assist in determining whether the environment needs to be adjusted.

Images