Visual intelligent agricultural planting control system

Abstract

The invention relates to the technical field of intelligent agriculture, and discloses a visual intelligent agricultural planting control system. The system comprises an environment data acquisition module used for acquiring multi-source environment sensing data; the growth feature modeling module is used for extracting crop growth state features through a morphological analysis algorithm; the environment regulation and control decision module is used for generating an environment regulation and control instruction set by utilizing a dynamic threshold matching algorithm; the visual interaction module is used for generating a three-dimensional farmland live-action simulated diagram through multi-dimensional data fusion processing; the strategy execution module is used for driving agricultural facilities to execute actions by adopting a self-adaptive control algorithm; an abnormity early warning module is further arranged, and abnormity early warning signals are generated through correlation analysis. The system realizes comprehensive acquisition and analysis of agricultural planting environment data, precise environment regulation and control, visual and visual display and abnormal early warning, effectively improves the intelligent and precise level of agricultural planting, improves the crop yield and quality, and assists the development of intelligent agriculture.

Description

Technical Field

[0001] The present invention relates to the technical field of smart agriculture, and particularly to a visual smart agriculture planting control system. Background Art

[0002] In the traditional agricultural planting mode, the planting process highly depends on manual experience. Farmers often manage farmland based on their long-term subjective cognition, which has many drawbacks. In terms of obtaining environmental data, it is difficult to achieve comprehensive, accurate, and real-time monitoring in the traditional way. For example, when monitoring soil humidity, only local measurements may be taken with simple tools, and it is impossible to obtain the accurate distribution of soil humidity across the entire farmland. For light intensity, only a rough judgment can be made, and it is difficult to precisely quantify the light differences in different areas. In the monitoring of air temperature and humidity, manual measurement is not only inefficient but also unable to achieve round-the-clock continuous monitoring, resulting in the inability to promptly grasp the subtle changes in the environment.

[0003] In the assessment of crop growth status, traditional methods also have deficiencies. Farmers mainly judge the growth status by visually observing the appearance of crops. There are no scientific measurement methods for quantitative indicators such as leaf expansion area and stem height growth rate, and it is even more difficult to accurately evaluate the fruit maturity index. This extensive assessment method makes it impossible for farmers to accurately grasp the crop growth stage, and thus they cannot take targeted planting management measures at critical times.

[0004] From the perspective of environmental regulation, traditional agriculture lacks scientific basis and precise control. During irrigation, it is usually carried out according to fixed time and water volume, without considering the actual soil humidity and the water requirements of crops, which easily causes problems such as water resource waste or insufficient irrigation. The opening and closing of sunshade nets and the use of ventilation equipment also mostly rely on manual experience and cannot be dynamically adjusted according to real-time environmental data and crop growth needs, making it difficult to create the most suitable environmental conditions for crop growth.

[0005] In addition, during the traditional agricultural planting process, the presentation and management of data are also relatively backward. It is difficult for farmers to intuitively and comprehensively understand the overall situation of the farmland and conduct effective traceability and analysis of the planting process. In large-scale planting scenarios, the lack of such information management will lead to a significant increase in management costs, while the planting efficiency and quality are difficult to improve. With the advancement of agricultural modernization, the traditional planting mode can no longer meet the growing market demand and sustainable development requirements. There is an urgent need to introduce advanced technical means to build an intelligent and visual planting control system to achieve precise management and efficient development of agricultural production.

[0006] With the development of information technology, sensor technology and artificial intelligence technology, smart agriculture has gradually emerged. However, there are still some problems with the existing smart agriculture system. Although some systems can collect environmental data, their data processing and analysis capabilities are limited. They cannot deeply explore the value behind the data and make scientific and reasonable regulatory decisions. In terms of visualization, the images generated by some systems are not real and intuitive enough, and cannot provide users with comprehensive and accurate farmland information. Moreover, most systems have deficiencies in the collaborative work of multiple modules, and lack effective integration between modules, resulting in poor stability and reliability of the entire planting control system. These problems have restricted the further development of smart agriculture and need to be solved urgently. Summary of the invention

[0007] The purpose of the present invention is to provide a visual intelligent agricultural planting control system to solve the problems raised in the above-mentioned background technology.

[0008] To achieve the above object, the present invention provides the following technical solution: a visual smart agricultural planting control system, the system comprising:

[0009] Environmental data acquisition module: used to obtain a multi-source environmental sensor data set; the multi-source environmental data includes soil moisture data, light intensity data, air temperature and humidity data, and crop growth image data;

[0010] Growth feature modeling module: based on the crop growth image data, the crop growth state characteristics are extracted through a morphological analysis algorithm, and the state characteristics include leaf expansion area, stem height growth rate and fruit maturity index;

[0011] Environmental control decision module: used to generate an environmental control instruction set based on the soil moisture data and the air temperature and humidity data through a dynamic threshold matching algorithm, wherein the instruction set includes irrigation water gradient, sunshade net opening and closing degree, and ventilation equipment start and stop parameters;

[0012] Visualization interaction module: multi-dimensional data fusion processing is performed on the light intensity data and the crop growth status characteristics to generate a three-dimensional farmland real scene simulation map; the multi-dimensional data fusion processing includes spatial coordinate mapping and light rendering synthesis;

[0013] Strategy execution module: integrates the environmental control instruction set with the three-dimensional farmland real-life simulation map through an adaptive control algorithm to generate a device execution action sequence and drive the agricultural facilities to execute.

[0014] Preferably, the morphological analysis algorithm includes:

[0015] Performing background noise filtering on the crop growth image data to extract the crop main body contour area;

[0016] Divide different organ growth regions based on a pixel clustering algorithm, and calculate the change rate of morphological parameters for each region;

[0017] Generate a maturity evaluation coefficient based on the change rate of morphological parameters and a preset growth stage threshold;

[0018] Combine the maturity evaluation coefficient with the stem height growth rate to form multi-dimensional growth state characteristics.

[0019] Preferably, the dynamic threshold matching algorithm includes:

[0020] Perform time window sampling on the soil moisture data and extract the humidity fluctuation trend for each time slice;

[0021] Use an interval covering algorithm to match the air temperature and humidity data with a preset suitable crop growth interval to generate a temperature and humidity deviation level;

[0022] Combine the humidity fluctuation trend and the temperature and humidity deviation level to perform instruction priority sorting and output an environmental control instruction set.

[0023] Preferably, the multi-dimensional data fusion processing includes:

[0024] According to the spatial distribution of the light intensity data, use a ray tracing algorithm to reconstruct a three-dimensional farmland light model;

[0025] Based on the spatial coordinate mapping relationship of the crop growth state characteristics, superimpose and render the three-dimensional light model and the crop morphological parameters;

[0026] Generate a three-dimensional farmland real scene simulation map by synthesizing different levels of visual elements through a transparency layering algorithm.

[0027] Preferably, the adaptive control algorithm includes: detecting action conflicts in the environmental control instruction set, screening device instructions that can be executed in parallel, dynamically adjusting the device response delay parameters according to the real-time update status of the three-dimensional farmland real scene simulation map; fusing the device response delay parameters with the action conflict detection results to generate a device execution action sequence.

[0028] Preferably, the optimization method of the pixel clustering algorithm includes:

[0029] Calculate the initial clustering center and the minimum pixel similarity threshold according to the historical image data distribution;

[0030] Traverse different clustering parameter combinations through a silhouette coefficient evaluation index and select the partitioning result with the highest silhouette coefficient;

[0031] Dynamically adjust the clustering center and the similarity threshold according to the partitioning result to optimize the calculation accuracy of the maturity evaluation coefficient.

[0032] Preferably, the implementation steps of the interval coverage algorithm include:

[0033] Perform moving average filtering on the air temperature and humidity data to generate a smooth environmental parameter curve;

[0034] Use the piecewise linear interpolation algorithm to divide the boundaries of the temperature and humidity suitable intervals, and calculate the distance weights between the current data points and the interval boundaries; optimize the deviation level division path through the dynamic programming algorithm to generate the temperature and humidity deviation levels.

[0035] Preferably, the construction method of the ray tracing algorithm includes:

[0036] Define the attenuation coefficient mapping relationship between the light intensity data and the three-dimensional space coordinates;

[0037] Generate a set of ray reflection paths based on the farmland terrain elevation data, and calculate the light contribution values of each area through the path integral algorithm;

[0038] Overlay and fuse the light contribution values with the crop morphological parameters to reconstruct the three-dimensional light model.

[0039] Preferably, the action conflict detection method includes: establishing a conflict rule library according to the physical constraint conditions of the device execution actions, traversing the execution paths of the current instruction set through the state machine model, identifying the instruction combinations that violate the conflict rules, and using the priority replacement algorithm to adjust the execution order of the conflict instructions to generate a conflict-free action sequence.

[0040] Preferably, it further includes:

[0041] Abnormal warning module: Generate an abnormal warning signal through the deviation degree accumulation algorithm according to the correlation analysis between the environmental regulation instruction set and the crop growth state characteristics; the correlation analysis includes the calculation of the dynamic correlation between the environmental parameters and the growth indicators, and triggers the warning level division when the accumulated deviation degree exceeds the preset threshold.

[0042] Compared with the prior art, the beneficial effects of the present invention are:

[0043] At the level of data collection and analysis, the environmental data collection module can obtain rich and comprehensive multi-source environmental sensing data, covering soil humidity, light intensity, air temperature and humidity, and crop growth image data, etc. This enables growers to comprehensively understand the farmland environmental information and provides a solid data foundation for subsequent precise analysis. The growth feature modeling module uses the morphological analysis algorithm to accurately extract key growth state characteristics such as the leaf expansion area, stem height growth rate, and fruit maturity index from the crop growth image data. Through in-depth analysis of these data, growers can grasp the crop growth dynamics in real time, accurately judge the crop growth stage, and thus take corresponding measures in a timely manner at the key growth nodes, greatly improving the scientificity and accuracy of planting management.

[0044] In terms of environmental regulation, the environmental regulation decision-making module generates a precise set of environmental regulation instructions based on soil moisture data and air temperature and humidity data using a dynamic threshold matching algorithm. This means that parameters such as the amount of irrigation water, the opening and closing degree of sunshade nets, and the start and stop parameters of ventilation equipment can be dynamically adjusted according to the actual environmental changes and crop requirements. For example, when the soil moisture is below the appropriate range, the system automatically calculates the reasonable amount of irrigation water, avoiding water waste and insufficient irrigation; according to the light intensity and the crop's light requirements, the system can intelligently control the opening and closing degree of sunshade nets to provide suitable light conditions for the crops; in terms of air temperature and humidity, the system can also start or stop the ventilation equipment in a timely manner to create a microclimate environment most conducive to crop growth, effectively improving the growth quality and yield of the crops.

[0045] The visual interaction module is of great significance. It performs multi-dimensional data fusion processing on light intensity data and crop growth state characteristics to generate a realistic three-dimensional virtual farmland simulation map. Through technical means such as spatial coordinate mapping and light rendering synthesis, growers can intuitively and vividly observe the real-time situation of the farmland as if they were on the spot. This visual display not only facilitates growers to understand the overall layout of the farmland and the growth trend of the crops, but also assists growers in decision-making analysis. For example, by observing the three-dimensional simulation map, growers can quickly identify areas with abnormal growth, promptly investigate the root causes of the problems, and take targeted solutions.

[0046] The strategy execution module uses an adaptive control algorithm to effectively integrate the set of environmental regulation instructions and the three-dimensional virtual farmland simulation map. Through action conflict detection, it screens out device instructions that can be executed in parallel, avoiding action conflicts between devices and improving the system operation efficiency. At the same time, it dynamically adjusts the device response delay parameters according to the real-time update status of the three-dimensional virtual farmland simulation map to ensure that the devices can execute the instructions in a timely and accurate manner, achieving precise control of agricultural facilities. This enables the entire planting control system to operate efficiently and stably, ensuring the smooth progress of agricultural production.

[0047] The abnormal warning module provides strong guarantee for agricultural production. Through the correlation analysis of the set of environmental regulation instructions and crop growth state characteristics, it generates abnormal warning signals using a deviation accumulation algorithm. When the dynamic correlation between environmental parameters and growth indicators is abnormal and the accumulated deviation exceeds the preset threshold, the system will promptly trigger the warning level division. Growers can quickly respond according to the warning information, take corresponding remedial measures, reduce crop losses caused by abnormal situations, and ensure the economic benefits and stability of agricultural production.

[0048] The visualization intelligent agricultural planting control system of the present invention realizes the full-process intelligent management from data collection, analysis, regulation to visualization display and abnormal warning through the collaborative work of each module, effectively improving the efficiency, quality and sustainable development ability of agricultural planting, providing strong technical support for the development of intelligent agriculture, and having important significance for promoting the process of agricultural modernization. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Figure 1 It is the working principle diagram of a visualization intelligent agricultural planting control system described in the present invention;

[0050] Figure 2 It is the working principle diagram of multi-dimensional data fusion processing;

[0051] Figure 3 It is the working principle diagram of the optimization method of pixel clustering algorithm;

[0052] Figure 4 It is the working principle diagram of the implementation steps of the interval coverage algorithm. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0054] Please refer to Figures 1 - 4 , the present invention provides a visualization intelligent agricultural planting control system, and its specific implementation method will be elaborated in detail below.

[0055] The system mainly consists of an environmental data collection module, a growth feature modeling module, an environmental regulation decision module, a visualization interaction module, and a strategy execution module. Each module works collaboratively to achieve precise control and visualization management of intelligent agricultural planting.

[0056] The environmental data acquisition module obtains a multi-source environmental sensing data set through various sensors. In farmland, soil moisture sensors are buried in the soil at different depths to collect soil moisture data in real time, which is used to reflect the moisture condition of the soil and provide a basis for subsequent irrigation decisions. The light intensity sensor is installed in a suitable position to accurately measure the light intensity data to understand the light conditions of the crop growth environment. The air temperature and humidity sensor monitors the temperature and humidity data of the air in real time, and these data are crucial for judging the comfort level of the crop growth environment. At the same time, crop growth image data is collected by cameras installed at different positions in the farmland. The cameras take pictures at preset time intervals to obtain continuous image information during the crop growth process.

[0057] The growth feature modeling module extracts crop growth state features based on crop growth image data using morphological analysis algorithms. This module first receives the crop growth image data transmitted by the environmental data acquisition module and then processes the images.

[0058] The environmental regulation decision module generates an environmental regulation instruction set through a dynamic threshold matching algorithm based on the soil moisture data and the air temperature and humidity data. This module obtains the soil moisture data and the air temperature and humidity data collected by the environmental data acquisition module in real time. After receiving the data, it samples the soil moisture data with a time window to analyze the change trend of the soil moisture in different time periods. By analyzing these data, the fluctuation situation of the soil moisture can be judged, such as whether it is gradually decreasing, remaining stable or suddenly rising, etc.

[0059] The visualization interaction module performs multi-dimensional data fusion processing on the light intensity data and the crop growth state features to generate a three-dimensional virtual farmland simulation map. This module first receives the light intensity data and the crop growth state feature data. For the light intensity data, according to its spatial distribution, a ray tracing algorithm is used to reconstruct the three-dimensional light model of the farmland. During the reconstruction process, the attenuation coefficient mapping relationship between the light intensity data and the three-dimensional space coordinates is defined. This relationship is used to describe the attenuation of light when propagating in the three-dimensional space. Based on the farmland terrain elevation data, a set of light reflection paths is generated, and the light contribution value of each area is calculated through the path integral algorithm. The light contribution value is superimposed and fused with the crop morphological parameters to reconstruct a three-dimensional light model that can truly reflect the light conditions of the farmland.

[0060] The policy execution module integrates the environmental regulation instruction set and the three-dimensional farmland real-scene simulation map through an adaptive control algorithm, generates a device execution action sequence, and drives the agricultural facilities to execute. This module receives the environmental regulation instruction set generated by the environmental regulation decision module and the three-dimensional farmland real-scene simulation map generated by the visual interaction module. First, it performs action conflict detection on the environmental regulation instruction set, establishes a conflict rule library based on the physical constraint conditions of the device execution actions. The conflict rule library contains various conflict situations that may occur when the devices execute actions and the corresponding processing rules.

[0061] The following further illustrates the technical solution of the present invention with specific embodiments.

[0062] Embodiment 1:

[0063] This embodiment focuses on elaborating the specific implementation process of the morphological analysis algorithm in the growth feature modeling module. The morphological analysis algorithm aims to accurately extract the crop growth state features based on the crop growth image data. In actual operation, first, the background noise of the crop growth image data is filtered to extract the crop main body contour area. The Gaussian filtering algorithm is used to process the image, and the Gaussian filtering formula is:

[0064] ;

[0065] where G(x, y) represents the value of the Gaussian filter at the coordinate (x, y), (x0, y0) is the center coordinate of the Gaussian kernel, is the standard deviation of the Gaussian kernel, which determines the smoothness of the filtering, the larger the value, the smoother the filtered image. Through Gaussian filtering, the noise in the image can be effectively removed, making the crop main body contour clearer.

[0066] Based on the pixel clustering algorithm, different organ growth regions are divided, and the change rate of the morphological parameters of each region is calculated. In this embodiment, the K-Means clustering algorithm is used, and its calculation formula is:

[0067] ;

[0068] where, is the sum of squared errors of clustering, n is the number of sample points, k is the number of clustering categories, c i is the i-th sample point, μ j is the j-th clustering center, ‖c i -μ j ‖ 2 represents the distance from the sample c i to the clustering center μ jThe square of the Euclidean distance. Through the K-Means clustering algorithm, the pixel points in the image are divided into different clusters, corresponding to different organ growth regions of the crop. For example, the leaf region, the stem region, and the fruit region are clustered separately. Then, morphological parameters such as the area and perimeter of each region are calculated, and by comparing consecutive images, the change rate of the morphological parameters is calculated, such as the growth rate of the leaf area and the change rate of the stem perimeter.

[0069] Generate a maturity evaluation coefficient based on the change rate of the morphological parameters and the preset growth stage threshold. Preset the morphological parameter thresholds for different growth stages. For example, during the flowering stage of the crop, the growth rate of the leaf area should be within a certain range, and when the fruit begins to swell, there are also corresponding thresholds for the change rate of the fruit region area. Compare the actually calculated change rate of the morphological parameters with these thresholds, and generate a maturity evaluation coefficient according to the comparison results. If the change rate of the morphological parameters is within the threshold range, the maturity evaluation coefficient increases; otherwise, it decreases.

[0070] Combine the maturity evaluation coefficient with the stem height growth rate to form a multi-dimensional growth state feature. The stem height growth rate is calculated by measuring and comparing the stem heights in images taken at different times. Considering the maturity evaluation coefficient and the stem height growth rate comprehensively can more comprehensively reflect the growth state of the crop. For example, when the maturity evaluation coefficient is high and the stem height growth rate is also within a reasonable range, it indicates that the crop is growing well; if the maturity evaluation coefficient is high but the stem height growth rate is extremely low, there may be growth problems and further analysis of the reasons is needed. Through the above steps, the accurate extraction of the growth state characteristics of the crop is achieved, providing an important basis for subsequent environmental regulation and management.

[0071] Example 2:

[0072] In this example, the dynamic threshold matching algorithm of the environmental regulation decision module is specifically implemented as follows.

[0073] Perform time window sampling on the soil moisture data. For example, set the time window to 1 hour and collect the soil moisture data every 10 minutes. In this way, 6 soil moisture data values will be obtained within 1 hour, denoted as S1, S2, S3, S4, S5, and S6 respectively. By calculating the differences between adjacent data, such as ΔS1 = S2 - S1, ΔS2 = S3 - S2, and so on, to judge the fluctuation trend of the humidity. If ΔS i is continuously positive, it indicates that the soil moisture is on the rise; if it is continuously negative, it means that the soil moisture is decreasing. For example, when ΔS1 = 2, ΔS2 = 1.5, and ΔS3 = 3, it can be judged that the current soil moisture is in an upward state.

[0074] Next, the interval coverage algorithm is used to match the air temperature and humidity data with the preset suitable interval for crop growth, and the temperature and humidity deviation level is generated. Suppose the preset suitable temperature interval for a certain crop growth is [20°C, 30°C], and the suitable humidity interval is [50%, 70%]. The air temperature T and air humidity H data are collected in real time.

[0075] First, perform a simple moving average process on the air temperature and humidity data. For example, take the average of the data at the last 3 moments. If the temperatures at the current moment and the previous two moments are T1, T2, and T3 respectively, then the temperature after moving average ; similarly for humidity, if the humidity data are H1, H2, and H3, then the humidity after moving average .

[0076] Judge the relationship between the temperature and humidity data after moving average and the suitable interval. If , then the temperature deviates from the suitable interval, and the deviation degree is recorded as ; if , the deviation degree is recorded as . The humidity judgment method is the same. If , the deviation degree is ; if , the deviation degree is . Comprehensively evaluate the temperature and humidity deviation level according to the deviation degrees of temperature and humidity. When both the temperature deviation degree and the humidity deviation degree are small, the temperature and humidity deviation level is low; if one or both of the deviation degrees are large, the temperature and humidity deviation level is correspondingly increased.

[0077] Combine the humidity fluctuation trend with the temperature and humidity deviation level to sort the instruction priorities, and output the environmental control instruction set. If the soil humidity shows a downward trend and the temperature and humidity deviation level is high, give priority to generating irrigation instructions to supplement the soil moisture and regulate the microclimate of the farmland; if the temperature and humidity deviation level is mainly due to too high temperature and the light intensity is large, then give priority to generating instructions for opening and closing the sunshade net and starting and stopping the ventilation equipment to reduce the temperature and improve the crop growth environment. In this way, according to the actual environmental data and the crop growth requirements, environmental control instructions are reasonably generated to ensure that the crops grow in a suitable environment.

[0078] Example 3:

[0079] This example details the process of multi-dimensional data fusion processing in the visual interaction module.

[0080] Reconstruct the three-dimensional farmland lighting model according to the spatial distribution of the light intensity data. The light intensity data are collected by sensors distributed at different positions in the farmland. Suppose multiple light sensors are set in the farmland, located at positions A, B, C, etc., and the light intensity values collected by each sensor are I A 、IB , I C , etc. Based on the position information of these sensors (such as coordinates (x A , y A , z A ), (x B , y B , z B ), (x A , y C , z C )) and the propagation characteristics of light in space, a three-dimensional light model is constructed. The light changes during propagation with the distance from the light source and the influence of obstacles. For example, the light intensity is relatively high in the area closer to the light source, and the light intensity will decrease where there are obstacles. By comprehensively considering these factors, the distribution of light in the three-dimensional space of the farmland is simulated.

[0081] Based on the spatial coordinate mapping relationship of crop growth state characteristics, the three-dimensional light model is superimposed and rendered with crop morphological parameters. Crop growth state characteristics include leaf expansion area, stem height, etc. Taking the leaf as an example, assume that the area of a certain leaf is B obtained through image processing technology, and the spatial position of this leaf in the farmland is determined (coordinates are (x leaf , y leaf , z leaf ). The spatial position information of the leaf is combined with the three-dimensional light model, and the leaf is rendered according to the light intensity. In the area with high light intensity, the color of the leaf rendering is brighter to indicate that the leaf in this area receives sufficient light; in the area with low light intensity, the color of the leaf is relatively darker. For the stem height, similarly, according to its spatial position and combined with the light model, the stem presents corresponding light and shadow effects under different light conditions. For example, on the side directly irradiated by the sun, there will be a more obvious bright part on the stem, while the shaded side is darker.

[0082] Synthesize visualization elements at different levels through the transparency layering algorithm to generate a three-dimensional virtual farmland simulation map. Divide the crop model, lighting model, and other elements representing soil, ridges, etc. into different levels. For example, place the crop model on the top layer because it is the focus of attention; the lighting model is in the middle layer to provide lighting effects; background elements such as soil and ridges are placed on the bottom layer. Set appropriate transparency for each level. The transparency of the crop model is set lower to clearly display the shape and growth status of the crops; the transparency of the background elements is set higher to avoid excessive interference with the display of the crops and lighting effects. By synthesizing these visualization elements at different levels in a certain order and transparency, a three-dimensional virtual farmland simulation map that can truly reflect the actual situation of the farmland is finally generated. By viewing this simulation map, users can intuitively understand the distribution of lighting, the growth status of the crops, and the relationships between them in the farmland, providing strong support for agricultural production decisions.

[0083] Example 4:

[0084] This example elaborates in detail the specific application of the adaptive control algorithm in the policy execution module. The adaptive control algorithm is mainly used to integrate the environmental regulation instruction set and the three-dimensional virtual farmland simulation map to generate a reasonable sequence of device execution actions.

[0085] Detect action conflicts in the environmental regulation instruction set and establish a conflict rule library based on the physical constraint conditions of device execution actions. In agricultural facilities, there may be conflicts in the execution actions of different devices. For example, when the irrigation device is working, the ventilation device may not be able to operate normally because ventilation will accelerate water evaporation and affect the irrigation effect. Establish a conflict rule library, such as stipulating that "when the irrigation device is turned on, the ventilation device shall not be started during irrigation".

[0086] Traverse the execution path of the current instruction set through the state machine model to identify instruction combinations that violate the conflict rules. The state machine model is transformed according to the execution order of the instructions and the state of the devices. Let the state of the state machine be , and the initial state is . When an instruction is received, perform a state transformation according to the instruction type and the current state. For example, when in the initial state and the irrigation device turn-on instruction is received, the state is transformed to (irrigation device on state). During the state transformation process, check whether there are instruction combinations that violate the conflict rules.

[0087] The priority replacement algorithm is adopted to adjust the execution order of conflicting instructions and generate a conflict-free action sequence. If conflicting instructions are detected, such as receiving irrigation equipment opening instructions and ventilation equipment opening instructions simultaneously, they are adjusted according to the preset priority rules. Assume that the priority of the irrigation instruction is higher than that of the ventilation instruction. When a conflict is detected, the irrigation instruction is executed first, and the ventilation instruction is delayed. The formula for the priority replacement algorithm is:

[0088] ;

[0089] where P i is the priority weight of the i-th instruction, W i is the priority coefficient of the i-th instruction, and m is the total number of the instruction set. By calculating the priority weight of each instruction, the execution order of the instructions is determined, thereby generating a conflict-free action sequence.

[0090] Dynamically adjust the device response delay parameters according to the real-time update status of the three-dimensional farmland real-scene simulation map. The three-dimensional farmland real-scene simulation map reflects the actual situation of the farmland in real time. When the light intensity in a certain area of the crop suddenly changes in the simulation map, it may be necessary to adjust the opening and closing degree of the sunshade net in a timely manner. According to this real-time change, the device response delay parameters are dynamically adjusted. For example, when the light intensity increases rapidly, shorten the response delay time of the sunshade net opening and closing device so that it can react faster and ensure the stability of the crop growth environment.

[0091] Fuse the device response delay parameters with the action conflict detection results to generate a device execution action sequence. Considering the conflict-free action sequence obtained from the action conflict detection and the device response delay parameters comprehensively, determine the specific execution time and action of each device. For example, first determine the opening time of the irrigation equipment, and then determine their execution times according to the priorities and response delay parameters of the sunshade net and ventilation equipment. Finally, generate a complete device execution action sequence to drive the agricultural facilities to execute the corresponding actions accurately and efficiently, realizing the precise regulation of the farmland environment.

[0092] Example 5:

[0093] This example focuses on describing the specific implementation process of the abnormal warning module. The abnormal warning module generates an abnormal warning signal through the deviation accumulation algorithm based on the correlation analysis between the environmental regulation instruction set and the crop growth state characteristics.

[0094] Perform dynamic correlation calculation of environmental parameters and growth indicators. Environmental parameters include soil humidity, air temperature and humidity, light intensity, etc., and growth indicators such as leaf expansion area, stem height growth rate, fruit maturity index, etc. The Pearson correlation coefficient is used to calculate the correlation between environmental parameters and growth indicators. The formula for the Pearson correlation coefficient is:

[0095]

[0096] where r is the Pearson correlation coefficient, a i and b i are the values of the environmental parameter and the growth index at the i-th sample point respectively, and are the average values of the environmental parameter and the growth index respectively, and d is the number of samples. By calculating the Pearson correlation coefficient, the degree of linear correlation between the environmental parameter and the growth index can be judged. For example, when the Pearson correlation coefficient between soil humidity and leaf expansion area is high and positive, it indicates that the increase in soil humidity is beneficial to leaf expansion.

[0097] According to the dynamic correlation calculation results, an abnormal warning signal is generated through the deviation accumulation algorithm. Let the actual value of the environmental parameter be , and the theoretical value corresponding to the growth index calculated according to the dynamic correlation be . The deviation calculation formula is:

[0098]

[0099] where D is the deviation. When the cumulative deviation exceeds the preset threshold, the warning level division is triggered, and t is the number of samples. For example, the preset threshold is T s , when , different warning levels are divided according to the size of the cumulative deviation. A smaller cumulative deviation may trigger a lower-level warning, indicating that attention needs to be paid to the relevant environmental parameters and crop growth status; a larger cumulative deviation triggers a higher-level warning, indicating that there may be problems seriously affecting crop growth and timely measures need to be taken for adjustment. Through the abnormal warning module, abnormal situations in the farmland environment and crop growth process can be detected in a timely manner, providing strong support for ensuring the normal growth of crops.

[0100] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "including", "comprising" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0101] Although embodiments of the present invention have been shown and described, those of ordinary skill in the art will appreciate that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A visual intelligent agricultural planting control system, characterized in that, Including: Environmental data acquisition module: used to obtain a multi-source environmental sensing data set; the multi-source environmental data includes soil humidity data, light intensity data, air temperature and humidity data, and crop growth image data; Growth characteristic modeling module: based on the crop growth image data, extract the crop growth state characteristics through a morphological analysis algorithm, and the state characteristics include leaf expansion area, stem height growth rate, and fruit maturity index; Environmental regulation decision-making module: used to generate an environmental regulation instruction set through a dynamic threshold matching algorithm according to the soil humidity data and air temperature and humidity data, and the instruction set includes irrigation water volume gradient, sunshade net opening and closing degree, and ventilation equipment start-stop parameters; Visualization interaction module: perform multi-dimensional data fusion processing on the light intensity data and crop growth state characteristics to generate a three-dimensional farmland real-scene simulation map; the multi-dimensional data fusion processing includes spatial coordinate mapping and light rendering synthesis; Strategy execution module: integrate the environmental regulation instruction set and the three-dimensional farmland real-scene simulation map through an adaptive control algorithm, generate a device execution action sequence, and drive agricultural facilities to execute.

2. The visualized intelligent agricultural planting control system according to claim 1, wherein The morphological analysis algorithm includes: Filter the background noise of the crop growth image data to extract the main crop contour area; Based on the pixel clustering algorithm, divide different organ growth areas and calculate the change rate of morphological parameters of each area; Generate a maturity evaluation coefficient according to the change rate of morphological parameters and a preset growth stage threshold; Combine the maturity evaluation coefficient with the stem height growth rate to form multi-dimensional growth state characteristics.

3. The visualized intelligent agricultural planting control system according to claim 1, characterized in that The dynamic threshold matching algorithm includes: Perform time window sampling on the soil humidity data to extract the humidity fluctuation trend of each time slice; Use the interval coverage algorithm to match the air temperature and humidity data with a preset suitable crop growth interval to generate a temperature and humidity deviation level; Combine the humidity fluctuation trend and the temperature and humidity deviation level to sort the instruction priorities and output an environmental regulation instruction set.

4. The visual intelligent agricultural planting control system according to claim 1, characterized in that, The multi-dimensional data fusion processing includes: According to the spatial distribution of the light intensity data, reconstruct the three-dimensional light model of the farmland using the ray tracing algorithm; Based on the spatial coordinate mapping relationship of the crop growth state characteristics, superimpose and render the three-dimensional light model and the crop morphological parameters; Synthesize different levels of visualization elements through the transparency layering algorithm to generate a three-dimensional farmland real-scene simulation map.

5. The visualization intelligent agricultural planting control system according to claim 1, wherein The adaptive control algorithm includes: detect action conflicts in the environmental regulation instruction set, screen device instructions that can be executed in parallel, and dynamically adjust the device response delay parameters according to the real-time update status of the three-dimensional farmland real-scene simulation map; fuse the device response delay parameters and the action conflict detection results to generate a device execution action sequence.

6. The visualization intelligent agricultural planting control system according to claim 2, characterized in that, The optimization method of the pixel clustering algorithm includes: Calculate the initial clustering center and the minimum pixel similarity threshold according to the historical image data distribution; Traverse different clustering parameter combinations through the silhouette coefficient evaluation index, and select the division result with the highest silhouette coefficient; Dynamically adjust the clustering center and similarity threshold according to the division result to optimize the calculation accuracy of the maturity evaluation coefficient.

7. The visualized intelligent agricultural planting control system according to claim 3, wherein The implementation steps of the interval coverage algorithm include: Perform sliding mean filtering on the air temperature and humidity data to generate a smooth environmental parameter curve; The piecewise linear interpolation algorithm is used to divide the boundaries of the suitable temperature and humidity intervals, and the distance weights between the current data points and the interval boundaries are calculated; the deviation level division path is optimized by the dynamic programming algorithm to generate the temperature and humidity deviation levels.

8. The visualized intelligent agricultural planting control system according to claim 4, wherein The construction method of the ray tracing algorithm includes: Defining the mapping relationship between the light intensity data and the attenuation coefficient of the three-dimensional space coordinates; Generating a set of ray reflection paths based on the elevation data of the farmland terrain, and calculating the light contribution values of each area through the path integral algorithm; Superposing and fusing the light contribution values with the crop morphological parameters to reconstruct the three-dimensional light model.

9. The visualization intelligent agricultural planting control system according to claim 5, characterized in that The action conflict detection method includes: establishing a conflict rule library according to the physical constraint conditions of the device execution actions, traversing the execution paths of the current instruction set through the state machine model, identifying the instruction combinations that violate the conflict rules, and using the priority replacement algorithm to adjust the execution order of the conflict instructions to generate a conflict-free action sequence.

10. The visualized intelligent agricultural planting control system according to claim 1, wherein It also includes: Anomaly warning module: generating an anomaly warning signal through the deviation accumulation algorithm according to the correlation analysis between the environmental regulation instruction set and the crop growth state characteristics; the correlation analysis includes the calculation of the dynamic correlation between the environmental parameters and the growth indicators, and triggering the warning level division when the accumulated deviation exceeds the preset threshold.

Images