Multi-variety fusion grafting control method of China rose, server, medium and product

Abstract

The invention provides a multi-variety fusion grafting control method for Chinese rose flowers, a server, a medium and a product, and relates to the technical field of Chinese rose fusion control application. The method comprises the following steps: firstly, acquiring a plurality of candidate grafting schemes, collecting rootstock monitoring data and to-be-scion Chinese rose plant information in combination with the schemes, inputting the rootstock monitoring data and the to-be-scion Chinese rose plant information into a growth feature matching prediction model formed by deep learning training according to historical grafting records, scion variety features, grafting results and expert experience, and performing scion grafting. Calculating the growth characteristic matching degree of the to-be-scion and the rootstock, and screening out candidate to-be-scion plant information according to the growth characteristic matching degree; combining the rootstock data and the candidate information to obtain a growth condition score through a fusion growth prediction model, and selecting a scheme with the highest score as an optimal scheme; and finally generating guidance data and sending the guidance data to a user side. By implementing the method, the Chinese rose growth characteristic matching degree can be accurately determined, so that the combination with high matching degree is screened out, the grafting success rate is greatly improved, and grafting failures caused by compatibility problems are reduced.

Description

Technical Field

[0001] The present application relates to the technical field of rose fusion control applications, and in particular to a method, system, medium and product for controlling multi-variety fusion grafting of roses. Background Art

[0002] Roses are beloved for their beautiful shapes and rich colors. To cultivate more ornamental rose varieties, gardeners often use grafting to combine different varieties. Multi-variety rose grafting not only allows for the production of diverse blooms on a single plant, but also allows for the integration of the best characteristics of multiple varieties.

[0003] Currently, rose grafting relies primarily on the experience of horticulturists. By observing the growth of the rootstock, horticulturists select areas with robust growth and thick branches as grafting sites. During grafting, horticulturists use their experience to determine the appropriateness of the scion and rootstock, manually making the incision and grafting. Post-grafting care primarily relies on manual observation of the plant's growth to adjust watering, fertilization, and other maintenance measures.

[0004] However, with the development of large-scale planting, rose grafting based on manual experience faces many challenges. Due to the lack of scientific data support, relying solely on empirical judgment often makes it difficult to accurately assess the matching degree of growth characteristics between scion and rootstock, resulting in unstable grafting survival rate. At the same time, manual observation makes it difficult to timely detect and predict growth interference between multiple varieties, which can easily cause some varieties to grow too strongly and inhibit the normal development of other varieties. Therefore, existing methods are difficult to efficiently achieve precise control of the grafting process of multiple varieties of roses, and thus it is difficult to improve the grafting survival rate and ensure the coordinated growth of multiple varieties. Summary of the Invention

[0005] The present application provides a multi-variety fusion grafting control method, server, medium and product for rose, which are used to achieve precise control of the multi-variety grafting process of rose, thereby improving the grafting survival rate.

[0006] In the first aspect, the present application provides a multi-variety fusion grafting control method for roses, which is applied to a server. The method includes: obtaining multiple candidate grafting schemes; obtaining rootstock monitoring data and information of rose plants to be grafted in combination with the candidate grafting schemes; inputting the rootstock monitoring data and the information of the rose plants to be grafted into a pre-trained growth feature matching prediction model to determine the matching degree of multiple growth characteristics of the rose plants to be grafted and the rootstock, and the growth feature matching degree is used to characterize the growth compatibility of the rose plants to be grafted and the rootstock. The growth feature matching prediction model is pre-trained based on the rootstock monitoring data, scion variety feature data, grafting result data in the historical grafting record database and the rose variety fusion rules in the expert experience knowledge base, and is trained using a deep learning method. The invention provides a method for determining a plurality of growth condition score information by combining the rootstock monitoring data and the information of the candidate scion rose plants with growth condition score information, and constructing the model through deep learning in advance based on a plurality of rootstock monitoring data and information sets of the scion rose plants annotated with growth condition score information; determining an optimal grafting scheme based on the candidate grafting scheme with the highest growth condition score information; generating multi-variety fused grafting guidance data including grafting operation steps, precautions and expected effects according to the optimal grafting scheme, and sending the multi-variety fused grafting guidance data to the user end for display, wherein the precautions at least include the environmental data required for grafting.

[0007] By employing this technical solution, the server first obtains multiple candidate grafting schemes, gathering a wide range of possible combinations and providing a foundation for subsequent, precise screening. These schemes are then combined to obtain rootstock monitoring data and information about the rose plants to be grafted, leveraging this multi-source information to comprehensively understand the grafting entity. Both are then fed into a growth characteristic matching prediction model, built based on historical grafting records and expert experience. This model deeply mines data associations and accurately determines the degree of growth characteristic matching, thereby selecting highly compatible combinations. This significantly improves the grafting success rate and reduces grafting failures due to compatibility issues.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of obtaining multiple candidate grafting schemes specifically includes: receiving candidate grafting schemes sent by the user terminal, each candidate grafting scheme recording a combination of different rose varieties to be grafted and rootstocks.

[0009] By adopting the above technical solution, candidate grafting solutions are received from the user end. This method of allowing users to participate in providing solutions can, on the one hand, make full use of user experience and needs. Different users propose various combinations based on their own practices or assumptions, enriching the candidate solution pool; on the other hand, the server subsequently evaluates and screens these solutions from users based on professional models, combining user wisdom with the server's professional analysis, which not only enhances user participation, but also ensures the scientific nature of selecting the optimal grafting solution.

[0010] In combination with some embodiments of the first aspect, in some embodiments, in combination with the candidate grafting scheme, the step of obtaining rootstock monitoring data and information of the rose plant to be grafted specifically includes: receiving rootstock monitoring data collected by multiple types of sensors set at different heights and orientations of the rootstock plant, the rootstock monitoring data including rootstock growth environment parameters and rootstock variety information that characterize the growth status of the rootstock; obtaining information of multiple rose plants to be grafted sent by the user end.

[0011] By adopting this technical solution, we receive rootstock monitoring data collected by multiple sensors, comprehensively covering the rootstock plant at different heights and orientations, accurately capturing the rootstock's growth environment parameters and variety information, ensuring the real-time status of the rootstock. At the same time, we obtain information about the rose plant to be grafted, sent by the user. By combining the two, we can comprehensively consider the actual growth conditions of the rootstock and the characteristics of the scion plant.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the step of determining the candidate rose plant information to be grafted based on the growth characteristic matching degree specifically includes: if the growth characteristic matching degree is greater than the set matching degree threshold, the corresponding rose plant information to be grafted is determined as the candidate rose plant information to be grafted.

[0013] By adopting the above technical solution, a growth characteristic matching threshold is set. Only when the matching degree exceeds this threshold is the corresponding rose plant information identified as a candidate for scion. This threshold acts as a precise filter, quickly eliminating scion plants with poor growth compatibility with the rootstock based on the matching degree calculated by the growth characteristic matching prediction model. This avoids wasting resources in the subsequent complex fusion growth prediction process, focusing on optimizing combinations with high matching degrees, making the entire grafting solution optimization process more efficient and directly focusing on the grafting combinations with the greatest potential for success.

[0014] In combination with some embodiments of the first aspect, in some embodiments, after the step of determining the optimal grafting scheme among the candidate grafting schemes with the highest growth condition score information, it also includes: obtaining grafting effect diagram information uploaded by the user; determining the number of scion rose plants and scion position information through a target detection algorithm; and determining the grafting operation steps based on the number of scion rose plants and scion position information.

[0015] By employing this technical solution, after obtaining user-uploaded grafting images, the system uses an object detection algorithm to determine the number of rose scion plants and their placement. This actual grafting feedback provides first-hand information for optimizing grafting procedures. Combined with the previously selected optimal grafting plan, detailed adjustments can be made to the actual grafting process, such as optimizing the incision angle and tying position based on the graft's position. This allows grafting guidance to be more closely aligned with on-site practice, further improving grafting success rates and ensuring optimal implementation of multi-variety fusion grafting.

[0016] In combination with some embodiments of the first aspect, in some embodiments, after generating multi-variety fusion grafting guidance data including grafting operation steps, precautions and expected effects according to the optimal grafting plan, and sending the multi-variety fusion grafting guidance data to the user terminal for display, it also includes: acquiring grafting video data in real time; determining the grafting operation score information through a grafting specification model in combination with the grafting video data, and the grafting specification model is constructed in advance based on multiple grafting video data sets annotated with grafting operation score information through deep learning; if the grafting operation score information is lower than the set grafting specification threshold, the user terminal issues a reminder.

[0017] By employing this technical solution, grafting video data is captured in real time and analyzed using a deep learning-based grafting standardization model to generate grafting operation scores. This allows for real-time monitoring of grafting compliance. If the grafting score falls below the set threshold, an alert is immediately sent to the user, allowing them to promptly correct any errors. This prevents the impact of improper grafting on survival rates, nipping issues in the bud, ensuring grafting quality throughout the entire process and ensuring a rigorous and orderly multi-variety fusion grafting process.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the step of sending a reminder to the user terminal if the grafting operation score information is lower than the set grafting specification threshold, it also includes: obtaining environmental data in real time through multiple sensors, and the environmental data includes at least temperature, humidity and light intensity data; combining the precautions and the environmental data, sending adjustment information to the user terminal.

[0019] By employing this technical solution, after alerting the user when irregular grafting procedures are detected, multiple sensors are used to collect real-time environmental data covering key factors such as temperature, humidity, and light intensity. This data is compared with the actual environmental data, and adjustment information is sent to the user. This ensures a suitable environment throughout the grafting process, allowing the scion and rootstock to grow and fuse under ideal conditions. This further consolidates the grafting results from an environmental perspective and reduces the negative impact of environmental factors on the grafting success rate.

[0020] In a second aspect, the present application provides a server comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory being used to store computer program code, the computer program code comprising computer instructions, the one or more processors calling the computer instructions to cause the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0021] In a third aspect, the present application provides a computer-readable storage medium comprising instructions, which, when executed on a server, cause the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, the present application provides a computer program product, which, when executed on a server, enables the server to execute the method described in the first aspect and any possible implementation of the first aspect.

[0023] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. Due to the adoption of a technical means of obtaining multiple candidate grafting schemes, collecting all-round information of rootstocks and plants to be grafted based on the schemes, inputting a growth characteristic matching prediction model that integrates historical data and expert experience to calculate the matching degree, and then screening the optimal scheme based on the matching degree and combining the integrated growth prediction model to generate guidance data, it effectively solves the technical problems in the existing technology of lack of scientificity and accuracy of grafting schemes and low success rate of grafting based on experience, thereby achieving the technical effect of improving the success rate of multi-variety fusion grafting of roses and ensuring the grafting effect.

[0024] 2. Due to the use of a technical means to collect all-round information on the growth environment and variety of rootstocks at different heights and orientations through multiple types of sensors, and combine it with the information of the plants to be grafted sent by the user end for subsequent accurate evaluation of the matching degree and growth status, it effectively solves the technical problems in the existing technology that lead to grafting decision-making errors and high failure rates due to one-sided information, and thus achieves the technical effect of improving the pertinence and scientific nature of the grafting plan and reducing the risk of grafting.

[0025] 3. Since the technical means of obtaining the grafting effect pictures uploaded by users after determining the optimal grafting plan, using the target detection algorithm to analyze the number and position of the scion plants, and optimizing the grafting operation steps accordingly is adopted, it effectively solves the technical problems in the existing technology that the grafting operation steps are divorced from reality and lack guidance, and thus achieves the technical effect of improving the accuracy and operability of the grafting operation and increasing the success rate of the grafting. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1This is a flow chart of a multi-variety fusion grafting control method for roses in an embodiment of the present application; Figure 2 This is another flow chart of the multi-variety fusion grafting control method of rose in the embodiment of the present application; Figure 3 This is a schematic diagram of the physical device structure of the server in an embodiment of the present application. DETAILED DESCRIPTION

[0027] The terms used in the following examples of the present application are for the purpose of describing specific embodiments only and are not intended to limit the present application. As used in the specification and appended claims of the present application, the singular expressions "a," "an," "said," "above," "the," and "this" are intended to include plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term "and / or" used in the present application refers to and encompasses any or all possible combinations of one or more of the listed items.

[0028] In the following, the terms "first" and "second" are used for descriptive purposes only and should not be understood to imply or suggest relative importance or implicitly indicate the number of the technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the embodiments of this application, unless otherwise specified, "plurality" means two or more.

[0029] For ease of understanding, the following describes the process of the method provided by this implementation. Figure 1 , is a flow chart of the multi-variety fusion grafting control method of rose in the embodiment of this application.

[0030] S101, obtaining multiple candidate grafting schemes; At the network communication level, the server has built a stable and secure communication interface to establish a real-time two-way communication channel with the user end. After the user selects different rose varieties and rootstock combinations to be grafted on the operation interface and clicks the Submit button, the user-end software will package these combination information in a specific data format and send it to the server.

[0031] To facilitate users in providing candidate solutions, a user-side application with rich functions and convenient operation has been developed. The design of the user-side application adopts the graphical user interface (GUI) design concept, providing an intuitive and easy-to-understand operation interface. In the rose variety selection area, by calling the rose variety data stored in the server database, all optional rose varieties to be grafted and rootstock varieties are presented in the form of a list or drop-down menu. Users can easily complete the combination selection with a simple check or click. At the same time, the application also has a parameter setting area, where users can set parameters such as the expected grafting time and the desired grafting effect (such as flower color matching, plant growth form, etc.) according to their own needs. In addition, to ensure the accuracy and validity of the data entered by the user, the application has integrated a data validation function. Before the user submits the solution, the program will automatically check whether the selected combination conforms to the basic logic, such as whether a variety combination unsuitable for grafting has been selected or whether key parameters have been omitted. If there is a problem, a prompt box will pop up in time to guide the user to make changes.

[0032] S102, based on the candidate grafting scheme, obtaining rootstock monitoring data and information of the rose plant to be grafted; During the multi-variety grafting process of roses, the rootstock's growth status plays a crucial role in the grafting success rate and subsequent plant growth. To comprehensively and accurately understand the rootstock's growth status, it is necessary to receive rootstock monitoring data collected by multiple sensors placed at different heights and orientations along the rootstock plant. To ensure comprehensive and accurate information on the rootstock's growth status, multiple sensors are installed at different heights and orientations along the rootstock plant. The placement of sensors at different heights reflects the fact that the rootstock's growth environment varies vertically. For example, soil moisture and temperature near the ground may differ from those at the top of the plant. Placing sensors at different orientations accounts for horizontal unevenness in factors such as light and ventilation. For example, the south-facing side of the plant may receive more sunlight, while the north-facing side may be relatively shaded. By deploying sensors at different heights and orientations, we can more comprehensively capture the changing rootstock environment.

[0033] Sensors include, but are not limited to, temperature sensors, which measure the temperature of the rootstock's surroundings. Temperature significantly impacts the rootstock's growth, physiological activity, and post-graft healing process. Different rootstock varieties have varying temperature adaptation ranges. A suitable temperature facilitates rootstock cell division and growth, improving the success rate of grafting. Light sensors can detect the intensity and duration of light received by the rootstock. Light is the energy source for photosynthesis in plants, and rootstocks at different growth stages have different light requirements. Reasonable light conditions help the rootstock accumulate nutrients and enhance its growth potential. Soil nutrient sensors can measure the content of key nutrients such as nitrogen, phosphorus, and potassium in the soil, as well as the soil's pH. The availability of soil nutrients directly impacts the growth and health of the rootstock.

[0034] Rootstock variety information can be obtained through tags pre-installed in sensors or by linking to a database. Different rootstock varieties have different biological characteristics, such as growth rate, disease resistance, and environmental adaptability. Understanding this rootstock variety information, combined with growth environment parameters, allows for a more accurate assessment of the rootstock's growth status, providing a scientific basis for subsequent grafting operations. For example, certain disease-resistant rootstock varieties may be more suitable as a grafting base under the same environmental conditions, while faster-growing rootstock varieties may require more nutrients and management after grafting. Multiple sensors installed on the rootstock plants transmit the collected rootstock monitoring data via wireless communication technology to a data aggregation node. The data aggregation node performs preliminary processing and integration, and then transmits the data to a server via wireless network. The server stores and analyzes this data, which can be used as historical data for subsequent growth trend analysis and comparison. During the analysis process, the server evaluates the collected data based on the growth characteristics of different rootstock varieties to determine whether the rootstock's growth status is normal and whether appropriate measures are needed to optimize the growing environment to improve the success rate and effectiveness of the grafting.

[0035] To obtain information about the rose plants to be grafted, the server receives relevant information sent by the user. The client application provides a detailed information entry interface, where users can enter information such as the variety name, growth stage (such as seedling stage, growing season, flowering period, etc.), health status (whether it is susceptible to pests and diseases, whether it is growing vigorously, etc.), plant height, stem thickness, etc. To facilitate user input, the application has set up intelligent prompts and auto-completion functions. When the user enters partial information, the system automatically associates and matches related options to reduce user input errors. After receiving the rose plant information sent by the user, the server will perform data verification and cleaning.

[0036] S103, inputting the rootstock monitoring data and the information of the rose plant to be grafted into a pre-trained growth characteristic matching prediction model to determine the matching degree of multiple growth characteristics of the rose plant to be grafted and the rootstock, the growth characteristic matching degree is used to characterize the growth compatibility of the rose to be grafted and the rootstock, and the growth characteristic matching prediction model is previously established by training using a deep learning method based on the rootstock monitoring data, scion variety characteristic data, grafting result data in the historical grafting record database, and rose variety fusion rules in the expert experience knowledge base; After obtaining the rootstock monitoring data and the information of the rose plant to be grafted, the server inputs this data into a pre-trained growth feature matching prediction model to determine the degree of matching of the growth characteristics of the two. The growth feature matching prediction model is a complex model built based on deep learning. Its training data comes from the rootstock monitoring data, scion variety characteristic data, grafting result data in the historical grafting record database, and the rose variety fusion rules in the expert experience knowledge base. In the early stage of model training, the server will perform in-depth preprocessing on these data. For numerical data, such as temperature, humidity, plant height, etc., standardization or normalization methods are used to map them to specific intervals to eliminate the impact of dimensions and ensure the stability and accuracy of model training. For categorical data, such as rose variety names, unique hot encoding technology is used to convert them into vector form that is easy for computers to process.

[0037] The server uses a neural network architecture, such as a multi-layer perceptron (MLP) or convolutional neural network (CNN), to build a growth feature matching prediction model. For example, the MLP consists of an input layer, multiple hidden layers, and an output layer. The input layer receives preprocessed rootstock monitoring data and information about the rose to be grafted. The hidden layer uses nonlinear activation functions to extract and transform features from the input data, mining the complex underlying relationships between the data. The output layer outputs multiple growth feature matching values, characterizing the growth compatibility between the rose to be grafted and the rootstock from multiple dimensions, such as physiological and metabolic compatibility, growth rate coordination, and synergistic pest and disease resistance.

[0038] During model training, the server uses a backpropagation algorithm to adjust model parameters. Through continuous iterative training, the error between predicted and actual grafting results is gradually reduced. During training, a large amount of historical data is divided into training, validation, and test sets. The model is trained on the training set, and its performance is evaluated on the validation set. Hyperparameters such as the number of hidden layers and neurons are adjusted based on the validation results. When the model achieves high performance metrics such as accuracy and recall on the test set, model training is considered complete and ready for practical application.

[0039] When the server receives new rootstock monitoring data and information about the rose plants to be grafted, it first performs the same preprocessing as the training data to ensure it meets the model's input requirements. The preprocessed data is then fed into the trained growth feature matching prediction model, which analyzes and predicts the input data, outputting multiple growth feature matching degrees.

[0040] S104, determining information of candidate rose plants to be grafted based on the growth characteristic matching degree; A reasonable matching threshold is pre-set within the server. This threshold is not set in stone, but rather is derived from in-depth analysis of a large amount of historical grafting data and comprehensive analysis of expert experience. The server statistically analyzes indicators such as grafting survival rate and plant growth status at different matching levels in historical grafting cases, constructs a data model, and simulates the screening effects under different threshold settings. At the same time, referring to the practical experience and professional opinions of experts in the rose cultivation field, the strictness of the screening is weighed against the diversity of grafting schemes. Ultimately, a threshold is determined that ensures that the selected scions and rootstocks have a high degree of growth compatibility while ensuring a sufficient number of candidate schemes for further evaluation.

[0041] After the server obtains the growth characteristic matching data, it uses an efficient data processing algorithm to traverse the matching information of all rose plants to be grafted. Taking parallel computing technology as an example, the server can divide the matching data into multiple data blocks and assign them to multiple processor cores for simultaneous processing, greatly improving the screening efficiency. For each rose plant to be grafted, the server compares its growth characteristic matching with the set threshold. If the matching degree is greater than the threshold, the server will extract the corresponding information of the rose plant to be grafted and determine it as the candidate rose plant information to be grafted. This information includes but is not limited to key data such as the variety name, growth stage, health status, and plant morphological characteristics of the rose to be grafted.

[0042] To facilitate subsequent processing and management, the server stores the information of the confirmed candidate rose plants for grafting in a dedicated database table. During the storage process, the server adds a unique identifier to each piece of information for quick query and retrieval. At the same time, the server also associates and stores the numerical value of the growth characteristic matching between the candidate plant and the rootstock, facilitating subsequent sorting, comparison, and analysis. In addition, the server will also preliminarily organize and annotate the candidate information, such as grading it according to the degree of matching, to facilitate subsequent fusion growth prediction and optimal solution screening.

[0043] After the screening is completed, the server will conduct a quality assessment and quantity check on the candidate scion rose plant information, and by checking the completeness and accuracy of the candidate information, ensure that the selected candidate plant information has no missing or incorrect data. If the number of candidate information is found to be too small, it may mean that the set threshold is too high. The server will appropriately lower the threshold and re-screen; if the number of candidate information is too large, the server will further refine the screening conditions and combine other auxiliary indicators, such as the market demand and ornamental value of the scion rose plants, to conduct a second screening of the candidate information to ensure that the final candidate scion rose plant information has a high degree of growth characteristic matching and can meet actual production and ornamental needs.

[0044] S105, combining the rootstock monitoring data and the candidate scion rose plant information, determining multiple growth status score information through a fusion growth prediction model, wherein the fusion growth prediction model is constructed in advance based on multiple rootstock monitoring data and scion rose plant information sets annotated with growth status score information through deep learning; After determining the candidate rose scion candidates, the server combines this information with rootstock monitoring data and uses a fusion growth prediction model to calculate multiple growth status scores, thereby comprehensively assessing the growth potential of different grafting combinations. This fusion growth prediction model is built using deep learning technology. Its training relies on a large amount of rose rootstock monitoring data and scion rose plant information sets annotated with growth status scores. The server preprocesses this training data during the initial model construction.

[0045] The server uses recurrent neural networks (RNNs) and their variants, such as long short-term memory (LSTM) networks, to build an integrated growth prediction model. Taking LSTM as an example, it effectively processes time series data and captures the long-term dependencies between rose rootstocks and scions during growth. The model's input layer receives preprocessed rootstock monitoring data and information about candidate scion rose plants. The hidden layer uses LSTM units to extract features from the input data and conduct deep interactive learning, exploring the complex internal connections between the data. The output layer outputs a corresponding growth status score, which comprehensively reflects key growth indicators such as the grafted rose plant's survival rate, growth rate (new branch length and leaf number), flower quantity and quality (flower size, color vividness, number of petal layers), and pest and disease resistance (resistance to common rose pests and diseases such as aphids and spider mites). During model training, the server uses a backpropagation algorithm combined with an adaptive learning rate optimization algorithm to adjust model parameters. Through continuous iterative training, the error between the predicted and actual growth status scores is reduced. During training, the large amount of training data is divided into training, validation, and test sets. The model is trained on the training set, and its performance is evaluated on the validation set. The model's hyperparameters are adjusted based on the validation results. Model training is considered complete when the model achieves high performance indicators such as accuracy and mean squared error on the test set.

[0046] When the server wants to determine the growth status score information, it first performs the same preprocessing operations on the rootstock monitoring data to be processed and the information of the candidate rose plants to be grafted as the training data, so that they meet the input requirements of the fusion growth prediction model. The preprocessed data is then input into the trained fusion growth prediction model. The model analyzes and predicts the input data based on the learned rules and characteristics, and outputs multiple growth status score information. The server will analyze and process the growth status score information output by the model, calculate the statistical characteristics of the growth status scores of different candidate grafting schemes, such as the mean and standard deviation, to intuitively understand the overall performance and stability of each scheme. At the same time, the growth status score information is compared and analyzed with historical data to determine whether the current prediction results are reasonable. If an abnormal score is found, the server will re-check the data input and model operation process, and fine-tune or retrain the model if necessary to ensure the accuracy and reliability of the growth status score information.

[0047] S106, determining the optimal grafting scheme based on the candidate grafting scheme with the highest growth status score information; After obtaining the growth status scores for each candidate grafting solution, the server compares these scores and determines the solution with the highest score as the optimal grafting solution. The server uses efficient data sorting algorithms, such as quick sort and heap sort, to sort the growth status scores of all candidate grafting solutions. These algorithms can quickly sort the scores from high to low, facilitating the rapid identification of the highest-scoring solution. During sorting, the server not only ranks the scores but also simultaneously adjusts the corresponding candidate grafting solution information to ensure an accurate correspondence between scores and solutions.

[0048] After the sorting is completed, the server selects the candidate grafting scheme with the highest growth score as the optimal grafting scheme. To ensure the accuracy and reliability of the selection, the server will conduct multiple verifications on the highest-scoring scheme. On the one hand, it checks whether the data of the scheme is complete and accurate to avoid incorrect selection due to data errors. On the other hand, it refers to the historical data and expert experience of rose grafting to judge the feasibility and potential risks of the scheme in actual production. For example, if the combination of scion and rootstock in the highest-scoring scheme has a low historical grafting success rate, or experts believe that the combination may have problems with growth coordination, the server will further analyze the reasons and re-evaluate other candidate schemes if necessary.

[0049] After determining the optimal grafting plan, the server will record and back up the plan in detail. All relevant information about the plan, including rootstock variety (such as the commonly used wild rose rootstock, which has strong adaptability and a well-developed root system), rootstock growth environment parameters, rose variety to be grafted (such as ornamental varieties such as "Pink Peace" and "Feishan"), growth characteristic matching degree, growth status score, etc., will be stored in a dedicated database table.

[0050] In some embodiments, after determining the optimal grafting scheme, the grafting effect diagram information uploaded by the user is obtained, the relevant information is determined through the target detection algorithm, and the grafting operation steps are determined accordingly. This series of processes is a further optimization of the grafting scheme, which makes the grafting operation more practical and helps to improve the success rate and effect of grafting. Specifically, the server provides users with a special upload interface that supports multiple image formats, such as JPEG, PNG, etc. Users can take photos of the grafted plants through mobile phones, cameras and other devices, and then upload them to the server through the user-side application or web platform. During the upload process, the system will perform preliminary processing on the pictures, compress the picture size to increase the upload speed, and ensure that the picture quality does not affect subsequent analysis. After receiving the pictures, the server will classify and store the pictures, and archive them according to information such as user identification and grafting time to facilitate subsequent query and use.

[0051] The server uses advanced object detection algorithms, trained on numerous rose plant images, to accurately identify the scion rose plants in the images. The server feeds the user's uploaded grafting image into the object detection algorithm model. The model analyzes the pixels in the image to identify the outline and features of the scion rose plant, thereby determining the number of scion rose plants. Furthermore, the server calculates the scion's coordinate position in the image to determine its location. The server adjusts resource allocation for subsequent grafting operations based on the number of scion rose plants. A large number of grafting plants may require more grafting tools and materials, leading to more time for the grafting process. A smaller number may require adjustments to the plan. Regarding scion location information, the server analyzes the distribution of the scion on the rootstock. If the scion distribution is uneven, the server recommends adjusting the scion's position during subsequent grafting to ensure balanced and aesthetically pleasing plant growth. For example, if the scion is too dense on one side, the server will guide the user to select an appropriate grafting location on the other side. The server will also optimize the specific grafting operation details based on the scion position, such as determining the angle and depth of the incision according to the scion position, so that the scion and rootstock fit better and improve the grafting survival rate.

[0052] S107: Generate multi-variety fusion grafting guidance data including grafting operation steps, precautions and expected effects according to the optimal grafting plan, and send the multi-variety fusion grafting guidance data to the user terminal for display. The precautions at least include environmental data required for grafting.

[0053] The server generates detailed grafting operation steps based on the rootstock and scion information in the optimal grafting plan, combined with professional knowledge and practical experience in rose grafting. Taking "branch grafting" as an example, in terms of the operation steps, the server will guide users to choose the appropriate grafting time, which is generally more appropriate before the rose buds in spring. When selecting a rootstock, choose branches that are strong and free of pests and diseases, cut them at 10-15 cm from the ground, and flatten the cross section. For the scion, choose a one-year-old branch with 2-3 full buds, cut it into a wedge shape at the bottom of the scion, and the cut surface is 2-3 cm long. Insert the scion into the incision of the rootstock, align the cambium of the two, and tie them tightly with plastic film strips. The interface can be coated with grafting wax or vaseline to prevent moisture evaporation. These operation steps are presented in clear and easy-to-understand text descriptions, and are accompanied by intuitive diagrams or animations, and displayed to users through the user-side application.

[0054] Regarding precautions, the server comprehensively considers various factors during the grafting process and generates comprehensive prompts. In addition to the environmental data required for grafting, such as maintaining a temperature of 18-22 degrees Celsius, a relative humidity of 70%-80%, and avoiding direct sunlight, users are also reminded to disinfect grafting tools to prevent bacterial infection, as roses are susceptible to pathogens and can affect grafting survival rates. During the post-grafting maintenance phase, users are reminded to water the roses appropriately, keeping the soil moist but avoiding standing water, as this can easily lead to root rot. Regarding fertilization, adhere to the principle of "thin fertilizer, frequent application," applying a thin liquid fertilizer once a month during the rose's peak growing season. Also, regularly inspect the grafting site for healing. If the scion is observed to be wilting or infested with pests or diseases, take appropriate measures, such as spraying pesticides for pest control. The server organizes these precautions by importance and chronological order, presenting them to users in a list for easy review and compliance.

[0055] In the embodiment of the present application, due to the adoption of a technical means of obtaining multiple candidate grafting schemes, collecting all-round information of the rootstock and the plant to be grafted based on the schemes, inputting a growth characteristic matching prediction model that integrates historical data and expert experience to calculate the matching degree, and then screening the optimal scheme based on the matching degree and combining the integrated growth prediction model to generate guidance data, the feasibility of the grafting combination can be comprehensively and accurately evaluated, effectively solving the technical problems in the existing technology of the lack of scientificity and accuracy of the grafting schemes and the low success rate of grafting based on experience, thereby achieving the technical effect of improving the success rate of multi-variety fusion grafting of roses and ensuring the grafting effect.

[0056] After combining the above content, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the multi-variety fusion grafting control method of rose in the embodiment of this application.

[0057] S201, real-time acquisition of grafted video data; After generating and sending guidance data based on the optimal grafting solution, the server initiates the process of acquiring grafting video data in real time to monitor the grafting operation. The server establishes a stable connection with the cameras installed at the grafting site. These cameras are located in key locations that can fully capture the grafting operation, such as filming the grafting personnel's operating area from different angles to ensure that every operational detail is fully recorded. The connection utilizes wireless communication technologies such as Wi-Fi and 4G / 5G to ensure that video data can be transmitted to the server stably and quickly. A dedicated video acquisition program is set up on the server to continuously monitor the video signals sent by the cameras. At the same time, the video data is marked with a timestamp to accurately record the acquisition time of each frame of data, facilitating subsequent chronological analysis of the grafting operation process.

[0058] S202, combining the grafting video data, and determining grafting operation score information through a grafting specification model, where the grafting specification model is constructed in advance based on multiple grafting video datasets annotated with grafting operation score information through deep learning; After acquiring the grafting video data, the server analyzes it using a grafting specification model to determine the grafting operation score. The grafting specification model is built based on deep learning, and its training relies on a large dataset of grafting videos annotated with grafting operation scores.

[0059] During the model training phase, the server preprocesses the grafting videos in the training dataset. Video image enhancement technology is used to improve the clarity and contrast of the videos, enabling the model to better identify operational details. A video segmentation algorithm breaks the videos down into frames, facilitating frame-by-frame analysis by the model. For each frame, an object detection algorithm, such as one based on a convolutional neural network (CNN), identifies key elements in the image, such as grafting tools (such as grafting knives and binding tape), the positions of the rootstock and scion, and the grafter's hand movements. These elements are then annotated and, based on pre-defined grafting operation specifications, each frame is assigned a score, which serves as a label for model training. The server uses a convolutional neural network (CNN) architecture to build the grafting specification model and adjusts and optimizes it based on actual needs. These models have powerful image feature extraction capabilities and can automatically learn complex features in images. The model's input layer receives preprocessed video frame data. The hidden layer, through components such as convolutional layers, pooling layers, and fully connected layers, performs multiple feature extraction and transformations on the input data to uncover the operational information contained in the image. The output layer outputs the corresponding grafting operation score information. The score range can be set to 0-100 points. The higher the score, the more standardized the operation.

[0060] To determine the grafting operation score, the server splits and preprocesses the real-time grafting video data frame by frame to ensure it meets the input requirements of the grafting standardization model. The preprocessed video frame data is then fed into the trained grafting standardization model frame by frame. The model analyzes and predicts each frame and outputs a corresponding score. The server integrates and analyzes the scores output by the model, for example, calculating the average score over a period of time (e.g., one minute) as the overall score for the grafting operation during that period. This allows the server to accurately assess the grafting operation's compliance in real time, providing a basis for timely detection and correction of irregularities.

[0061] S203: If the grafting operation score information is lower than the set grafting standard threshold, the user terminal sends a reminder; After determining the grafting operation score using the grafting standard model, the server compares the score with a pre-set grafting standard threshold. If the grafting operation score falls below the threshold, the server immediately sends a reminder to the user, allowing them to promptly correct any irregularities and ensure grafting quality.

[0062] A reasonable grafting threshold is pre-set within the server. This threshold is not fixed but is dynamically adjusted based on analysis of a large amount of historical grafting video data and the experience of rose grafting experts. The server analyzes indicators such as the actual grafting survival rate and plant growth status corresponding to grafting operations in different score ranges. Combined with industry standards and best practices for rose grafting, it determines a threshold that ensures the quality of grafting operations while adapting to different grafting scenarios. For example, after extensive experiments and data analysis, it was found that when the operation score reaches 80 points, the grafting survival rate is high and the plant growth is good. Therefore, 80 points can be set as the initial threshold.

[0063] When the server determines that the grafting operation score is below the set threshold, it activates the reminder mechanism. The server sends a reminder message to the user through network communication technology. The content of the reminder message includes specific information about the non-standard operation, such as "Using the grafting knife at an improper angle may affect the fit between the scion and the rootstock" and "Binding too loosely may cause the scion to loosen." At the same time, to allow users to more intuitively understand the problem, the server can attach a screenshot or short video of the non-standard operation to the reminder message, clearly indicating the time and specific manifestations of the problem.

[0064] On the user side, a dedicated reminder receiving and display program was developed. Upon receiving a reminder message from the server, the program would alert the user through various methods, including pop-up windows, vibration, and sound. The pop-up window would display detailed information about the irregular operation and improvement suggestions for easy viewing by the user.

[0065] S204, acquiring environmental data in real time through multiple sensors, the environmental data including at least temperature, humidity, and light intensity data; After real-time monitoring of the grafting operation and issuing reminders based on the operation score, the server needs to further acquire environmental data to provide users with more comprehensive grafting guidance and ensure that grafting is carried out in an appropriate environment. The server first establishes a stable connection with multiple sensors distributed throughout the grafting site. These sensors collect key environmental data such as temperature, humidity, and light intensity. The temperature sensor can be a high-precision thermistor sensor, which accurately measures ambient temperature within the ideal temperature range for rose growth, such as 5°C-35°C, with an accuracy of ±0.5°C. The humidity sensor uses a capacitive or resistive humidity sensor, which accurately measures relative humidity within a range of 20%-95% RH with an accuracy of ±3% RH. The light intensity sensor can use a silicon photocell or photodiode, measuring light intensity from 0 to 20,000 lux, meeting the light intensity monitoring needs of roses at different growth stages.

[0066] The sensors transmit collected data to the aggregation node using low-power wireless communication technologies. These wireless communication technologies, with their low power consumption and long-distance transmission capabilities, are suitable for use in outdoor planting environments. They can effectively reduce sensor energy consumption and extend their service life. The aggregation node performs preliminary processing on the received sensor data and transmits the processed data to the server via the network.

[0067] S205: Send adjustment information to the user terminal based on the precautions and environmental data.

[0068] After obtaining real-time environmental data, the server compares and analyzes it with the environmental requirements in the grafting precautions previously provided to the user, and then sends adjustment information to the user to ensure that the grafting environment is always suitable. The server first reads the precautions related to the current grafting task from the database, which contains the environmental data range suitable for grafting. For example, the suitable temperature range for grafting is set at 18℃-22℃, the relative humidity range is 70%-80%, and the light intensity range varies depending on the rose variety and growth stage, generally between 5000lux-15000lux. Then, the real-time environmental data is compared with these standard ranges to generate adjustment information and send it to the user.

[0069] For example, when the temperature data exceeds the appropriate range, the server will generate corresponding adjustment information based on the actual situation. If the temperature is too high, exceeding 22°C, the server will prompt the user to take cooling measures, such as building a sunshade, using a spray device to increase air humidity to lower the temperature, etc., and inform the user that excessively high temperatures may cause the scion's moisture to evaporate too quickly, affecting the grafting survival rate. If the temperature is too low, below 18°C, the server will remind the user to take warming measures, such as covering with insulation film, placing heating equipment around, etc., and explain that low temperatures may slow down wound healing and be detrimental to the growth of grafted plants.

[0070] On the user side, the application will display these adjustment information in the form of pop-up windows, message push, etc. The pop-up window details the difference between the current environmental data and the suitable range, as well as specific adjustment suggestions and reasons, to facilitate user understanding and operation. At the same time, the user-side application will also provide operational guidance, such as linking to the control interface of related devices (if there is a smart device), to help users quickly adjust the environment. In addition, the server will continuously monitor changes in environmental data. If the environmental data still does not reach the suitable range after the user's adjustment, the server will send adjustment information again until the environmental data meets the requirements, thereby providing a stable and suitable environmental guarantee for the fusion grafting of multiple varieties of roses and improving the success rate of grafting.

[0071] In the embodiment of the present application, due to the use of real-time acquisition of grafting video data and evaluation of operation scores through a grafting specification model, a user is reminded when the score is lower than a threshold, and environmental data is acquired in real time and adjustment information is sent in combination with precautions. Therefore, the grafting process can be monitored in real time and comprehensively, effectively solving the technical problems in the prior art of lack of standardized supervision of grafting operations and difficulty in controlling environmental factors, resulting in a high grafting failure rate, thereby achieving the technical effect of improving the standardization of grafting operations, ensuring the suitability of the grafting environment, and further improving the success rate of multi-variety fusion grafting of roses.

[0072] The following describes the server in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , is a schematic diagram of a physical device structure of a server in an embodiment of the present application.

[0073] It should be noted that Figure 3 The structure of the server shown is only an example and should not limit the functions and scope of use of the embodiments of the present invention.

[0074] like Figure 3 As shown, the server includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes, such as the methods described in the above embodiments, based on programs stored in a read-only memory (ROM) 302 or programs loaded from a storage unit 308 into a random access memory (RAM) 303. RAM 303 also stores various programs and data required for system operation. CPU 301, ROM 302, and RAM 303 are interconnected via a bus 304. An input / output (I / O) interface 305 is also connected to bus 304.

[0075] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, push button switches, and the like; an output section 307 including a liquid crystal display (LCD), an audio output device, indicator lights, and the like; a storage section 308 including a hard disk and the like; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. Removable media 311, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 310 as needed, so that computer programs read from the removable media can be installed in the storage section 308 as needed.

[0076] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for executing the methods illustrated in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication section 309 and / or installed from removable media 311. When executed by the central processing unit (CPU) 301, the computer program performs the various functions defined in the present invention.

[0077] It should be noted that specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0078] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings.

[0079] Specifically, the server of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the multi-variety fusion grafting control method of rose provided in the above embodiment is implemented.

[0080] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the server described in the above embodiments, or may exist independently and not incorporated into the server. The storage medium carries one or more computer programs, which, when executed by a processor of the server, enable the server to implement the rose multi-variety fusion grafting control method provided in the above embodiments.

[0081] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0082] As used in the above embodiments, the term “when” may be interpreted to mean “if” or “after” or “in response to determining that” or “in response to detecting that”, depending on the context. Similarly, the phrases “upon determining that” or “if (stated condition or event) is detected” may be interpreted to mean “if determining that” or “in response to determining that” or “upon detecting (stated condition or event)” or “in response to detecting (stated condition or event)”, depending on the context.

[0083] Those skilled in the art will appreciate that all or part of the process steps in the above-described method embodiments can be implemented by a computer program instructing the relevant hardware. The program can be stored in a computer-readable storage medium, and when executed, the program can include the process steps in the above-described method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A multi-variety fusion grafting control method for roses, applied to a server, characterized in that: The method comprises: Obtain multiple candidate grafting solutions; In combination with the candidate grafting scheme, obtaining rootstock monitoring data and information of the rose plant to be grafted; Inputting the rootstock monitoring data and the information of the rose plant to be grafted into a pre-trained growth characteristic matching prediction model to determine the matching degree of multiple growth characteristics of the rose plant to be grafted and the rootstock, wherein the growth characteristic matching degree is used to characterize the growth compatibility of the rose plant to be grafted and the rootstock, and the growth characteristic matching prediction model is previously established by training using a deep learning method based on the rootstock monitoring data, scion variety characteristic data, grafting result data in a historical grafting record database, and rose variety fusion rules in an expert experience knowledge base; Determining information of candidate rose plants to be grafted based on the growth characteristic matching degree; Combining the rootstock monitoring data and the candidate scion rose plant information, determining multiple growth status score information through a fusion growth prediction model, wherein the fusion growth prediction model is constructed in advance based on multiple rootstock monitoring data and scion rose plant information sets annotated with growth status score information through deep learning; Determine the optimal grafting scheme based on the candidate grafting scheme with the highest growth status score information; According to the optimal grafting scheme, multi-variety fusion grafting guidance data including grafting operation steps, precautions and expected effects is generated, and the multi-variety fusion grafting guidance data is sent to the user end for display, wherein the precautions at least include environmental data required for grafting.

2. The method according to claim 1, characterized in that The steps of obtaining multiple candidate grafting solutions specifically include: The candidate grafting schemes sent by the user end are received, each of which records the combination of different rose varieties to be grafted and rootstocks.

3. The method according to claim 1, characterized in that In combination with the candidate grafting scheme, the steps of obtaining rootstock monitoring data and information of the rose plant to be grafted include: Receiving rootstock monitoring data collected by multiple types of sensors set at different heights and orientations of the rootstock plant, the rootstock monitoring data including rootstock growth environment parameters and rootstock variety information characterizing the growth status of the rootstock; Get the information of multiple rose plants to be grafted sent by the user.

4. The method according to claim 1, wherein The step of determining the information of candidate rose plants to be grafted based on the growth characteristic matching degree specifically includes: If the growth characteristic matching degree is greater than the set matching degree threshold, the corresponding rose plant information to be grafted is determined as candidate rose plant information to be grafted.

5. The method according to claim 1, wherein After determining the optimal grafting scheme from the candidate grafting scheme with the highest growth status score information, the method further includes: Get the grafting effect picture information uploaded by the user; The number of scion rose plants and scion location information are determined through target detection algorithms; The grafting operation steps are determined based on the number of scion rose plants and the scion position information.

6. The method according to claim 1, wherein After generating multi-variety fusion grafting guidance data including grafting operation steps, precautions, and expected effects according to the optimal grafting plan, and sending the multi-variety fusion grafting guidance data to the user terminal for display, the method further includes: Obtain grafted video data in real time; Determining grafting operation score information using a grafting specification model based on the grafting video data, wherein the grafting specification model is constructed by deep learning based on a plurality of grafting video datasets annotated with grafting operation score information; If the grafting operation score information is lower than the set grafting standard threshold, the user terminal sends a reminder.

7. The method according to claim 6, characterized in that After the step of sending a reminder to the user terminal if the grafting operation score information is lower than the set grafting standard threshold, the method further includes: Acquire environmental data in real time through multiple sensors, the environmental data including at least temperature, humidity and light intensity data; The adjustment information is sent to the user terminal in combination with the precautions and the environmental data.

8. A server, characterized in that: The server includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to cause the server to execute the method according to any one of claims 1 to 7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instructions are executed on a server, the server is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product is run on a server, the server is caused to perform the method according to any one of claims 1 to 7.

Images