A plant growth stage monitoring method, device and computer equipment

Abstract

The application relates to a plant growth stage monitoring method and device and a computer device. The method comprises the following steps: acquiring multi-modal data of a target plant, wherein the multi-modal data features comprise environmental data features and image data features; inputting the image data features into a plant category recognition model to obtain a target type of the target plant; dynamically adjusting the image data features based on the environmental data features, and generating fusion features of the environmental data features and the dynamically adjusted image data features; and inputting the fusion features into a target growth stage classification network corresponding to the target type, and outputting a current growth stage classification result of the target plant.

Description

Technical Field

[0001] This application relates to the field of intelligent agricultural technology, and in particular to a method, device, and computer equipment for monitoring plant growth stages. Background Technology

[0002] With the development of smart agriculture, the requirements for precise environmental control in closed planting systems such as plant cabins and greenhouses are becoming increasingly stringent. Crops have significantly different needs for light, temperature, humidity, and nutrients at different growth stages. Accurately identifying growth stages and dynamically adjusting environmental parameters is key to improving yield and quality.

[0003] Traditional monitoring of growth stages relies on manual observation and experience, which suffers from low efficiency, strong subjectivity, and significant time lag, making it difficult to meet the needs of large-scale, automated planting. Even with the introduction of image recognition technology in some systems, challenges remain, such as high sample annotation costs, weak model generalization ability, and difficulty in adapting to complex planting environments. These challenges result in insufficient accuracy in identifying growth stages, affecting the precision of subsequent environmental control. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, device, and computer equipment for monitoring plant growth stages to address the aforementioned technical problems.

[0005] Firstly, this application provides a method for monitoring plant growth stages, including:

[0006] Acquire multimodal data of the target plant, wherein the multimodal data features include environmental data features and image data features;

[0007] The image data features are input into a plant category recognition model to obtain the target type of the target plant;

[0008] The image data features are dynamically adjusted based on the environmental data features, and a fusion feature of the environmental data features and the dynamically adjusted image data features is generated.

[0009] The fused features are input into the target growth stage classification network corresponding to the target type, and the classification result of the current growth stage of the plant is output.

[0010] In one embodiment, the step of inputting the fused features into a target growth stage classification network corresponding to the target type and outputting the classification result of the current growth stage of the plant includes:

[0011] Multiple fusion features of the target plant are input into the target growth stage classification network. The target growth stage classification network dynamically adjusts the stage classification decision boundary according to the environmental data features so that the same target plant outputs different stage classification results. The stage classification results include stage labels and confidence scores.

[0012] In one embodiment, the method further includes:

[0013] Obtain historical environmental time-series data of the target plant;

[0014] The historical environmental time series data is input into the mapping model to obtain the stage transition rate. The mapping model is trained based on the historical environmental data and the historical stage transition time data.

[0015] Based on the stage transfer rate, the time points of different stages of the target plant are determined;

[0016] Based on the current time and the time points of the different stages, the prediction stage result corresponding to the current time is obtained;

[0017] The prediction stage results are fused with the current growth stage classification results to obtain the final stage classification results.

[0018] In one embodiment, the target plant is obtained based on the image to be detected, and the method further includes:

[0019] Based on the image to be detected, a rectangular box region is determined in the image to be detected, and the rectangular box is used to mark the target plant in the image to be detected;

[0020] Text labels are generated based on the current growth stage classification results of the plant. The text labels are located within a rectangular box area, and the growth stage classification results include confidence levels.

[0021] In one embodiment, the method further includes

[0022] If the confidence level is less than a preset threshold, an alarm message is sent to prompt the image to be detected to be re-detected.

[0023] In one embodiment, prior to acquiring the multimodal data of the target plant, the process includes:

[0024] Obtain unlabeled images;

[0025] Based on the unlabeled image, identify the occluders in the unlabeled image;

[0026] Plant regions are segmented in unlabeled images based on a semantic segmentation model to obtain preset plant regions.

[0027] The actual plant area is obtained based on the percentage of overlap between the obstruction and the preset plant area.

[0028] Secondly, this application also provides a monitoring device for plant growth stages, comprising:

[0029] The acquisition module is used to acquire multimodal data of the target plant, wherein the multimodal data features include environmental data features and image data features;

[0030] The recognition module is used to input the image data features into the plant category recognition model to obtain the target type of the target plant;

[0031] The fusion module is used to dynamically adjust the image data features based on the environmental data features, and generate a fused feature of the environmental data features and the dynamically adjusted image data features;

[0032] The output module is used to input the fused features into the target growth stage classification network corresponding to the target type, and output the classification result of the current growth stage of the plant.

[0033] Fourthly, this disclosure also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of a method for monitoring plant growth stages.

[0034] Fifthly, this disclosure also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of a method for monitoring plant growth stages.

[0035] Sixthly, this disclosure also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of a method for monitoring plant growth stages.

[0036] The aforementioned monitoring of plant growth stages simultaneously acquires both environmental and image data features, taking into account both external growth conditions and the plant's morphological characteristics. Compared to single-image monitoring, this provides more complete information and stronger resistance to interference. Dynamically adjusting image data features using environmental data effectively eliminates the impact of environmental changes on visual features, maintaining stable recognition even in complex field scenarios. Deeply fusing environmental data features with corrected image data features creates a more informative fusion feature set, enabling the model to learn more comprehensive growth patterns. First, the plant target type is determined through image features, then a growth stage classification network for the corresponding variety is used, avoiding confusion between features of different plant growth stages and significantly improving the accuracy of growth stage identification. Attached Figure Description

[0037] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0038] Figure 1 This is an application environment diagram of a plant growth stage monitoring method in one embodiment;

[0039] Figure 2 This is a flowchart illustrating a method for monitoring plant growth stages in one embodiment;

[0040] Figure 3 This is a structural block diagram of a plant growth stage monitoring device in one embodiment;

[0041] Figure 4 This is an internal structural diagram of a computer device in one embodiment;

[0042] Figure 5 This is an internal structure diagram of a server in one embodiment. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0044] It should be noted that the terms "first," "second," etc., used in this application can be used to describe various elements, but these elements are not limited by these terms. These terms are only used to distinguish the first element from the second element. The terms "comprising" and "having," and any variations thereof, used in this application, are intended to cover non-exclusive inclusion. The term "multiple" used in this application refers to two or more. The term "and / or" used in this application refers to one of the embodiments, or any combination of multiple embodiments.

[0045] The plant growth stage monitoring method provided in this application embodiment can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on the cloud or other network servers. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, drones, low-altitude aircraft, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0046] In one exemplary embodiment, such as Figure 2 As shown, a method for monitoring plant growth stages is provided. Wherein:

[0047] S202: Acquire multimodal data of the target plant, wherein the multimodal data features include environmental data features and image data features.

[0048] Multimodal data refers to combinations of data from different sensors and in different formats, including various environmental data and image data. Environmental data refers to external environmental parameters affecting plant growth collected by sensors, including air temperature, air humidity, light intensity, CO2 concentration, soil temperature and humidity, wind speed, and rainfall. These are used to determine the external conditions for plant growth and help models more accurately distinguish growth stages. Image data can be multispectral images and RGB images of plants acquired through spectrometers and vision cameras. Image data features include morphological features, structural features, and spectral features. Morphological features include plant outline, leaf shape, leaf number, and leaf texture; structural features include plant height, foliage density, and growth distribution; and spectral features include reflectance at different wavelengths and vegetation indices.

[0049] By constructing multimodal data containing image and environmental information, we can achieve centralized storage, standardized processing, and one-to-one correspondence of plant visual feature data and environmental sensing data, ensuring that each frame of image data matches the environmental parameters at the time of acquisition, thus providing a standardized data foundation for subsequent multimodal feature fusion and model training.

[0050] After constructing the multimodal data, preprocessing can be performed, including image normalization, size unification and format conversion, image denoising and quality screening. Standardization improves data consistency and usability. To enhance the generalization ability of subsequent models, data augmentation strategies such as random rotation, flipping, cropping, brightness and contrast adjustment, and random scaling are further employed to enrich sample diversity and reduce the impact of changes in illumination, angle, and scale on model recognition performance.

[0051] S204: Input the image data features into the plant category recognition model to obtain the target type of the target plant.

[0052] To accurately identify plant growth stages, both deep learning and traditional machine learning methods can be used simultaneously to build a plant category recognition model. Deep learning models, including CNN (Convolutional Neural Networks), ResNet, EfficientNet, VisionTransformer (ViT), and Swin Transformer, are used to automatically extract deep visual features from images. Traditional machine learning models, including Random Forest and Support Vector Machine (SVM), are used to process structured data such as environmental data. First, the multimodal dataset is divided into training, validation, and test sets. Then, model training and parameter updates are performed, and cross-validation is used to further improve model stability and generalization ability, ensuring reliable recognition results.

[0053] The performance of the plant category recognition model is comprehensively evaluated using multiple metrics, including accuracy, recall, precision, and F1 score, to fully measure the correctness, completeness, and balance of growth stage identification. Simultaneously, hyperparameter tuning is performed through methods such as learning rate adjustment, batch size optimization, and network layer adjustment to improve model convergence speed, recognition accuracy, and generalization ability, ensuring the stability and reliability of the plant category recognition model in real-world agricultural scenarios.

[0054] S206: Dynamically adjust the image data features based on the environmental data features, and generate a fusion feature of the environmental data features and the dynamically adjusted image data features.

[0055] Images are easily distorted by light, temperature, and humidity; for example, they appear washed out in strong light and dark on cloudy days. Environmental data can be used as correction parameters to automatically adjust image features. For instance, if the light is too strong, the image brightness feature can be automatically reduced; if the light is too weak, the feature can be automatically enhanced, resulting in more realistic, stable, and unaffected by environmental interference. Dynamic correction of image data features based on environmental data eliminates the interference of environmental factors such as light, temperature, and humidity on visual features. The corrected image features are then fused with the original environmental features to generate a fused feature that combines plant morphology and environmental condition information, providing a more comprehensive and stable feature input for subsequent growth stage identification.

[0056] S208: Input the fused features into the target growth stage classification network corresponding to the target type, and output the classification result of the current growth stage of the plant.

[0057] First, identify the type of crop to be monitored, then determine its typical growth stages. For example, in the seed stage, focus on seed morphology and germination status; in the seedling stage, focus on leaf quantity, leaf unfolding degree, and leaf texture; during the rapid growth period, focus on leaf area growth, branch and leaf density, and plant height changes; in the maturity stage, focus on leaf color changes, flower and fruit formation, and spectral reflectance characteristics. The characteristic differences of different stages will serve as an important basis for the learning of the target growth stage classification network. The fused features are input into the target growth stage classification network corresponding to the target type, and the current growth stage classification result of the target plant is output.

[0058] The aforementioned methods for monitoring plant growth stages simultaneously acquire both environmental and image data features, taking into account both external growth conditions and the plant's morphological characteristics. Compared to single-image monitoring, this provides more complete information and stronger resistance to interference. Dynamically adjusting image data features using environmental data effectively eliminates the impact of environmental changes on visual features, maintaining stable recognition even in complex field scenarios. Deeply fusing environmental data features with corrected image data features creates richer fusion features, enabling the model to learn more comprehensive growth patterns. First, the plant target type is determined through image features, then a growth stage classification network for the corresponding variety is used, avoiding confusion between features of different plant growth stages and significantly improving the accuracy of growth stage identification.

[0059] In an exemplary embodiment, the step of inputting the fused features into a target growth stage classification network corresponding to the target type and outputting the classification result of the current growth stage of the plant includes:

[0060] Multiple fusion features of the target plant are input into the target growth stage classification network. The target growth stage classification network dynamically adjusts the stage classification decision boundary according to the environmental data features so that the same target plant outputs different stage classification results. The stage classification results include stage labels and confidence scores.

[0061] Multiple fusion features represent multiple sets of features across multiple time points, angles, and environments. The dynamic adjustment of decision boundaries allows the target growth stage classification network to automatically change judgment criteria based on temperature, humidity, light, and CO2. For example, if low temperature leads to slow growth, the judgment criteria can be relaxed. Because the environment is constantly changing, the same plant will be classified into different growth stages under different conditions, which is more in line with real agricultural practices.

[0062] The target growth stage classification network can dynamically and adaptively adjust the growth stage classification decision boundary based on environmental data characteristics, so that the same plant can output different stage classification results under different environmental conditions, and finally output complete classification information including growth stage labels and confidence scores.

[0063] In one exemplary embodiment, the method further includes:

[0064] Obtain historical environmental time-series data of the target plant;

[0065] The historical environmental time series data is input into the mapping model to obtain the stage transition rate. The mapping model is trained based on the historical environmental data and the historical stage transition time data.

[0066] Based on the stage transfer rate, the time points of different stages of the target plant are determined;

[0067] Based on the current time and the time points of the different stages, the prediction stage result corresponding to the current time is obtained;

[0068] The prediction stage results are fused with the current growth stage classification results to obtain the final stage classification results.

[0069] Historical environmental time-series data is obtained by collecting continuous environmental data over a period of time. This data is then arranged chronologically to form a time-series sequence, used to determine whether the plant grew quickly or slowly in the past. This historical environmental time-series data is input into a mapping model to obtain the stage transition rate. The mapping model is trained using historical environmental data and historical stage transition times, and environmental trends are used to calculate the speed at which the plant grows from one stage to the next. Based on historical environmental data, the growth rate can be calculated, and the future entry into the next stage can be predicted.

[0070] Stage transition rate refers to the rate or time change of a plant transitioning from one growth stage to the next under current environmental conditions. It is used to quantify the influence of environmental factors on the growth rate of plants. For example, plants grow faster and have a faster stage transition rate when the temperature is high, the light is sufficient, and the humidity is suitable. It describes the time from the seedling stage to the growth stage and from the growth stage to the maturity stage.

[0071] By using the stage transition rate, the expected time points of each stage from seed stage to seedling stage, from seedling stage to growth stage, and from growth stage to maturity stage can be calculated.

[0072] By comparing the current time with the aforementioned stage time points, the predicted growth stage based on the stage transition rate is output. This is a trend prediction result, not real-time image recognition.

[0073] The prediction results based on historical environmental time-series data are fused with the current growth stage classification results obtained through real-time inference based on multimodal data. This combined approach corrects the identification bias of single-path methods, outputting a stable, accurate, and consistent final growth stage classification result that aligns with plant growth patterns. Real-time identification, due to abnormal lighting conditions today (leading to overly bright images), might misclassify a stage as mature. The prediction stage, based on past temperature and light levels, determines that the stage is still in the growth period. The fusion process, by comprehensively judging the final growth stage, prevents errors from single identifications and ensures the results better reflect agricultural patterns.

[0074] In one exemplary embodiment, the target plant is obtained based on the image to be detected, and the method further includes:

[0075] Based on the image to be detected, a rectangular box region is determined in the image to be detected, and the rectangular box is used to mark the target plant in the image to be detected;

[0076] Text labels are generated based on the current growth stage classification results of the plant. The text labels are located within a rectangular box area, and the growth stage classification results include confidence levels.

[0077] The plants to be monitored by the system are obtained from the images to be detected. The system uses object detection algorithms to locate the plant regions in the images to be detected, generates rectangular boxes to label the target plants, generates text labels with confidence scores based on the classification results of the current growth stage of the plants, and places the text labels within the rectangular boxes to complete the visual labeling of the plant growth stages.

[0078] In one exemplary embodiment, the method further includes:

[0079] If the confidence level is less than a preset threshold, an alarm message is sent to prompt the image to be detected to be re-detected.

[0080] When the prediction confidence of the target growth stage classification network falls below a set threshold, the sample automatically enters the manual review queue. Reviewers can correct erroneous annotations through the annotation system interface, and the corrected data will be re-added to the training set for model iteration. The system can assist the model in identifying abnormal annotations, such as unreasonable bounding box sizes, incorrect growth stage predictions, and abnormal image quality. The system automatically filters and marks abnormal data. For images that fail to be automatically annotated, the system will analyze the reasons, such as plant occlusion, excessively strong or dim lighting, or image blurring. For images that still cannot be automatically annotated, manual annotation will be performed.

[0081] In an exemplary embodiment, prior to acquiring the multimodal data of the target plant, the process includes:

[0082] Obtain unlabeled images;

[0083] Based on the unlabeled image, identify the occluders in the unlabeled image;

[0084] Plant regions are segmented in unlabeled images based on a semantic segmentation model to obtain preset plant regions.

[0085] The actual plant area is obtained based on the percentage of overlap between the obstruction and the preset plant area.

[0086] Unlabeled images are raw field photos that have not yet been manually labeled, bounding boxes selected, or categorized. The goal is to identify non-plant elements or those that might obscure plant areas, such as weeds, soil, or overlapping leaves. A semantic segmentation model is used to mark possible plant regions in the image, resulting in a preliminary plant region. Then, the extent to which these preliminary plant regions are covered by occlusion is calculated. Excessively obscured or invalid parts are removed, leaving clean, complete, and unobstructed areas to obtain the true plant regions.

[0087] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages in other steps. It is understood that the steps in different embodiments can be freely combined as needed, and all non-contradictory solutions formed by such combinations are within the scope of protection of this application.

[0088] Based on the same inventive concept, this application also provides a plant growth stage monitoring device for implementing the above-described plant growth stage monitoring method. The solution provided by this device is similar to the solution described in the above-described method; therefore, the specific limitations of one or more plant growth stage monitoring device embodiments provided below can be found in the limitations of the plant growth stage monitoring method described above, and will not be repeated here.

[0089] In one exemplary embodiment, such as Figure 3 As shown, a plant growth stage monitoring device 300 is provided, comprising:

[0090] The acquisition module 302 is used to acquire multimodal data of the target plant, wherein the multimodal data features include environmental data features and image data features;

[0091] The recognition module 304 is used to input the image data features into the plant category recognition model to obtain the target type of the target plant;

[0092] The fusion module 306 is used to dynamically adjust the image data features based on the environmental data features, and generate a fusion feature of the environmental data features and the dynamically adjusted image data features;

[0093] The output module 308 is used to input the fused features into the target growth stage classification network corresponding to the target type, and output the classification result of the current growth stage of the plant.

[0094] Each module in the aforementioned plant growth stage monitoring device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0095] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0096] The modules in the aforementioned device for monitoring plant growth stages can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the computer device's memory as software, so that the processor can call and execute the corresponding operations of each module.

[0097] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows:Figure 4 As shown, the computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The database stores data. The network interface communicates with external terminals via a network connection. When executed by the processor, the computer program implements a method for monitoring plant growth stages.

[0098] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 5 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a method for monitoring plant growth stages. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0099] Those skilled in the art will understand that Figure 4 , Figure 5 The structure shown is merely a block diagram of a portion of the structure related to the present disclosure and does not constitute a limitation on the computer device to which the present disclosure is applied. A specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0100] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the methods described in any embodiment of this disclosure.

[0101] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the methods described in any embodiment of this disclosure.

[0102] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this disclosure can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this disclosure may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this disclosure may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0103] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0104] The embodiments described above are merely illustrative of several implementations of this disclosure, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent disclosure. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this disclosure, and these all fall within the protection scope of this disclosure. Therefore, the protection scope of this disclosure should be determined by the appended claims.

Claims

1. A method for monitoring plant growth stages, characterized in that, The method includes: Acquire multimodal data of the target plant, wherein the multimodal data features include environmental data features and image data features; The image data features are input into a plant category recognition model to obtain the target type of the target plant; The image data features are dynamically adjusted based on the environmental data features, and a fusion feature of the environmental data features and the dynamically adjusted image data features is generated. The fused features are input into the target growth stage classification network corresponding to the target type, and the current growth stage classification result of the target plant is output.

2. The method according to claim 1, characterized in that, The step of inputting the fused features into the target growth stage classification network corresponding to the target type and outputting the classification result of the current growth stage of the plant includes: Multiple fusion features of the target plant are input into the target growth stage classification network. The target growth stage classification network dynamically adjusts the stage classification decision boundary according to the environmental data features so that the same target plant outputs different stage classification results. The stage classification results include stage labels and confidence scores.

3. The method according to claim 1, characterized in that, The method further includes: Obtain historical environmental time-series data of the target plant; The historical environmental time series data is input into the mapping model to obtain the stage transition rate. The mapping model is trained based on the historical environmental data and the historical stage transition time data. Based on the stage transfer rate, the time points of different stages of the target plant are determined; Based on the current time and the time points of the different stages, the prediction stage result corresponding to the current time is obtained; The prediction stage results are fused with the current growth stage classification results to obtain the final stage classification results.

4. The method according to claim 1, characterized in that, The target plant is obtained based on the image to be detected, and the method further includes: Based on the image to be detected, a rectangular box region is determined in the image to be detected, and the rectangular box is used to mark the target plant in the image to be detected; Text labels are generated based on the current growth stage classification results of the plant. The text labels are located within a rectangular box area, and the growth stage classification results include confidence levels.

5. The method according to claim 4, characterized in that, The method further includes: If the confidence level is less than a preset threshold, an alarm message is sent to prompt the image to be detected to be re-detected.

6. The method according to claim 1, characterized in that, Before acquiring the multimodal data of the target plant, the following steps are included: Obtain unlabeled images; Based on the unlabeled image, identify the occluders in the unlabeled image; Plant regions are segmented in unlabeled images based on a semantic segmentation model to obtain preset plant regions. The actual plant area is obtained based on the percentage of overlap between the obstruction and the preset plant area.

7. A monitoring device for plant growth stages, characterized in that, The device includes: The acquisition module is used to acquire multimodal data of the target plant, wherein the multimodal data features include environmental data features and image data features; The recognition module is used to input the image data features into the plant category recognition model to obtain the target type of the target plant; The fusion module is used to dynamically adjust the image data features based on the environmental data features, and generate a fused feature of the environmental data features and the dynamically adjusted image data features; The output module is used to input the fused features into the target growth stage classification network corresponding to the target type, and output the classification result of the current growth stage of the plant.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images