Self-adaptive green belt pruning system based on multi-sensor fusion

Abstract

The invention relates to the technical field of green belt pruning. The invention relates to a self-adaptive green belt pruning system based on multi-sensor fusion. The system comprises a sensor screening module, a region division module, an environment data fusion module, a model verification module and a pruning plan making module. The sensor screening module obtains the geographic position of a green belt to be pruned and screens out a temperature sensor and an image sensor whose detection range is overlapped with the geographic position. Multi-period image data is collected through a region division module, a three-dimensional simulation model is constructed in combination with geographic positions, exposed and sheltered regions are divided, then parameters such as green plant categories, illumination and temperatures are accurately supplemented to corresponding spatial positions of the model through an environment data fusion module, a dynamic model with morphological characteristics and growth characteristics is formed, and meanwhile, the dynamic model is subjected to dynamic modeling. And the model verification module realizes adaptive adjustment of model parameters through difference rate analysis of a predicted image and an actual image, so that the simulation model is always highly matched with the actual growth state of the green plant.

Description

Technical Field

[0001] This invention relates to the field of green belt pruning technology, and more specifically, to an adaptive green belt pruning system based on multi-sensor fusion. Background Technology

[0002] In the maintenance of landscaping and greening, green belt pruning is a core part. Existing pruning techniques include manual pruning, semi-automatic vehicle-mounted pruning equipment, and simple sensor-assisted pruning devices. Their function is to maintain the neat and beautiful shape of green belts, ensure the visibility of drivers on municipal roads and highways, and regulate the growth of green plants to promote their healthy and lush growth.

[0003] In existing green belt pruning techniques, manual pruning relies on the operator's experience and judgment, resulting in poor consistency in pruning height and amplitude, high labor intensity, and low work efficiency, making it difficult to meet the rapid maintenance needs of large-area green belts. Semi-automatic vehicle-mounted pruning equipment adopts a one-cut pruning mode with preset fixed parameters, which cannot distinguish the growth characteristics of different types of green plants, nor can it adjust the pruning strategy according to the real-time growth status of green plants. This easily leads to problems such as over-pruning that damages the root system of green plants or under-pruning that affects the landscape effect. In order to reduce this situation, an adaptive green belt pruning system based on multi-sensor fusion is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide an adaptive green belt pruning system based on multi-sensor fusion to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, an adaptive greenbelt pruning system based on multi-sensor fusion is provided, including a sensor selection module, a region division module, an environmental data fusion module, a model validation module, and a pruning plan formulation module.

[0006] The sensor screening module acquires the geographical location of the green belt to be pruned and filters out temperature sensors and image sensors whose detection range overlaps with that geographical location.

[0007] The region division module obtains the types of green plants, collects image data from different time periods based on the image sensor, establishes a blank model based on the geographical location, and inputs image data from different time periods to transform the green belt simulation model. It also obtains volume data based on the green belt simulation model and divides the exposed image area and the occluded image area.

[0008] The environmental data fusion module extracts light data for different time periods from image data, acquires temperature data for different time periods through temperature sensors, and then supplements the green belt simulation model with green plant type, temperature data, and light data.

[0009] The model verification module performs growth prediction simulation using the supplemented green belt simulation model, obtains predicted image data, and verifies the predicted image data of the same period by combining it with image data collected by the image sensor, and adjusts the green belt simulation model according to the verification results.

[0010] The pruning plan formulation module sets the pruning target shape, simulates the growth state of the occluded image area through the adjusted green belt simulation model, and generates a complete green belt model by combining the exposed image area. Then, a pruning plan is generated based on the complete green belt model and the pruning target shape.

[0011] As a further improvement to this technical solution, the sensor screening module establishes a communication connection with the green belt management terminal to obtain the geographical location of the green belt to be built.

[0012] The system searches for nearby sensors based on geographical location, obtains a sensor list, and acquires the detection range of each sensor. The detection range of each sensor is then compared with the geographical location of the green belt for spatial overlap filtering.

[0013] If the detection range spatially overlaps with the geographical location of the green belt, the sensor data is saved.

[0014] Conversely, if the detection range does not spatially overlap with the geographical location of the green belt, the sensor will be discarded.

[0015] The sensor types include temperature sensors and image sensors.

[0016] As a further improvement to this technical solution, in the area division module, the type of green plants corresponding to the green belt is obtained at the green belt management terminal;

[0017] The planting time of the green belt is obtained from the green belt management terminal. Then, the planting time is combined with the real time as the time range. The time range is then divided into multiple time periods. At the same time, the multiple time periods are combined and images are collected by the image sensor after filtering to obtain image data corresponding to different time periods, forming image datasets for different time periods.

[0018] A blank model is established based on the geographical location. Image datasets from different time periods are substituted into the blank model to complete the mapping and restoration of texture, contour and spatial form, and the blank model is transformed into a green belt simulation model.

[0019] As a further improvement to this technical solution, based on the green belt simulation model, spatial calculations are performed on the green belt, and volume data of the green belt at different time periods are obtained based on the calculation results.

[0020] By combining the image data and volume data corresponding to each time period, the green belt simulation model is divided into regions based on the image acquisition clarity and the occlusion of green space, thus dividing the exposed image region and the occluded image region.

[0021] The exposed image area is clear and unobstructed by greenery;

[0022] The occluded image area is characterized by a blurred image or the presence of vegetation.

[0023] As a further improvement to this technical solution, the environmental data fusion module performs illumination analysis on the green belt based on image data from different time periods to obtain illumination data of the green belt at different time periods.

[0024] Temperature data of the green belt was collected at different times using temperature sensors;

[0025] The light and temperature data are matched according to time period, and the light and temperature data of the same period are correlated. Then, the correlated light and temperature data are combined with the types of green plants in the green belt and added to the corresponding spatial location of the green belt simulation model.

[0026] As a further improvement to this technical solution, the model verification module, based on the supplemented green belt simulation model, calls the growth simulation algorithm and combines the inherent growth characteristics of green plant species to complete the simulation of the natural growth state of the green belt.

[0027] The growth prediction algorithm is used to predict the growth status of the green belt in a randomly selected period, generating predicted image data of the green belt for that period. Then, based on the selected period, image data of the same period is extracted from the filtered image sensor.

[0028] As a further improvement to this technical solution, a difference threshold is set in the model verification module;

[0029] Image data from the same time period is extracted and combined with predicted image data to perform difference rate analysis and obtain the image difference rate.

[0030] Then, the difference rate is compared with the difference threshold;

[0031] When the difference rate is greater than the difference threshold, the growth parameters and environmental parameters of the green belt simulation model are adjusted.

[0032] Conversely, when the difference rate is less than the difference threshold, the parameters of the green belt simulation model remain unchanged.

[0033] As a further improvement to this technical solution, the pruning plan formulation module sets the pruning target shape through the green belt management terminal;

[0034] Based on the green belt simulation model after parameter adjustment, the growth simulation algorithm is called to simulate the growth status of green plants in the occluded image area divided in the latest time period, obtain the growth status of green plants in the occluded image area, and then combine the light data, temperature data, green plant type and growth status of adjacent exposed areas in the occluded image area to determine the green belt model corresponding to the occluded image area.

[0035] The green belt model of the occluded image area is spatially stitched and fused with the exposed image area in the green belt simulation model. At the same time, the stitched and fused model is subjected to edge optimization, spatial completion and morphological calibration to form a complete green belt model.

[0036] As a further improvement to this technical solution, the pruning plan formulation module performs spatial difference analysis between the complete green belt model and the pruning target morphology to obtain spatial difference data between the complete green belt model and the pruning target morphology.

[0037] The operational capabilities of the pruning equipment are obtained, and then the spatial difference data are combined with the growth tolerance of the green plant species and the operational capabilities of the pruning equipment to conduct pruning plan analysis. The optimal pruning parameters are determined from the analysis results, and a pruning plan for the green belt is generated.

[0038] The pruning plan is converted into executable pruning operation instructions, which are then sent to the pruning execution equipment. The control equipment then prunes the green belt according to the operation instructions.

[0039] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0040] 1. In this adaptive green belt pruning system based on multi-sensor fusion, the region division module collects image data from multiple time periods, combines it with geographical location to construct a three-dimensional simulation model and divide it into exposed and shaded areas. Then, the environmental data fusion module accurately supplements the model with parameters such as plant type, light, and temperature to the corresponding spatial location, forming a dynamic model that combines morphological features and growth characteristics. At the same time, the model verification module achieves adaptive adjustment of model parameters by analyzing the difference rate between predicted and actual images, so that the simulation model always highly matches the actual growth state of the green plants, providing accurate basis for the growth simulation of shaded areas and the construction of a complete model.

[0041] 2. In this adaptive green belt pruning system based on multi-sensor fusion, the sensor screening module accurately matches temperature sensors and image sensors that overlap with the geographical location of the green belt, eliminating invalid sensor data. At the same time, the feature-level fusion method is used to process multi-source sensor data, which can dynamically adjust the sensor weights according to changes in outdoor environment such as light and rainfall. This greatly improves the accuracy and reliability of green belt perception data in complex environments, laying a solid data foundation for the subsequent construction of simulation models. Attached Figure Description

[0042] Figure 1 This is a schematic diagram of the adaptive green belt pruning system based on multi-sensor fusion of the present invention;

[0043] Figure 2 This is a flowchart illustrating the sensor screening module of the present invention;

[0044] Figure 3 This is a flowchart illustrating the region division module of the present invention;

[0045] Figure 4 This is a flowchart illustrating the environmental data fusion module of the present invention.

[0046] Figure 5 This is a flowchart illustrating the model verification module of the present invention;

[0047] Figure 6 This is a flowchart illustrating the pruning plan formulation module of the present invention. Detailed Implementation

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] Please see Figures 1-6 As shown, the purpose of this embodiment is to provide an adaptive green belt pruning system based on multi-sensor fusion, including a sensor screening module, a region division module, an environmental data fusion module, a model validation module, and a pruning plan formulation module.

[0050] The sensor screening module acquires the geographical location of the green belt to be pruned and filters out temperature sensors and image sensors whose detection range overlaps with that geographical location.

[0051] This enables the acquisition of the geographical location of the green belt to be pruned and the accurate selection of effective sensors, defining the sensor range and data type boundaries for the data source collection of all subsequent modules, eliminating invalid sensor data from the source, and avoiding interference from redundant data to subsequent model construction and growth simulation.

[0052] In the sensor screening module, a communication connection is established with the green belt management terminal to obtain the geographical location of the green belt to be built.

[0053] Establish a communication connection between the system and the green belt management terminal, and retrieve the geographical location information of the green belt to be pruned from the green belt management terminal through this communication link;

[0054] The system searches for nearby sensors based on geographical location, obtains a sensor list, and acquires the detection range of each sensor. The detection range of each sensor is then compared with the geographical location of the green belt for spatial overlap filtering.

[0055] Using the geographical location of the retrieved green belt as the retrieval benchmark, search for all sensors deployed around that geographical location, generate an initial sensor list, and obtain the effective detection range information of each sensor in the list from the sensor database;

[0056] If the detection range spatially overlaps with the geographical location of the green belt, the sensor data is saved.

[0057] Conversely, if the detection range does not spatially overlap with the geographical location of the green belt, the sensor will be discarded.

[0058] The sensor types include temperature sensors and image sensors.

[0059] The region division module obtains the types of green plants, collects image data from different time periods based on the image sensor, establishes a blank model based on the geographical location, and inputs image data from different time periods to transform the green belt simulation model. It also obtains volume data based on the green belt simulation model and divides the exposed image area and the occluded image area.

[0060] Based on the effective image sensor of the sensor screening module, the basic data collection and three-dimensional simulation model construction of the green belt are completed. At the same time, the time-period volume measurement and exposure / shading area division of the green belt are realized. It is a key transition module from data collection to model analysis of the whole system, providing a basic three-dimensional model carrier and area division basis for subsequent growth simulation and model completion.

[0061] In the area division module, the types of green plants corresponding to the green belt are obtained from the green belt management terminal;

[0062] The planting time of the green belt is obtained from the green belt management terminal. Then, the planting time is combined with the real time as the time range. The time range is then divided into multiple time periods. At the same time, the multiple time periods are combined and images are collected by the image sensor after filtering to obtain image data corresponding to different time periods, forming image datasets for different time periods.

[0063] The planting time (start time) of the green belt is obtained from the green belt management terminal, and the real-time time of the system (end time) is collected at the same time. The total time range of image acquisition is determined by using these two times as boundaries. Based on the above total time range, multiple image acquisition periods are divided according to the preset equidistant time intervals.

[0064] A blank model is established based on the geographical location. Image datasets from different time periods are substituted into the blank model to complete the mapping and restoration of texture, contour and spatial form, and the blank model is transformed into a green belt simulation model.

[0065] The valid image sensors output by the sensor filtering module are called to collect green belt image data in each acquisition period. The data is stored according to the time period to form image datasets for different time periods. Then, based on the geographical location spatial coordinates of the green belt to be pruned, a blank 3D model matching the spatial dimension of the geographical location is built. The image datasets of different time periods are then substituted into the blank 3D model one by one to complete the texture mapping, contour restoration and spatial morphology reconstruction of the model, and the blank 3D model is transformed into a green belt simulation model.

[0066] ;

[0067] in, Let be the volume of the green belt in time period t. , , This represents the total number of voxels in the three-dimensional space (x / y / z axes) of the green belt simulation model. The filling state of the voxel at position (x,y,z) in the model at time t (ρ=1 indicates the presence of greenery, ρ=0 indicates the absence of greenery). The fixed volume of a single voxel;

[0068] Based on the spatial coordinate system of the green belt simulation model, the green belt is spatially measured, and the volume data of the green belt at different time periods is obtained based on the measurement results.

[0069] By combining the image data and volume data corresponding to each time period, the green belt simulation model is divided into regions based on the image acquisition clarity and the occlusion of green space, thus dividing the exposed image region and the occluded image region.

[0070] By combining image datasets and corresponding volume data from different time periods, and using preset image clarity thresholds and green plant occlusion percentage thresholds as criteria, the green belt simulation model is divided into regions.

[0071] When the image clarity is greater than or equal to the clarity threshold and the proportion of green plants obscuring the image is 0, it is determined to be an exposed image area.

[0072] When the image clarity is less than the clarity threshold or the percentage of greenery obstruction is greater than 0, it is determined to be an obstructed image area;

[0073] The exposed image area is clear and unobstructed by greenery;

[0074] The occluded image area is characterized by a blurred image or the presence of vegetation.

[0075] The environmental data fusion module extracts light data for different time periods from image data, acquires temperature data for different time periods through temperature sensors, and then supplements the green belt simulation model with green plant type, temperature data, and light data.

[0076] Based on the effective temperature sensor of the sensor screening module and the image dataset and simulation model of the region division module, the system completes the collection and fusion of two core environmental data: light and temperature. It also supplements the corresponding spatial locations of the green belt simulation model with green plant category, light data, and temperature data, upgrading the original basic simulation model that only has spatial morphology into a complete simulation model that integrates category characteristics and environmental parameters, providing accurate model parameter support for subsequent growth simulation and model verification.

[0077] In the environmental data fusion module, the illumination analysis of the green belt is performed based on image data from different time periods to obtain illumination data of the green belt at different time periods;

[0078] Based on the image datasets of different time periods output by the region division module, pixel-level illumination feature analysis is performed on the green belt images of each time period. Illumination data such as illumination intensity and duration of illumination are extracted for each time period and spatial location. The data are classified and stored according to the dimensions of time period and spatial location to form a structured illumination dataset.

[0079] Temperature data of the green belt was collected at different times using temperature sensors;

[0080] The effective temperature sensors output by the sensor filtering module are called, and the ambient temperature data of different spatial locations in the green belt are collected at a preset acquisition frequency (synchronized with the image acquisition period). The data are also classified and stored according to time period and spatial location to form a structured temperature dataset.

[0081] The light and temperature data are matched according to time period, and the light and temperature data of the same period are correlated. Then, the correlated light and temperature data are combined with the types of green plants in the green belt and added to the corresponding spatial location of the green belt simulation model.

[0082] Using the collection time period as the sole matching dimension, data from the same time period and spatial location in the light and temperature datasets are associated to generate a "time period-spatial location-light-temperature" associated dataset, ensuring the spatiotemporal consistency of the two types of environmental data. Then, based on the three-dimensional spatial coordinate system (x,y,z) of the green belt simulation model, the light and temperature data in the associated datasets are combined with the attribute parameters of the green plant category corresponding to the spatial location to accurately map and supplement the corresponding coordinate points of the model, completing the spatial filling of the environmental parameters of the simulation model.

[0083] The model verification module performs growth prediction simulation using the supplemented green belt simulation model, obtains predicted image data, and verifies the predicted image data of the same period by combining it with image data collected by the image sensor, and adjusts the green belt simulation model according to the verification results.

[0084] The green belt simulation model is supplemented by the environmental data fusion module to complete the simulation of natural growth and prediction of growth status of the green belt. By comparing the predicted image data and the actual image data, the parameters of the simulation model are dynamically adjusted according to the preset difference threshold to ensure that the growth characteristics of the simulation model are highly matched with the actual growth status of the green belt. This avoids the distortion of subsequent shading area simulation and pruning plan formulation due to model parameter deviation. It is the core module for realizing system self-adaptation.

[0085] In the model verification module, based on the supplemented green belt simulation model, the growth simulation algorithm is called, and combined with the inherent growth characteristics of the green plant species, the natural growth state of the green belt is simulated.

[0086] The environmental data fusion module is used to supplement the complete green belt simulation model. The Logistic growth simulation algorithm, adapted to the growth characteristics of green plants, is then invoked. The model is input with the inherent growth parameters of each plant species (such as maximum growth and growth rate coefficient), light data, and temperature data. This completes the simulation of the natural growth state of the green belt across the entire space and at all times. The formula is as follows:

[0087] ;

[0088] in, Let represent the growth of greenery at spatial location (x, y, z) during time period t. This represents the maximum growth rate (inherent characteristic parameter) of the plant species at this location. This represents the basic growth rate coefficient (inherent characteristic parameter) of the plant species at this location. This refers to the start time of plant growth (planting time). For environmental correction functions (light and temperature);

[0089] The growth prediction algorithm is used to predict the growth status of the green belt in a randomly selected period, generating predicted image data of the green belt for that period. Then, based on the selected period, image data of the same period is extracted from the filtered image sensor.

[0090] Based on the growth simulation results, the growth prediction algorithm is called to randomly select one or more subsequent prediction periods for the green belt, predict the growth status (plant height, crown width, volume, etc.) of the green belt during the prediction period, and convert the predicted growth status into visualized prediction image data.

[0091] Based on the selected prediction period, the actual image data of the green belt collected during that period is extracted from the valid image sensors output by the sensor filtering module. At the same time, the judgment threshold of the image difference rate is set (usually 0.1~0.2, i.e. 10%~20%), taking into account the growth characteristics of green plants and the image recognition accuracy requirements.

[0092] In the model validation module, a difference threshold is set;

[0093] Image data from the same time period is extracted and combined with predicted image data to perform difference rate analysis and obtain the image difference rate.

[0094] Then, the difference rate is compared with the difference threshold, as shown in the following formula:

[0095] ;

[0096] ;

[0097] in, Image difference rate (value 0~1, 0 for no difference, 1 for complete difference). To predict image data and actual image data, It is a structural similarity index. To predict the pixel mean of the image and the actual image, To calculate the pixel standard deviation between the predicted image and the actual image, To predict the pixel covariance between the image and the actual image, and is a constant, and α is the difference threshold;

[0098] >α When the difference rate is greater than the difference threshold, the growth parameters and environmental parameters of the green belt simulation model are adjusted; the parameters of the green belt simulation model (growth rate coefficient, light response coefficient, temperature response coefficient, etc.) are specifically corrected, and the growth simulation is re-executed.

[0099] If ≤α, then when the difference rate is less than the difference threshold, the parameters of the green belt simulation model remain unchanged.

[0100] The pruning plan formulation module sets the pruning target shape, simulates the growth state of the occluded image area through the adjusted green belt simulation model, and generates a complete green belt model by combining the exposed image area. Then, a pruning plan is generated based on the complete green belt model and the pruning target shape.

[0101] Integrating the results of all previous modules, the model of the shading area is completed and the complete model of the green belt is generated. Based on the complete model and the pruning target shape, a personalized, precise and feasible green belt pruning plan is formulated. Finally, the plan is transformed into execution instructions and sent to the pruning equipment, realizing the implementation from model analysis to actual pruning execution.

[0102] In the pruning plan formulation module, the pruning target shape is set through the green belt management terminal;

[0103] By inputting the landscape design requirements (such as crown shape, height, crown width, etc.) of the green belt to be pruned into the green belt management terminal, the abstract pruning target shape is transformed into a standardized target model with three-dimensional spatial coordinates;

[0104] Based on the green belt simulation model after parameter adjustment, the growth simulation algorithm is called to simulate the growth status of green plants in the occluded image area divided in the latest time period, obtain the growth status of green plants in the occluded image area, and then combine the light data, temperature data, green plant type and growth status of adjacent exposed areas in the occluded image area to determine the green belt model corresponding to the occluded image area.

[0105] After loading the model verification module and completing the parameter adjustment of the green belt simulation model, the Logistic growth simulation algorithm adapted to the growth characteristics of green plants is called. For the occluded image area divided in the latest time period, the light data, temperature data, and inherent growth parameters of the green plant type are input for the area. At the same time, the growth status data of the adjacent exposed area are integrated to simulate the growth status of green plants in the occluded area.

[0106] Based on the growth simulation results of the shaded area, the three-dimensional morphological parameters (plant height, crown width, volume, etc.) of the shaded area are corrected to generate an accurate green belt model corresponding to the shaded image area, as shown in the following formula:

[0107] ;

[0108] in, Let represent the amount of vegetation growth at location (x, y, z) in the shaded area during time period t. This is a neighborhood correction coefficient (values ​​range from 0.9 to 1.1, adapted to the growth correlation between shaded and exposed areas). To cover the area adjacent to the exposed area The actual growth;

[0109] The green belt model of the occluded image area is spatially stitched and fused with the exposed image area in the green belt simulation model. At the same time, the stitched and fused model is subjected to edge optimization, spatial completion and morphological calibration to form a complete green belt model.

[0110] The occluded area model and the exposed image area model in the simulation model are aligned in three-dimensional space coordinates to complete the basic spatial stitching and fusion. At the same time, the stitched model is sequentially subjected to edge optimization (eliminating stitching gaps), spatial completion (filling missing three-dimensional data), and morphological calibration (matching the overall growth pattern) to finally generate a complete green belt model.

[0111] In the pruning plan formulation module, spatial difference analysis is performed between the complete model of the green belt and the pruning target morphology to obtain spatial difference data between the complete model of the green belt and the pruning target morphology.

[0112] Perform pixel-level 3D spatial comparison to extract spatial difference data such as height difference, crown difference, and volume difference at each spatial location;

[0113] The operational capabilities of the pruning equipment are obtained, and then the spatial difference data are combined with the growth tolerance of the green plant species and the operational capabilities of the pruning equipment to conduct pruning plan analysis. The optimal pruning parameters are determined from the analysis results, and a pruning plan for the green belt is generated.

[0114] The system retrieves the operational capability parameters of the pruning equipment (maximum pruning height, pruning accuracy, operating speed, blade compatibility range, etc.) from the equipment database to form a set of equipment operational capabilities. At the same time, it retrieves the growth tolerance parameters of the plant species (maximum pruning range, suitable pruning height, pruning cycle, etc.).

[0115] By performing multi-constraint coupling analysis on spatial difference data, plant growth tolerance, and equipment operation capabilities, the optimal pruning parameters (pruning area, pruning height, pruning sequence, operation speed, etc.) that meet the requirements of plant growth health and equipment execution are selected.

[0116] The pruning plan is converted into executable pruning operation instructions, which are then sent to the pruning execution equipment. The control equipment then prunes the green belt according to the operation instructions.

[0117] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. An adaptive green belt pruning system based on multi-sensor fusion, characterized in that: It includes a sensor screening module, a region division module, an environmental data fusion module, a model validation module, and a pruning plan formulation module. The sensor screening module acquires the geographical location of the green belt to be pruned and filters out temperature sensors and image sensors whose detection range overlaps with that geographical location. The region division module obtains the types of green plants, collects image data from different time periods based on the image sensor, establishes a blank model based on the geographical location, and inputs image data from different time periods to transform the green belt simulation model. It also obtains volume data based on the green belt simulation model and divides the exposed image area and the occluded image area. The environmental data fusion module extracts light data for different time periods from image data, acquires temperature data for different time periods through temperature sensors, and then supplements the green belt simulation model with green plant type, temperature data, and light data. The model verification module performs growth prediction simulation using the supplemented green belt simulation model, obtains predicted image data, and verifies the predicted image data of the same period by combining it with image data collected by the image sensor, and adjusts the green belt simulation model according to the verification results. The pruning plan formulation module sets the pruning target shape, simulates the growth state of the occluded image area through the adjusted green belt simulation model, and generates a complete green belt model by combining the exposed image area. Then, a pruning plan is generated based on the complete green belt model and the pruning target shape.

2. The adaptive green belt pruning system based on multi-sensor fusion according to claim 1, characterized in that: In the sensor screening module, a communication connection is established with the green belt management terminal to obtain the geographical location of the green belt to be built. The system searches for nearby sensors based on geographical location, obtains a sensor list, and acquires the detection range of each sensor. The detection range of each sensor is then compared with the geographical location of the green belt for spatial overlap filtering. If the detection range spatially overlaps with the geographical location of the green belt, the sensor data is saved. Conversely, if the detection range does not spatially overlap with the geographical location of the green belt, the sensor will be discarded. The sensor types include temperature sensors and image sensors.

3. The adaptive green belt pruning system based on multi-sensor fusion according to claim 2, characterized in that: In the area division module, the types of green plants corresponding to the green belt are obtained from the green belt management terminal; The planting time of the green belt is obtained from the green belt management terminal. Then, the planting time is combined with the real time as the time range. The time range is then divided into multiple time periods. At the same time, the multiple time periods are combined and images are collected by the image sensor after filtering to obtain image data corresponding to different time periods, forming image datasets for different time periods. A blank model is established based on the geographical location. Image datasets from different time periods are substituted into the blank model to complete the mapping and restoration of texture, contour and spatial form, and the blank model is transformed into a green belt simulation model.

4. The adaptive green belt pruning system based on multi-sensor fusion according to claim 3, characterized in that: Based on the green belt simulation model, spatial calculations are performed on the green belt, and volume data corresponding to the green belt at different time periods are obtained based on the calculation results. By combining the image data and volume data corresponding to each time period, the green belt simulation model is divided into regions based on the image acquisition clarity and the occlusion of green space, thus dividing the exposed image region and the occluded image region. The exposed image area is clear and unobstructed by greenery; The occluded image area is characterized by a blurred image or the presence of vegetation.

5. The adaptive green belt pruning system based on multi-sensor fusion according to claim 1, characterized in that: In the environmental data fusion module, the illumination analysis of the green belt is performed based on image data from different time periods to obtain illumination data of the green belt at different time periods; Temperature data of the green belt was collected at different times using temperature sensors; The light and temperature data are matched according to time period, and the light and temperature data of the same period are correlated. Then, the correlated light and temperature data are combined with the types of green plants in the green belt and added to the corresponding spatial location of the green belt simulation model.

6. The adaptive green belt pruning system based on multi-sensor fusion according to claim 1, characterized in that: In the model verification module, based on the supplemented green belt simulation model, the growth simulation algorithm is called, and combined with the inherent growth characteristics of the green plant species, the natural growth state of the green belt is simulated. The growth prediction algorithm is used to predict the growth status of the green belt in a randomly selected period, generating predicted image data of the green belt for that period. Then, based on the selected period, image data of the same period is extracted from the filtered image sensor.

7. The adaptive green belt pruning system based on multi-sensor fusion according to claim 6, characterized in that: In the model validation module, a difference threshold is set; Image data from the same time period is extracted and combined with predicted image data to perform difference rate analysis and obtain the image difference rate. Then, the difference rate is compared with the difference threshold; When the difference rate is greater than the difference threshold, the growth parameters and environmental parameters of the green belt simulation model are adjusted. Conversely, when the difference rate is less than the difference threshold, the parameters of the green belt simulation model remain unchanged.

8. The adaptive green belt pruning system based on multi-sensor fusion according to claim 1, characterized in that: In the pruning plan formulation module, the pruning target shape is set through the green belt management terminal; Based on the green belt simulation model after parameter adjustment, the growth simulation algorithm is called to simulate the growth status of green plants in the occluded image area divided in the latest time period, obtain the growth status of green plants in the occluded image area, and then combine the light data, temperature data, green plant type and growth status of adjacent exposed areas in the occluded image area to determine the green belt model corresponding to the occluded image area. The green belt model of the occluded image area is spatially stitched and fused with the exposed image area in the green belt simulation model. At the same time, the stitched and fused model is subjected to edge optimization, spatial completion and morphological calibration to form a complete green belt model.

9. The adaptive green belt pruning system based on multi-sensor fusion according to claim 1, characterized in that: In the pruning plan formulation module, spatial difference analysis is performed between the complete model of the green belt and the pruning target morphology to obtain spatial difference data between the complete model of the green belt and the pruning target morphology. The operational capabilities of the pruning equipment are obtained, and then the spatial difference data are combined with the growth tolerance of the green plant species and the operational capabilities of the pruning equipment to conduct pruning plan analysis. The optimal pruning parameters are determined from the analysis results, and a pruning plan for the green belt is generated. The pruning plan is converted into executable pruning operation instructions, which are then sent to the pruning execution equipment. The control equipment then prunes the green belt according to the operation instructions.

Images