Soil improvement-based slope ecological greening construction method and monitoring system

Abstract

The invention provides a slope ecological greening construction method based on soil improvement and a monitoring system, and is suitable for the technical field of slope greening construction. The method comprises the steps of plant-fungus combination selection, slope protection structure construction, hypha system construction, unmanned aerial vehicle sowing, earth surface film mulching and dynamic maintenance regulation and control. A mycorrhizal network is constructed by taking arbuscular mycorrhizal fungi and ectomycorrhizal fungi as cores, a mycelium fiber network is formed, and the soil fixation capability is enhanced by the cohesive force between the fiber network and soil and the strength of fibers; the stability of soil aggregates is improved through glomalin secreted by hyphae, and the corrosion resistance of the soil is further improved; a water-retaining agent is added, so that the water-retaining property of the backfill soil is improved while early-stage growth of green vegetation and a hypha network is guaranteed; finally, the unmanned aerial vehicle carries a multispectral camera and a radar for cooperative scanning, rapidly and accurately monitors the soil moisture content, sets a soil moisture content warning value, reminds maintenance personnel to irrigate and maintains long-term survival of vegetation.

Description

Technical Field

[0001] The present invention relates to the technical field of slope greening construction, and in particular to a slope ecological greening construction method and a monitoring system based on soil improvement. Background Art

[0002] With the advancement of ecological civilization and the concept of green infrastructure development, slope greening restoration technology is gradually developing in an ecological and intelligent direction. Traditional lattice-type slope greening construction technology uses backfill soil and a single vegetation cover for greening restoration. However, before the plant roots grow and become more widespread, the overall soil solidification capacity is insufficient, resulting in rapid loss of backfill soil under rainy season. In addition, the backfill soil only has a shallow surface layer, and below the backfill is concrete used to reinforce the slope, resulting in poor soil water retention. Under long-term drought conditions without rain, the vegetation survival rate is very low.

[0003] Existing techniques can protect backfill soil with rope nets or other mesh materials to reduce backfill loss. However, this method not only affects vegetation growth, fails to address water retention and water shortages, but also conflicts with the concept of green restoration. Therefore, there is an urgent need for a slope ecological greening construction method that can address the problems of large-scale backfill loss and poor soil water retention. Summary of the Invention

[0004] The purpose of the present invention is to solve the problems of large-scale loss of backfill soil, poor soil water retention, and low long-term vegetation survival rate in traditional slope ecological greening construction. A slope ecological greening construction method and monitoring system based on soil improvement are proposed, which can improve the stability and water retention of backfill soil. At the same time, an intelligent soil moisture content monitoring function is set. When the soil moisture content is insufficient, the maintenance personnel are reminded to irrigate to ensure the long-term survival rate of vegetation. It can be widely used in the field of slope greening construction technology.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A slope ecological greening construction method based on soil improvement specifically comprises the following steps:

[0007] S100, plant-fungus combination selection, including selecting a bacterial system with arbuscular mycorrhizal fungi and ectomycorrhizal fungi as core species, Trichoderma and nitrogen-fixing bacteria as auxiliary bacterial communities, and biochar slow-release granules as the carrier of the bacterial agent. Herbs with well-developed root systems that can coexist with the bacterial system are selected as slope greening plants;

[0008] S200, slope protection structure construction, including the construction of a frame-type slope protection structure according to the slope reinforcement design, backfilling the cells surrounded by the frame with planting soil, and providing permeable holes in the slope concrete layer of each cell, and compacting the backfill to a density of z;

[0009] S300, constructing a mycelium system, including evenly arranging n mycelium culture tubes per square meter in the backfilled planting soil;

[0010] S400, drone sowing, using slope greening plant seeds, fungal agents and water-retaining agents to prepare seed bombs, placing the seed bombs in the magazine of the air pressure ejection device, and then using the drone equipped with the air pressure ejection device to sow;

[0011] S500, surface film covering, including laying straw degradable film on the surface of the planting soil and compacting the periphery of the straw degradable film;

[0012] S600, dynamic maintenance and control, including regular spraying of nutrient solution in the early stage of maintenance to ensure stable expansion of the mycelium until the mycelium network covers more than 80% of the planting soil surface. In the late stage of maintenance, the surface soil moisture content is collaboratively monitored based on the monitoring system's microwave-optical multi-source remote sensing technology. When the soil moisture content reaches the warning value, the maintenance personnel are reminded to perform watering maintenance.

[0013] As a preferred technical solution of the present invention, in step S300, each mycelium culture tube is filled with a mycelium matrix with a mass ratio of mushroom dregs: humus soil = 1:2, the moisture content in the tube is controlled at 45%-55%, and a degradable sealing cover is set on the top of the tube.

[0014] As a preferred technical solution of the present invention, in step S300, the number n of mycelium culture tubes arranged per square meter is determined based on the effective coverage area of ​​a single-tube mycelium. In the laboratory, under the conditions of the same slope as the on-site slope construction, the same permeability of the on-site soil, and the same maintenance method, the effective coverage area of ​​a single-tube mycelium is measured and recorded for archiving.

[0015] As a preferred technical solution of the present invention, in step S400, the seed bombs are sown avoiding the frame-type slope protection structure and the culture tubes, and the seed bombs are sown in the middle of adjacent culture tubes in each cell.

[0016] As a preferred technical solution of the present invention, in step S600, the microwave-optical multi-source remote sensing technology collaboratively monitors surface soil moisture, including microwave-optical data acquisition and preprocessing, multi-source data fusion modeling, and real-time monitoring and early warning;

[0017] The multi-source data fusion modeling includes using a generalized regression neural network model optimized by a cosine optimized particle swarm algorithm as a deep learning model, and using the time domain characteristics of effective radar signals and vegetation index as inputs of the deep learning model to predict soil moisture content;

[0018] The real-time monitoring and early warning system includes regularly updating data through collaborative scanning of the multispectral camera and radar carried by the drone, inputting the deep learning model into the inversion of the slope soil moisture content, and reminding maintenance personnel to irrigate when the soil moisture content is lower than the preset warning value.

[0019] As a preferred technical solution of the present invention, the construction process of the deep learning model includes microwave-optical data acquisition and preprocessing, soil moisture measurement, model training and verification;

[0020] The soil moisture content measurement includes gridding the area to be tested, inserting a soil thermometer and hygrometer into the soil of each grid, reading the soil moisture content value of each grid, and taking the average value after removing the maximum and minimum values ​​as the soil moisture content of the area to be tested;

[0021] The model training and verification includes taking the time domain characteristics, vegetation index and measured soil moisture content of the area to be detected obtained through microwave-optical data acquisition and preprocessing as a set of data, and also obtaining k groups of data from different areas to be detected to form a data set, dividing the data set into a training set and a validation set, and training and validating the deep learning model; in the deep learning model training process, the cosine optimized particle swarm algorithm is used to optimize the smoothing factor of the generalized regression neural network model, and the prediction accuracy of the deep learning model is measured by the determination coefficient R 2 , root mean square error RMSE and accuracy ACC are used for evaluation.

[0022]

[0023] Where y i - Measured value of soil moisture, y′ i -Predicted soil moisture content, y e -The mean of the measured values ​​of soil moisture content.

[0024] As a preferred technical solution of the present invention, the microwave-optical data acquisition and preprocessing includes using a drone equipped with an ultra-wideband radar to hover at an altitude of 1 meter above the surface of the area to be detected to collect the original echo signal of the surface radar, and using a multi-light slope camera equipped with the drone to obtain a multispectral image of the slope area; then, the original echo signal is intercepted according to the radar echo delay formula.

[0025]

[0026]

[0027] Where Δt is the time difference between the radar transmitting signal and receiving the surface echo, R is the distance between the radar antenna and the surface, V is the propagation speed of the radar wave in the medium, τ is the sampling frequency of the radar, and m is the distance between the radar antenna and the surface. -The number of sampling points; the interception starts from the mth sampling point of the original echo signal, and the intercepted second half of the radar echo signal is used as the effective radar signal; the time domain feature extraction of the effective radar signal is performed, including the calculation of the maximum amplitude, minimum amplitude, rectified mean, variance, root mean square, kurtosis, peak-to-peak value, skewness and peak factor; the surface vegetation index NDVI is calculated based on the multispectral image data,

[0028]

[0029] Where, ρ NR -Near-infrared band of multispectral image, ρ R -Infrared band of multispectral image; then the temporal characteristics and vegetation index are normalized to the maximum and minimum values.

[0030] A slope ecological greening monitoring system based on soil improvement, including a multi-source data acquisition module, a data preprocessing module, a deep learning model prediction module, an early warning module and a model iterative optimization module;

[0031] The multi-source data acquisition module is used to receive the original echo signal and multispectral image data collected by the ultra-wideband radar and multispectral camera carried by the drone through the network, as well as the coordinates and timestamps corresponding to the drone collecting the data, and save the original echo signal, multispectral image data, coordinates and timestamps in the corresponding slope project;

[0032] The data preprocessing module is used to extract time domain features of effective radar signals, calculate vegetation index based on multispectral image data, and perform maximum and minimum normalization processing on the time domain features and vegetation index;

[0033] The deep learning model prediction module is used to input the preprocessed temporal features and vegetation index into a trained generalized regression neural network model optimized by a cosine optimized particle swarm algorithm, and output the soil moisture content;

[0034] The early warning module is used to regularly update microwave-optical data through drone collaborative scanning, invert the slope soil moisture distribution through the data preprocessing module and the deep learning model prediction module, and trigger an early warning signal when the moisture content is detected to be lower than the preset warning value, reminding maintenance personnel to implement irrigation;

[0035] The model iteration optimization module is used to update the training set based on the newly added temporal characteristics, vegetation index and measured soil moisture content, and continuously optimize the accuracy of the deep learning model and its adaptability to different slope environments through periodic model training and verification.

[0036] The beneficial effects of the present invention are: 1. By constructing a mycorrhizal network with arbuscular mycorrhizal fungi and ectomycorrhizal fungi as the core, a mycelial fiber network is formed, and the soil fixing ability is enhanced by the adhesion between the fiber network and the soil and the strength of the fiber itself; 2. The stability of soil aggregates is increased by the glomalin secreted by the mycelium, and the soil erosion resistance is further improved; 3. By adding water-retaining agents, the early growth of green vegetation and mycelial networks is ensured while the water retention of backfill soil is improved; 4. Finally, the drone is equipped with a multi-spectral camera and radar for coordinated scanning to quickly and accurately monitor the soil moisture content and maintain the long-term survival of vegetation. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of the slope ecological greening construction method based on soil improvement of the present invention. DETAILED DESCRIPTION

[0038] The following describes in detail specific embodiments of the present invention in conjunction with the accompanying drawings. It should be understood that the specific embodiments provided herein are intended only to illustrate and explain the present invention and are not intended to limit the present invention. It should be noted that many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention may also have other embodiments and variations thereof. Therefore, the scope of protection of the present invention is not limited by the specific embodiments disclosed below.

[0039] Example 1, a slope ecological greening construction method based on soil improvement, specifically includes the following steps:

[0040] S100, selection of plant-fungus combinations, including selecting arbuscular mycorrhizal fungi and ectomycorrhizal fungi as core species, and by selecting fungi that coexist with plants (arbuscular mycorrhizal fungi and ectomycorrhizal fungi), the plants' ability to absorb water and nutrients is enhanced, especially the utilization efficiency of key elements such as phosphorus and nitrogen. The glomalin secreted by arbuscular mycorrhizal fungi increases the stability of soil aggregates and improves soil erosion resistance. Trichoderma and nitrogen-fixing bacteria are auxiliary bacterial communities, which further improve the soil microenvironment, inhibit the reproduction of pathogens, and increase soil fertility through nitrogen fixation. Biochar slow-release granules are used as bacterial carriers for the bacterial system, and herbaceous plants with well-developed root systems that coexist with the bacterial system are selected as slope greening plants. The formation of a mycelial network can stabilize soil structure and enhance the slope's resistance to erosion. The bacterial carrier provides a long-term and stable living environment for microorganisms, preventing the bacterial community from being inactivated due to environmental fluctuations. Herbaceous plants with well-developed root systems work synergistically with the bacterial community to form a stable ecological barrier, reducing the risk of slope degradation due to rain erosion or drought.

[0041] The present embodiment is located in a relatively arid area, and the arbuscular mycorrhizal fungus adopts heteromorphic root cyst mold, which is derived from the arbuscular mycorrhizal fungus developed by Beijing Jinbilai Agricultural Technology Co., Ltd., and its propagule number is ≥70 / mL; the ectomycorrhizal fungus is derived from the ectomycorrhizal fungus composite microbial agent, and is obtained by mixing colored pox puffball: pine lactarius: sticky cap boletus=1:1:1 in a mass ratio, and the Trichoderma adopts the harzianum produced by Zhongyi Keyi (Shandong) Ecological Agriculture Co., Ltd., and the nitrogen-fixing bacteria adopts the round brown nitrogen-fixing bacteria produced by Nanjing Celt Biotechnology Co., Ltd., and the mass ratio of 2g of Jinbilai mycorrhizal fungus, 1g of ectomycorrhizal fungus composite microbial agent, 0.5g of harzianum and 0.5g of round brown nitrogen-fixing bacteria per square meter of slope area is dissolved in 1kg to make a strain suspension, and the strain suspension is sprayed onto biochar slow-release particles to form microbial agent particles; the plant selection is Artemisia ordosica;

[0042] S200, slope protection structure construction, including the construction of a frame-type slope protection structure according to the slope reinforcement design, backfilling the cells enclosed by the frame with planting soil, and providing permeable holes in the slope concrete layer of each cell. After backfilling, the soil is compacted to a density of z. According to Article 5.1.2 of the "Code for Construction of Garden and Green Space Engineering" DB11 / T 1175-2015, the soil density for planting within the surface 1.5m should not exceed 0.80. On the one hand, high soil density is not conducive to vegetation survival, and on the other hand, low soil density is easily lost due to rainfall. In this embodiment, the density z ranges from 0.75 to 0.8.

[0043] S300, constructing a mycelium system, including uniformly distributing n mycelium culture tubes per square meter within the backfill planting soil; each mycelium culture tube is filled with a mycelium matrix having a mass ratio of mushroom dregs to humus soil of 1:2, the moisture content within the tube is controlled at 45%-55%, and a degradable sealing cap is provided on the top of the tube; the number n of mycelium culture tubes per square meter is determined based on the effective coverage area of ​​a single mycelium tube, and the effective coverage area of ​​a single mycelium tube is measured in a laboratory under the same slope, soil permeability, and curing method as the on-site slope construction, and the records are archived; when a sufficient number of archived items are obtained, a mathematical relationship between the coverage area and the slope, soil permeability, and curing conditions is established through regression analysis;

[0044] S400, drone seeding, including preparing seed bombs in a mass ratio of slope greening plant seeds: microbial agent granules: water-retaining agent = 6:3:1, placing the seed bombs in the magazine of a pneumatic ejection device, and then sowing using the pneumatic ejection device mounted on a drone; the seed bombs are sown avoiding the frame-type slope protection structure and culture tubes, and are sown in the middle of adjacent culture tubes in each cell; when the mycelium network expanding from the culture tubes cannot completely cover the area, the plant roots can quickly fill the uncovered area;

[0045] According to step S100, the slope greening plant seeds are Artemisia ordosica seeds, the inoculant particles are biochar slow-release particles that absorb the bacterial suspension, and the water-retaining agent is a highly absorbent resin purchased from Zhengzhou Guangyang Environmental Protection Technology Co., Ltd., which can absorb more than 400 times its own weight in deionized water;

[0046] S500, surface film mulching, involves laying straw-degradable film on the surface of the planting soil and compacting the edges of the film. The straw-degradable film reduces soil moisture evaporation, buffers rainwater erosion and wind erosion, and regulates soil temperature fluctuations through physical covering, providing a stable microenvironment for seed germination and mycelial network expansion. After degradation, the straw material is converted into humus, which directly replenishes the soil's organic matter content, improves soil fertility, and forms a synergistic effect with the microbial agent system to further improve the poor soil on the slope. The degradable film provides physical protection in the early stages. As the mycelial network coverage increases, its degradation rate matches the mycelial growth rhythm, gradually transferring the water-retention and soil-fixing functions to the biological system.

[0047] S600, dynamic maintenance and control, includes regular spraying of nutrient solution in the early stages of maintenance to ensure stable expansion of hyphae until the hyphae network covers more than 80% of the planting soil surface. In the late stages of maintenance, the monitoring system uses microwave-optical multi-source remote sensing technology to collaboratively monitor surface soil moisture, including microwave-optical data acquisition and preprocessing, multi-source data fusion modeling, and real-time monitoring and early warning.

[0048] Microwave-optical data acquisition and preprocessing involves using a drone equipped with an ultra-wideband radar to hover 1m above the surface of the area to be detected to collect the original echo signal of the surface radar, and using a multi-optical slope camera equipped with the drone to obtain multispectral images of the slope area; then, the original echo signal is intercepted according to the radar echo delay formula.

[0049]

[0050] Where Δt is the time difference between the radar transmitting signal and receiving the surface echo, R is the distance between the radar antenna and the surface, V is the propagation speed of the radar wave in the medium, τ is the sampling frequency of the radar, and m is the distance between the radar antenna and the surface. - The number of sampling points; the interception starts from the mth sampling point of the original echo signal, and the intercepted second half of the radar echo signal is used as the effective radar signal; the time domain feature extraction of the effective radar signal is performed, including the calculation of the maximum amplitude, minimum amplitude, rectified mean, variance, root mean square, kurtosis, peak-to-peak value, skewness and peak factor; the surface vegetation index NDVI is calculated based on the multispectral image data,

[0051]

[0052] Where, ρ NR -Near-infrared band of multispectral image, ρ R-Infrared band of multispectral image; then perform maximum and minimum normalization processing on temporal characteristics and vegetation index;

[0053] Multi-source data fusion modeling includes using a generalized regression neural network model optimized by a cosine-optimized particle swarm algorithm as a deep learning model, and using the time domain characteristics of effective radar signals and the vegetation index as inputs to the deep learning model to predict soil moisture. The construction process of the deep learning model includes microwave-optical data acquisition and preprocessing, soil moisture measurement, model training and verification. The soil moisture measurement includes gridding the area to be detected, inserting a soil thermometer and hygrometer into the soil of each grid, reading the soil moisture value of each grid, and taking the average value after removing the maximum and minimum values ​​as the soil moisture content of the area to be detected. Model training and verification includes using the time domain characteristics, vegetation index and measured soil moisture of the area to be detected obtained through microwave-optical data acquisition and preprocessing as a set of data, and also obtaining k groups of data from different areas to be detected to form a data set. The data set is divided into a training set and a validation set to train and verify the deep learning model. During the deep learning model training process, the cosine-optimized particle swarm algorithm is used to optimize the smoothing factor of the generalized regression neural network model. The prediction accuracy of the deep learning model is measured by the determination coefficient R. 2 , root mean square error RMSE and accuracy ACC are used for evaluation.

[0054]

[0055] Where y i - Measured value of soil moisture, y′ i -Predicted soil moisture content, y e -The mean of the measured values ​​of soil moisture content;

[0056] Real-time monitoring and early warning include regularly updating data through collaborative scanning of the multispectral camera and radar carried by drones, inputting deep learning models to invert the soil moisture content of the slopes, and reminding maintenance personnel to irrigate when the soil moisture content is lower than the preset warning value.

[0057] Embodiment 2, a slope ecological greening monitoring system based on soil improvement, includes a multi-source data acquisition module, a data preprocessing module, a deep learning model prediction module, an early warning module and a model iterative optimization module;

[0058] The multi-source data acquisition module is used to receive the original echo signals and multispectral image data collected by the UAV-mounted ultra-wideband radar and multispectral camera through the network, as well as the coordinates and timestamps corresponding to the UAV data collection, and save the original echo signals, multispectral image data, coordinates and timestamps in the corresponding slope project;

[0059] The data preprocessing module is used to extract the time domain features of the effective radar signal, calculate the vegetation index based on the multispectral image data, and perform maximum and minimum normalization processing on the time domain features and vegetation index;

[0060] The deep learning model prediction module is used to input the preprocessed temporal features and vegetation index into a trained generalized regression neural network model optimized by the cosine-optimized particle swarm algorithm, and output the soil moisture content;

[0061] The early warning module is used to regularly update microwave-optical data through drone collaborative scanning. The data preprocessing module and deep learning model prediction module invert the slope soil moisture distribution. When the moisture content is detected to be lower than the preset warning value, a warning signal is triggered to remind maintenance personnel to implement irrigation.

[0062] The model iteration optimization module is used to update the training set based on newly added temporal features, vegetation index, and measured soil moisture content. Through periodic model training and verification, the accuracy of the deep learning model and its adaptability to different slope environments are continuously optimized.

[0063] In summary, the slope ecological greening construction method and monitoring system based on soil improvement of the present invention have the characteristics of enhancing soil consolidation capacity, improving soil erosion resistance and water retention in the field of slope greening construction technology.

[0064] It should be understood that the above embodiments are one or more embodiments of the present invention, and there are many other embodiments and variations thereof based on the present invention; the variations and modifications made by ordinary technicians in this industry through the present invention without making groundbreaking innovations all fall within the scope of protection of the present invention.

Claims

1. A slope ecological greening construction method based on soil improvement, characterized in that: The following steps are involved: S100, plant-fungus combination selection, using arbuscular mycorrhizal fungi and ectomycorrhizal fungi as core species, Trichoderma and nitrogen-fixing bacteria as auxiliary bacteria, and biochar slow-release granules as the carrier of the bacterial agent. Herbs with well-developed root systems that coexist with the bacterial system are selected as slope greening plants; S200, slope protection structure construction: according to the slope reinforcement design, frame type slope protection structure construction is carried out. Planting soil is backfilled into the cells surrounded by the frame, and permeable holes are set in the slope concrete layer of each cell. After backfilling, the concrete layer is compacted to a density of z; S300, constructing a mycelium system, evenly distributing n mycelium culture tubes per square meter in the backfilled planting soil; S400, drone sowing, using slope greening plant seeds, fungal agents and water-retaining agents to prepare seed bombs, placing the seed bombs in the magazine of the air pressure ejection device, and using the drone equipped with the air pressure ejection device for sowing; S500, surface film covering, including laying straw degradable film on the surface of the planting soil and compacting the periphery of the straw degradable film; S600, dynamic maintenance and control. In the early stage of maintenance, nutrient solution is sprayed regularly to ensure the stable expansion of mycelium until the mycelium network covers more than 80% of the planting soil surface. In the later stage of maintenance, the surface soil moisture content is collaboratively monitored based on the monitoring system's microwave-optical multi-source remote sensing technology. When the soil moisture content reaches the warning value, maintenance personnel are reminded to perform watering maintenance.

2. The slope ecological greening construction method based on soil improvement according to claim 1 is characterized in that: In step S300, each of the mycelium culture tubes is filled with a mycelium matrix with a mass ratio of mushroom residue to humus soil of 1:2, the moisture content in the tube is controlled at 45%-55%, and a degradable sealing cover is provided on the top of the tube.

3. The slope ecological greening construction method based on soil improvement according to claim 1 is characterized in that: In step S300, the number n of mycelium culture tubes arranged per square meter is determined based on the effective coverage area of ​​a single mycelium tube. The effective coverage area of ​​a single mycelium tube is measured and recorded in a laboratory under the same slope, soil permeability, and maintenance conditions as those of the on-site slope construction.

4. The slope ecological greening construction method based on soil improvement according to claim 1 is characterized in that: In step S400, the seed bombs are sown avoiding the frame-type slope protection structure and the culture tubes, and are sown in the middle of the adjacent culture tubes in each cell.

5. The slope ecological greening construction method based on soil improvement according to claim 1 is characterized in that: In step S600, the microwave-optical multi-source remote sensing technology collaboratively monitors surface soil moisture, including microwave-optical data acquisition and preprocessing, multi-source data fusion modeling, and real-time monitoring and early warning; The multi-source data fusion modeling includes using a generalized regression neural network model optimized by a cosine optimized particle swarm algorithm as a deep learning model, and using the time domain characteristics of effective radar signals and vegetation index as inputs of the deep learning model to predict soil moisture content; The real-time monitoring and early warning system includes regularly updating data through collaborative scanning of the multispectral camera and radar carried by the drone, inputting the deep learning model into the inversion of the slope soil moisture content, and reminding maintenance personnel to irrigate when the soil moisture content is lower than the preset warning value.

6. The slope ecological greening construction method based on soil improvement according to claim 5 is characterized in that: The construction process of the deep learning model includes microwave-optical data acquisition and preprocessing, soil moisture measurement, model training and verification; The soil moisture content measurement includes gridding the area to be tested, inserting a soil thermometer and hygrometer into the soil of each grid, reading the soil moisture content value of each grid, and taking the average value after removing the maximum and minimum values ​​as the soil moisture content of the area to be tested; The model training and verification includes taking the time domain characteristics, vegetation index and measured soil moisture content obtained by microwave-optical data acquisition and preprocessing of the area to be detected as a set of data, and also obtaining k groups of data from different areas to be detected to form a data set, which is divided into a training set and a validation set, and the deep learning model is trained and verified. During the deep learning model training process, the cosine optimized particle swarm algorithm is used to optimize the smoothing factor of the generalized regression neural network model. The prediction accuracy of the deep learning model is measured by the determination coefficient R 2 , root mean square error RMSE and accuracy ACC are used for evaluation. Where y i - Measured value of soil moisture, y′ i -Predicted soil moisture content, y e -The mean of the measured values ​​of soil moisture content.

7. The slope ecological greening construction method based on soil improvement according to claim 5 or 6 is characterized in that: The microwave-optical data acquisition and preprocessing includes using a drone equipped with an ultra-wideband radar to hover at an altitude of 1 meter above the surface of the area to be detected to collect the original echo signal of the surface radar, and using a multi-light slope camera equipped with the drone to obtain multispectral images of the slope area; then, the original echo signal is intercepted according to the radar echo delay formula. Where Δt is the time difference between the radar transmitting signal and receiving the surface echo, R is the distance between the radar antenna and the surface, V is the propagation speed of the radar wave in the medium, τ is the sampling frequency of the radar, and m is the distance between the radar antenna and the surface. - The number of sampling points; the interception starts from the mth sampling point of the original echo signal, and the intercepted second half of the radar echo signal is used as the effective radar signal; the time domain feature extraction of the effective radar signal is performed, including the calculation of the maximum amplitude, minimum amplitude, rectified mean, variance, root mean square, kurtosis, peak-to-peak value, skewness and peak factor; the surface vegetation index NDVI is calculated based on the multispectral image data, Where, ρ NR -Near-infrared band of multispectral image, ρ R -Infrared band of multispectral image; then the temporal characteristics and vegetation index are normalized to the maximum and minimum values.

8. The slope ecological greening construction method based on soil improvement according to claim 1 is characterized in that: In step S200, the density z ranges from 0.75 to 0.

8.

9. A soil improvement-based slope ecological greening monitoring system, applied to the soil improvement-based slope ecological greening construction method according to any one of claims 1 to 8, characterized in that: It includes multi-source data acquisition module, data preprocessing module, deep learning model prediction module, early warning module and model iteration optimization module; The multi-source data acquisition module is used to receive the original echo signal and multispectral image data collected by the ultra-wideband radar and multispectral camera carried by the drone through the network, as well as the coordinates and timestamps corresponding to the drone collecting the data, and save the original echo signal, multispectral image data, coordinates and timestamps in the corresponding slope project; The data preprocessing module is used to extract time domain features of effective radar signals, calculate vegetation index based on multispectral image data, and perform maximum and minimum normalization processing on the time domain features and vegetation index; The deep learning model prediction module is used to input the preprocessed temporal features and vegetation index into a trained generalized regression neural network model optimized by a cosine optimized particle swarm algorithm, and output the soil moisture content; The early warning module is used to regularly update microwave-optical data through drone collaborative scanning, invert the slope soil moisture distribution through the data preprocessing module and the deep learning model prediction module, and trigger an early warning signal when the moisture content is detected to be lower than the preset warning value, reminding maintenance personnel to implement irrigation; The model iteration optimization module is used to update the training set based on the newly added temporal characteristics, vegetation index and measured soil moisture content, and continuously optimize the accuracy of the deep learning model and its adaptability to different slope environments through periodic model training and verification.

Images