Tree root system health intelligent early warning system and method integrated with multi-source sensing

Abstract

The invention provides a tree root system health intelligent early warning system and method fused with multi-source sensing, and the system comprises a multi-source sensing module which is composed of four types of heterogeneous sensors, and the four types of heterogeneous sensors form a distributed monitoring network. The system specifically comprises a ground penetrating radar, a dielectric property sensor array, a gas detection unit module and a geomagnetic anomaly detection module, the ground penetrating radar, the gas detection unit module and the geomagnetic anomaly detection module are all mounted in the cylindrical titanium alloy shell; the dielectric property sensor array is connected with the cylindrical titanium alloy shell through a cable; the dielectric property sensor array is vertically buried in a main distribution area of a tree root system; the problems that a traditional single sensor is single in monitoring dimension and high in false alarm rate are solved; the four types of heterogeneous sensors work cooperatively, multi-source data are fused in combination with the improved D-S evidence theory, and the false alarm rate and the omission ratio are effectively reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet of Things, in particular to a tree root health intelligent early warning system and method fusing multi-source sensing. BACKGROUND

[0002] In the process of urbanization, urban green belts, as an important part of the ecological environment, not only have the functions of purifying air and regulating climate, but also play a key role in beautifying the city landscape and improving the quality of life of residents. However, the health status of the roots of tall trees in the green belt is often neglected. As the underground support system of trees, the health status of the roots directly determines the resistance of the trees to lodging in extreme weather. Especially in areas where typhoons occur frequently, if the roots of trees are severely rotted due to diseases, when typhoon weather comes, large trees are prone to instability of root base and whole tree lodging, which not only causes damage to surrounding infrastructure, but also may endanger the safety of pedestrians, causing serious casualties and property losses.

[0003] Currently, the health monitoring of urban green tree roots mainly relies on manual inspection technology, which uses visual observation of crown shape changes combined with local drilling sampling methods. Specifically, technicians observe the degree of leaf yellowing, crown shrinkage and other apparent characteristics, and combine manual drilling to extract soil samples for laboratory physicochemical analysis (pH value, water content, etc.). However, this technology has significant drawbacks: in the early stage of root rot, the surface features are not obvious, and technicians cannot identify it in time through existing means, and the traditional experimental data shows that the detection lag period is long, often leading to untimely disease discovery, missing the best prevention opportunity; in addition, drilling sampling causes mechanical damage up to 35 cm in diameter, has a high secondary damage rate, and adversely affects the normal growth of trees, and the drilling sampling range is small, often failing to accurately reach the root lesion position of the tree, making it difficult to fully grasp the health status of the roots; more importantly, the existing technology cannot scientifically and accurately judge the overall health status of the tree roots. SUMMARY

[0004] The technical problem to be solved by the present application is to overcome the existing defects and provide a tree root health intelligent early warning system and method fusing multi-source sensing, which can effectively solve the problems in the background art.

[0005] To achieve the above-mentioned purpose, the present application proposes:

[0006] A tree root health intelligent early warning system fusing multi-source sensing, comprising a multi-source sensing module, wherein the multi-source sensing module is composed of four types of heterogeneous sensors, including a ground penetrating radar, a dielectric property sensor array, a gas detection unit module and a geomagnetic anomaly detection module.

[0007] The ground penetrating radar and the geomagnetic anomaly detection module are both installed in a cylindrical titanium alloy shell; the gas detection unit module and the dielectric property sensor array are both connected to the cylindrical titanium alloy shell through a cable;

[0008] The dielectric property sensor array is vertically embedded in the main distribution area of the tree root system;

[0009] The gas detection unit module includes multiple groups of MEMS gas sensors, which are arranged in a ring around the trunk for detecting the concentrations of CO2, CH4 and VOCs;

[0010] The geomagnetic anomaly detection module includes a three-axis magnetoresistance sensor, which is embedded with the cylindrical titanium alloy shell near the trunk;

[0011] It also includes a data processing unit, which includes a multi-modal data fusion engine, a health assessment model and a warning module model.

[0012] Preferably, the four types of heterogeneous sensors form a distributed monitoring network.

[0013] Preferably, it also includes a solar power supply unit.

[0014] A tree root system health intelligent early warning method fusing multi-source sensing, including the above-mentioned early warning system, comprising the following steps:

[0015] Step S1, multi-source data acquisition and preprocessing:

[0016] The ground penetrating radar is used to scan the root system area of the tree and generate imaging results;

[0017] The dielectric property sensor array real-time collects the soil complex dielectric constant real part ∈', which is used for subsequent dielectric stability calculation;

[0018] The gas detection unit module detects the concentrations of CO2, CH4 and VOCs, and calculates the gas diffusion flux gradient:

[0019]

[0020] The geomagnetic anomaly detection module collects geomagnetic field signals through a three-axis magnetoresistance sensor, and analyzes the frequency spectrum characteristics through fast Fourier transform:

[0021]

[0022] S2, multi-modal data fusion and decay probability determination:

[0023] The multi-modal data fusion engine of the data processing unit is used to determine the root decay probability of the tree;

[0024] The multi-modal data fusion engine of the data processing unit adopts an improved D-S evidence theory, specifically:

[0025] S21, define basic probability assignment function BPA:

[0026]

[0027] Wherein, A / represents the detection result of the i-th sensor (GPR, dielectric, gas, geomagnetic), w / is the sensor weight (w 探地雷达 = 0.3, w 介电 = 0.25, w 气体 = 0.2, w 地磁 = 0.25), S / is the sensor confidence;

[0028] S22, fuse multi-source evidence by synthesis rule:

[0029]

[0030] for comprehensive determination of root decay probability.

[0031] S3, root health index calculation:

[0032] Based on dielectric stability, strain uniformity, and gas diffusion entropy, an RHI model is constructed;

[0033] The calculation formula is: RHI = a · dielectric stability + β · strain uniformity + γ · gas diffusion entropy;

[0034] Wherein: a = 0.4, β = 0.3, γ = 0.3 are weight coefficients;

[0035] Dielectric stability is calculated by dielectric constant variance:

[0036] Where σ is the dielectric constant variance, μ is the mean

[0037] Strain uniformity is calculated from fiber optic strain sensor data:

[0038]

[0039] Wherein, ∈ / is the single-point strain, is the average strain;

[0040] Gas diffusion entropy is based on CO2 concentration gradient:

[0041]

[0042] S4, anti-wind fall mechanical model early warning:

[0043] Calculate the critical wind load and real-time root bearing capacity, trigger red early warning condition:

[0044] The critical wind load of the anti-wind fall mechanical model is calculated as: F B = 0.5·ρ·v + ·C C ·A 有效 ;

[0045] Where: ρ = 1.225 kg / m D Air density; v is the real-time wind speed; C C = 0.8 is the drag coefficient;

[0046] A 有效 is the effective anchoring area of the root system, calculated as: When the real-time root bearing capacity F $ <1.2F B , trigger red early warning;

[0047] S5, decision output and intervention instruction:

[0048] Generate differentiated decisions according to the fusion results:

[0049] When the rotten area of the tree root is greater than 40%, the pneumatic grouting reinforcement is started;

[0050] When the rotten depth is greater than 2m, the micro pile support is started;

[0051] When the root bearing capacity reduction rate is greater than 5% / week, the emergency felling instruction is generated.

[0052] Preferably, when And the ratio of CO2 / CH4 is greater than 20, it is determined that there is a fungal infection.

[0053] Preferably, when the energy ratio of the 0.1-2Hz frequency band in the spectrum feature is greater than 30%, it is determined that there is a root cavity.

[0054] Preferably, the ground penetrating radar adopts a dual-frequency antenna design, and the working frequencies are 500MHz and 1.5GHz, which are respectively used for shallow and deep root scanning, and the signal processing adopts the reverse time migration algorithm.

[0055] Preferably, a multi-protocol communication module is integrated in the cylindrical titanium alloy shell, supporting LoRaWAN, NB-IoT and Beidou short message, and the data packet encryption adopts the AES-256 algorithm.

[0056] Compared with the prior art, the beneficial effects of the present application are:

[0057] 1. The application solves the problem of high false alarm rate and inability to distinguish between normal transpiration and pathological water accumulation due to the single monitoring dimension of traditional single sensor; four types of heterogeneous sensors, including ground penetrating radar (GPR), dielectric property sensor, gas detection unit and geomagnetic anomaly detection module, work cooperatively, and combine improved D-S evidence theory to fuse multi-source data, effectively reducing false alarm rate and missed detection rate.

[0058] 2. The dielectric property sensor array is arranged radially, combined with GPR dual-frequency scanning, to realize three-dimensional profile reconstruction of the root rot area in depth; the gas detection unit triggers early warning when the rot area is less than 5% through CO2 / CH4 ratio and VOCs concentration, with long early warning time limit.

[0059] 3. The application solves the core problems of traditional technology detection lag, high false alarm rate and strong destructive power through multi-source sensor fusion, three-dimensional health assessment, dynamic risk early warning and ecological friendly design, significantly improves the safety management level of urban green trees, and provides reliable technical support for ecological protection of smart city. BRIEF DESCRIPTION OF DRAWINGS

[0060] Fig. 1 The flowchart of the application;

[0061] Fig. 2 The device structure diagram of the application;

[0062] Fig. 3 The internal structure diagram of the cylindrical titanium alloy shell.

[0063] In the figure: 1 solar cell panel, 2 cylindrical titanium alloy shell, 3 dielectric property sensor, 4 gas detection unit module, 5 geomagnetic anomaly detection module, 6 ground penetrating radar. DETAILED DESCRIPTION

[0064] The technical solutions in the embodiments of the application will be described clearly and completely below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0065] Please refer to Figs. 1-3 The application provides the following embodiments:

[0066] A tree root health intelligent early warning system fusing multi-source sensors, comprising a multi-source sensing module, which is composed of four types of heterogeneous sensors, including ground penetrating radar 6, dielectric property sensor array 3, gas detection unit module 4 and geomagnetic anomaly detection module 5.

[0067] The ground penetrating radar 6 and the geomagnetic anomaly detection module 5 are both installed inside the cylindrical titanium alloy shell 2; the gas detection unit module 4 and the dielectric property sensor array 3 are both connected to the cylindrical titanium alloy shell 2 through a cable.

[0068] The upper surface of the cylindrical titanium alloy shell 2 is installed with a solar panel, and the inside of the cylindrical titanium alloy shell 2 is installed with an energy storage lithium battery pack.

[0069] The number of the dielectric property sensor array 3 is four, and the four dielectric property sensor arrays 3 are vertically buried in the main distribution area of the tree root system.

[0070] The gas detection unit module 4 includes four groups of MEMS gas sensors, which are arranged in a ring around the trunk for detecting the concentrations of CO2, CH4 and VOCs.

[0071] The geomagnetic anomaly detection module 5 includes a three-axis magnetoresistance sensor, which is pre-buried around the trunk.

[0072] It also includes a data processing unit, which includes a multi-modal data fusion engine, a health assessment model and a warning module model.

[0073] Further, four types of heterogeneous sensors constitute a distributed monitoring network.

[0074] Further, it also includes a solar power supply unit.

[0075] A tree root system health intelligent early warning method fusing multi-source sensors, including the above-mentioned early warning system, the specific steps are as follows:

[0076] Step S1, multi-source data acquisition and preprocessing:

[0077] The ground penetrating radar is used to scan the root system area of the tree and generate imaging results;

[0078] The dielectric property sensor array real-time collects the soil complex dielectric constant real part ∈', which is used for subsequent dielectric stability calculation;

[0079] The gas detection unit module detects the concentrations of CO2, CH4 and VOCs, and calculates the gas diffusion flux gradient:

[0080]

[0081] The geomagnetic anomaly detection module collects geomagnetic field signals through a three-axis magnetoresistance sensor, and analyzes the frequency spectrum characteristics through fast Fourier transform:

[0082]

[0083] S2, multi-modal data fusion and decay probability determination:

[0084] The root rot probability of the tree is judged by a multi-modal data fusion engine of a data processing unit;

[0085] The multi-modal data fusion engine of the data processing unit adopts an improved D-S evidence theory, specifically:

[0086] S21, a basic probability assignment function BPA is defined:

[0087]

[0088] Wherein, A / represents the detection result of the i-th sensor (GPR, dielectric, gas, geomagnetic), w / is the sensor weight (w 探地雷达 =0.3, w 介电 =0.25, w 气体 =0.2, w 地磁 =0.25), S / is the sensor confidence;

[0089] S22, multi-source evidence is fused by a synthesis rule:

[0090]

[0091] for comprehensive determination of the root rot probability.

[0092] S3, root health index calculation:

[0093] Based on dielectric stability, strain uniformity, and gas diffusion entropy, an RHI model is constructed;

[0094] The calculation formula is: RHI=α·dielectric stability+β·strain uniformity+γ·gas diffusion entropy;

[0095] Wherein: α=0.4, β=0.3, γ=0.3 are weight coefficients;

[0096] Dielectric stability is calculated by dielectric constant variance:

[0097] Wherein σ is the dielectric constant variance, μ is the mean

[0098] Strain uniformity is calculated by fiber optic strain sensor data:

[0099]

[0100] Wherein, ∈ / is the single-point strain, is the average strain;

[0101] Gas diffusion entropy is based on CO2, concentration gradient:

[0102]

[0103] S4, anti-windfall mechanical model early warning:

[0104] Calculate the critical wind load and real-time root bearing capacity, trigger red warning condition:

[0105] The critical wind load of the anti-windfall mechanical model is calculated as: B = 0.5·ρ·v + ·C C ·A 有效 ;

[0106] Where: ρ = 1.225 kg / m D Air density; v is the real-time wind speed; C C = 0.8 is the drag coefficient;

[0107] A 有效 is the effective anchoring area of the root system, calculated as: When the real-time root bearing capacity F $ <1.2F B , trigger red warning;

[0108] S5, decision output and intervention instruction:

[0109] Generate differentiated decisions according to the fusion results:

[0110] When the rotten area of tree roots is greater than 40%, enable pneumatic grouting reinforcement;

[0111] When the depth of decay is greater than 2m, enable micro-pile support;

[0112] When the root bearing capacity decline rate is greater than 5% / week, generate emergency felling instructions.

[0113] Further, when And the ratio of CO2 / CH4 is >20, it is determined to be fungal infection.

[0114] Further, the energy ratio of the 0.1-2Hz frequency band in the spectral feature is >30%, indicating the presence of root cavities.

[0115] Further, the ground penetrating radar adopts a dual-frequency antenna design with working frequencies of 500MHz and 1.5GHz for shallow and deep root scanning respectively, and the signal processing uses the reverse time migration algorithm.

[0116] Further, the cylindrical titanium alloy shell integrates a multi-protocol communication module, supporting LoRaWAN, NB-IoT and Beidou short message, and the data packet encryption uses the AES-256 algorithm.

[0117] Specific implementation case: Urban street tree root health monitoring and typhoon early warning:

[0118] 1. System deployment;

[0119] Target trees: Select 10-year-old or older banyan trees along the main urban roads as the research object.

[0120] Deployment steps:

[0121] At 1m from the trunk, a 16cm diameter hollow drill is used to drill vertically to a depth of 3.2m, and the inner wall of the drill hole is sprayed with chitosan biological glue.

[0122] Send titanium alloy shell 2 into the drill hole, backfill soil and compact, and the top of the shell is exposed 5cm above the ground to install solar panels 1.

[0123] 2. Sensor deployment:

[0124] Dielectric property sensor: 4 groups of sensors are arranged radially along the trunk, with a group spacing of 40cm, and each group has a buried depth node of 0.5m, 1.0m, 1.5m, and 2.0m, connected to titanium alloy shell 2 through a cable.

[0125] Gas detection unit: 12 groups of MEMS sensors are arranged in a ring around the trunk at radii of 0.3m, 0.7m, and 1.2m, with 4 groups per layer, and wind speed sensors are installed synchronously.

[0126] 3. System activation:

[0127] After starting solar power supply, the geomagnetic module is left for 24 hours to complete background noise calibration.

[0128] Multi-source sensing data:

[0129] Dielectric property sensor: soil complex permittivity is collected every 10 minutes to generate dielectric stability parameters.

[0130] Gas detection unit: CO + concentration in the inner circle (0.3m) is 1200ppm, and in the outer circle (1.2m) is 400ppm, with a gradient CO + / CH F =25.

[0131] Geomagnetic module: 35% of the energy in the 0.1-2Hz frequency band is monitored, and the geomagnetic field strength changes ΔB=8nT.

[0132] Fiber strain: a strain rate of 12% / h is detected at 3 consecutive grating points at a depth of 1.2m.

[0133] Multimodal data fusion:

[0134] D-S evidence theory fusion: input sensor weights (w GHI = 0.3, w 介电 = 0.25, w 气体 = 0.2, w 地磁 = 0.25) and confidence:

[0135] S 介电 = 0.9, S 气体 = 0.85;

[0136] Calculate the rotten probability P 腐烂 = 0.82 after fusion; the threshold is 0.7.

[0137] Health assessment model:

[0138] Calculate the RHI index:

[0139] RHI = 0.4 x 0.68 + 0.3 x 0.12 + 0.3 x 1.45 = 0.62;

[0140] Where the dielectric stability is 0.68, the strain uniformity is 0.12, and the gas diffusion entropy is 1.45.

[0141] 3. Early warning output and response wind resistance mechanics model:

[0142] When the real-time wind speed is v = 15 m / s, calculate the critical wind load:

[0143] F B = 0.5 x 1.225 x 15 + x 0.8 x 2.4 = 264.6 kN

[0144] Effective anchorage area A 有效 = 2.4 m + ;

[0145] Real-time root bearing capacity F $ = 300 kN, F $ > 1.2 F B , no red warning triggered.

[0146] Four-level early warning mechanism:

[0147] RHI = 0.62 (yellow warning), determine the root damage area of about 20%, the system generates instructions: complete aerodynamic grouting reinforcement within 3 days, grouting point coordinates (X = 1.2 m, Y = 0.8 m, Z = 1.0-1.5 m).

[0148] Through multi-source sensing fusion and dynamic model early warning, the embodiment realizes precise monitoring and risk control of urban street tree root health, and verifies the significant advantages of the system in early decay detection, typhoon risk quantification and ecological friendliness.

[0149] While embodiments of the application have been shown and described, it is to be understood that the embodiments described are merely divergences of the principles and specific embodiments of the application and that numerous modifications, changes, substitutions, and alterations can be made thereto without departing from the spirit and scope of the application as defined by the appended claims and their equivalents.

Claims

1. An intelligent early warning system for tree root health integrating multi-source sensing, characterized by: It includes a multi-source sensing module, which is composed of four types of heterogeneous sensors: ground penetrating radar, dielectric property sensor array, gas detection unit module and geomagnetic anomaly detection module; The ground penetrating radar and geomagnetic anomaly detection module are both installed in a cylindrical titanium alloy shell; the gas detection unit module and the dielectric property sensor array are both connected to the cylindrical titanium alloy shell through cables; The dielectric property sensor array is vertically buried in the main distribution area of ​​tree roots; The gas detection unit module includes multiple groups of MEMS gas sensors, which are arranged in a ring around the trunk to detect the concentration of CO2, CH4 and VOCs; The geomagnetic anomaly detection module includes a three-axis magnetoresistive sensor, which is pre-buried near the tree trunk along with a cylindrical titanium alloy shell; It also includes a data processing unit, which includes a multimodal data fusion engine, a health assessment model and an early warning module model.

2. The tree root health intelligent early warning system integrating multi-source sensing according to claim 1 is characterized by: Four types of heterogeneous sensors constitute a distributed monitoring network.

3. The tree root health intelligent early warning system integrating multi-source sensing according to claim 1 is characterized by: A solar powered unit is also included.

4. An intelligent early warning method for tree root health integrating multi-source sensing, comprising the early warning system according to any one of claims 1 to 3, characterized in that: The steps include: Step S1, multi-source data collection and preprocessing: Use ground penetrating radar to scan the root area of ​​the tree and generate imaging results; The dielectric property sensor array collects the real part of the soil complex dielectric constant ∈ ′ , used for subsequent dielectric stability calculations; The gas detection unit module detects the concentrations of CO2, CH4 and VOCs, and calculates the gas diffusion flux gradient: The geomagnetic anomaly detection module collects geomagnetic field signals through a three-axis magnetoresistive sensor and analyzes the spectrum characteristics through fast Fourier transform: S2, multimodal data fusion and decay probability determination: The multimodal data fusion engine of the data processing unit is used to determine the probability of root decay of trees; The multimodal data fusion engine of the data processing unit adopts the improved DS evidence theory, specifically: S21, define the basic probability distribution function BPA: Among them, A / represents the detection result of the i-th type of sensor (GPR, dielectric, gas, geomagnetic), w / is the sensor weight (w 探地雷达 =0.3, w 介电 =0.25, w 气体 =0.2, w 地磁 =0.25), S / is the sensor confidence; S22, integrating multi-source evidence through synthesis rules: Used to comprehensively determine the probability of root rot. S3, root health index calculation: The RHI model is constructed based on dielectric stability, strain uniformity, and gas diffusion entropy; The calculation formula is: RHI = α·dielectric stability + β·strain uniformity + γ·gas diffusion entropy; Among them: α = 0.4, β = 0.3, γ = 0.3 are weight coefficients; Dielectric stability is calculated by the dielectric constant variance: Where σ is the dielectric constant variance and μ is the mean Strain uniformity is calculated from the optical fiber strain sensor data: Among them, ∈ / is the single point strain, is the average strain; Gas diffusion entropy based on CO2 concentration gradient: S4, wind-resistant mechanical model warning: Calculate critical wind loads and real-time root bearing capacity to trigger red alert conditions: The critical wind load of the wind-resistant mechanical model is calculated as: F B =0.5·ρ·v + ·C C ·A 有效 ; Where: ρ = 1.225 kg / m D is the air density; v is the real-time wind speed; C C =0.8 is the drag coefficient; A 有效 is the effective anchorage area of ​​the root system, which is calculated as: When the real-time root bearing capacity F $ <1.2F B A red alert is triggered when S5, decision output and intervention instructions: Generate differentiated decisions based on fusion results: When the root decay area of ​​a tree is greater than 40%, pneumatic grouting reinforcement is used; When the decay depth is greater than 2m, micro pile support is used; When the root carrying capacity decline rate is greater than 5% / week, an emergency felling order is generated.

5. The tree root health intelligent early warning method integrating multi-source sensing according to claim 4 is characterized by: when When the CO2 / CH4 ratio is >20, it is determined to be a fungal infection.

6. The tree root health intelligent early warning method integrating multi-source sensing according to claim 4 is characterized by: When the energy proportion of the 0.1-2 Hz frequency band in the spectral characteristics is greater than 30%, it is determined that a root cavity exists.

7. The tree root health intelligent early warning method integrating multi-source sensing according to claim 4 is characterized by: The ground penetrating radar adopts a dual-frequency antenna design with operating frequencies of 500 MHz and 1.5 GHz, which are used for shallow and deep root system scanning respectively. The signal processing adopts the reverse time migration algorithm.

8. The tree root health intelligent early warning method integrating multi-source sensing according to claim 4 is characterized by: The cylindrical titanium alloy shell integrates a multi-protocol communication module that supports LoRaWAN, NB-IoT and Beidou short messages, and the data packet encryption uses the AES-256 algorithm.

Images