Unmanned aerial vehicle borne p-wave band microwave and multispectral combined observation method and device for complex terrain conditions

Abstract

This invention relates to the field of remote sensing technology, and in particular provides a method and apparatus for UAV-borne P-band microwave and multispectral joint observation for complex terrain conditions. The method includes selecting a high-priority area as the basic flight path for the UAV, separating sub-paths from the basic path to survey areas with lower priority; the high-priority areas are the basic path, and the low-priority areas are the sub-paths; a multispectral camera collects the reflectance characteristics of vegetation across multiple spectral bands, and a P-band microwave sensor penetrates the vegetation to obtain terrain information beneath it; based on the fused data, a digital terrain model of the surveyed area is constructed, the relationship between vegetation distribution and terrain conditions is analyzed, the model simulates the terrain conditions, and an evaluation report is generated based on the terrain conditions. The apparatus includes a priority setting module, a decision fusion module, and a report output module. This invention improves the efficiency of terrain surveying under complex terrain conditions.

Description

Technical Field

[0001] This invention relates to the field of remote sensing technology, and in particular to an unmanned aerial vehicle (UAV)-borne P-band microwave and multispectral joint observation method and device for complex terrain conditions. Background Technology

[0002] In complex terrain environments (such as jungles and mountains), single optical remote sensing is easily obstructed by vegetation (with a missed detection rate of up to 40%), while traditional microwave remote sensing, due to wavelength limitations (such as the X-band), can only acquire shallow surface information. A 2015 NASA study showed that terrain modeling using only optical data had an elevation error of more than 5 meters, making it difficult to meet the needs of geological disaster early warning.

[0003] Prior art 1, Chinese Patent Application No. 2025101993185, discloses a method, device, equipment, and medium for UAV flight path layout in complex terrain, relating to the field of UAV aerial photogrammetry technology. The method includes: preliminary regional division of the terrain area to be surveyed, determining multiple regional units; clustering the regional units based on terrain observation data to obtain survey sub-regions; determining the survey priority corresponding to each survey sub-region, and matching the basic flight path corresponding to the UAV according to the survey priority; splicing the different types of basic flight paths corresponding to each survey sub-region, and optimizing and adjusting the spliced ​​basic flight paths to obtain the target UAV flight path for terrain surveying of the terrain area to be surveyed. Although this method enables the UAV flight path planning to adapt to various types of landform features, improves the completeness and accuracy of aerial survey data, and ensures the safety and efficiency of the UAV, it suffers from problems such as unclear preliminary regional division, ambiguous definition of terrain observation data, and lack of a method for determining survey priorities.

[0004] Prior art two, Chinese patent application number 2024118616670, discloses a method and system for geological disaster monitoring based on unmanned aerial vehicles (UAVs), relating to the field of geological disaster monitoring technology. The UAV-based geological disaster monitoring method includes: capturing initial image data and current image data of the area to be monitored using a UAV; preprocessing the initial image data to obtain a baseline image to be analyzed, and preprocessing the current image data to obtain a current image to be analyzed; extracting a baseline model and a current model; fusing the baseline model and the current model to obtain a monitoring model; analyzing the monitoring model to assess the geological disaster situation in the area to be monitored and obtaining a geological disaster assessment result; and outputting a geological disaster assessment report based on the geological disaster assessment result. While this method can effectively overcome the limitations of complex terrain and harsh working environments, and can effectively improve the safety of patrol personnel and the efficiency and accuracy of patrols, it suffers from problems such as unclear preprocessing methods and ambiguous model extraction methods.

[0005] Prior art three, Chinese patent application number: 2024116980179, discloses a ground deformation monitoring method and device based on unmanned aerial vehicles (UAVs), including: arranging several monitoring stations in the area to be monitored; placing the UAV on each monitoring station to obtain the initial coordinates of each monitoring station; designing the flight path of the UAV according to the location of the monitoring stations; the UAV sequentially landing at the same relative position of each monitoring station according to the flight path, and collecting the spatial coordinates of each monitoring station; the relative same position is the position where the UAV obtains the initial coordinates of each monitoring station; comparing the spatial coordinates and initial coordinates of each monitoring station to obtain the ground deformation of each monitoring station. This invention monitors the required locations and can be flexibly arranged according to various complex terrains or dangerous areas, reducing the complexity of data processing while ensuring the accuracy of data acquisition, and the collected data can be processed in a timely manner. However, it has the problems of not improving the accuracy of initial coordinate acquisition and lacking a flight path design method.

[0006] Current technologies 1, 2, and 3 suffer from issues such as insufficient accuracy in initial coordinate acquisition, ambiguous model extraction methods, and unclear preliminary region delineation. Therefore, this invention provides a method and apparatus for unmanned aerial vehicle (UAV)-borne P-band microwave and multispectral joint observation in complex terrain conditions. Summary of the Invention

[0007] To achieve the above objectives, the present invention adopts the following technical solution: One aspect of the present invention provides an unmanned aerial vehicle (UAV)-borne P-band microwave and multispectral joint observation method for complex terrain conditions, comprising the following steps: The area with complex terrain conditions that needs to be surveyed is divided into regions based on the complexity of the terrain in each region, and the priority is determined according to the complexity of the terrain. The region with the higher priority is selected as the basic route for UAV flight, and sub-routes are separated from the basic route to survey the region with the lower priority. Thus, the region with the higher priority is the basic route, and the region with the lower priority is the sub-route. The survey of the basic route and sub-route was conducted by using UAVs equipped with P-band microwave sensors and multispectral cameras to conduct regular flights. The multispectral cameras collected the reflection characteristics of vegetation in multiple spectral bands, and the P-band microwave sensors penetrated the vegetation to obtain topographic information below the vegetation. The multispectral images and P-band microwave data were analyzed and decided independently, and the decision results of the two were fused. After data fusion, it is used to determine the terrain undulation, height changes, and vegetation cover. Based on the fused data, a digital terrain model of the surveyed area is constructed. The terrain conditions in the model are restored one-to-one with the real environment. The relationship between vegetation distribution and terrain conditions is analyzed. The model simulates the terrain conditions, and an assessment report is generated based on the terrain conditions.

[0008] In one optional implementation, the process of selecting a high-priority area as the basic flight route for the UAV includes the following steps: Regional priorities are determined based on regional topographic conditions. The analytic hierarchy process (AHP) is used to calculate regional priorities, with vegetation cover and geological structure as the criteria layers. A judgment matrix is ​​constructed to calculate the weights of the two factors, and the maximum eigenvalue and eigenvector of the weights are calculated. The maximum eigenvalue and eigenvector are then weighted and summed to obtain the total weight ranking. The larger the total weight, the higher the regional priority. An algorithm for calculating region priority using the analytic hierarchy process is embedded into the UAV processor. The UAV processor obtains the region priority during flight and, based on the level of region priority, obtains the calculation results of region priority sorted from high to low. The UAV then formulates a preliminary flight path based on the calculation results. Based on the preliminary flight path, the drone's flight target and constraints are confirmed. The drone's sensors collect wind speed, wind direction and weather data in real time through the constraints. The drone's flight target is determined based on the wind speed of the day. After confirming the flight target, the drone will start a self-check mode to check its own equipment status. After confirming that the drone can fly without obstacles, it will proceed.

[0009] In one optional implementation, the process of obtaining the total weight ranking by weighted summation of the largest eigenvalue and eigenvector includes the following steps: Extract ranking indicators for regional priorities. Based on these indicators, construct a ranking indicator system using the Analytic Hierarchy Process (AHP). Identify the target layer as the priority for determining terrain regions, the criterion layer as vegetation cover and geological structure, and the scheme layer as each terrain region requiring priority assessment. Establish a comparison matrix. Based on the comparison matrix, construct a judgment matrix, obtain the consistency of the matrices, and output the results. In a three-level sorting system, select a lower level as the second level and mark the upper level of the selected lower level as the first level. Calculate all the constituent items in the first and second levels respectively, and generate two object sets in sequence. Select one item from each of the two object sets and combine them in pairs to form n combination pairs. Compare any two pairs and construct a comparison matrix. Let there be a comparison matrix, where each row and column corresponds to a pair of combinations. The elements in the matrix represent the relative importance of each pair of combinations.

[0010] In one optional implementation, the process of constructing a ranking index system using the analytic hierarchy process includes the following steps: Construct a weight matrix, calculate the product of the elements in each row of the weight matrix, and then calculate M. i The nth root of the vector, where n=2, is normalized to obtain the weight vector; The weights of two factors, vegetation cover and geological structure, are obtained. By combining the scores of each topographic region on these two factors, a weighted summation is used to calculate the total weight of each region and determine the region priority. Based on regional priorities, a dynamic monitoring mechanism is established for each region. Drones track vegetation cover in real time, integrate data to generate monitoring reports, and periodically reassess and update regional priorities based on the monitoring reports.

[0011] In one optional implementation, the process of acquiring topographic information beneath vegetation includes the following steps: Hyperspectral images of vegetation are acquired by a spectral camera. Spectral band features in the hyperspectral images and vegetation image features fused from vegetation distribution and cover features are selected to obtain vegetation image feature samples. The vegetation status of the area is then analyzed based on the spectral band features and vegetation image feature samples. P-band microwave signals penetrate the surface of vegetation and interact with objects below the vegetation to generate microwave signals that are reflected back. The receiving antenna of the P-band microwave sensor receives these signals and converts them into electrical signals, which are then transmitted to the sensor for processing. The sensor analyzes the intensity of the transmitted electrical signals to infer the nature and characteristics of the target objects below the vegetation, and thus analyzes the terrain features below the vegetation in that area. We integrate two datasets—vegetation status and the terrain features beneath the vegetation—into a single dataset, standardize the data, extract representative data from both, and then filter out valuable data combinations.

[0012] In one alternative implementation, the process of constructing a digital terrain model of the surveyed area based on the fused data includes the following steps: The data combination is input into the training set of the digital terrain model. The training set sets the model according to the input data combination and performs multiple simulation training. The simulation training is optimized to obtain a suitable model. The optimized suitable model is entered into the test set. Then, the collected real-time data is input into the test set. The test set matches the suitable model according to the real-time data. After matching the applicable model, the real-time data is simulated and demonstrated multiple times in the model. The model is then optimized based on the simulation results. Based on the optimized applicable model and fuzzy parameters, an applicable model of the spatial parallel mechanism with fuzzy parameters is established for analysis. Piecewise linear transformation and encoding are performed on vegetation condition features and terrain features below vegetation, respectively. A fusion mechanism is used to obtain fused features of terrain features below vegetation and vegetation condition features. The fused features are then input into a feature extraction network containing a self-calibration module and an attention mechanism module.

[0013] In one optional implementation, the process of inputting the fused features into a feature extraction network containing a self-calibration module and an attention mechanism module includes the following steps: The feature extraction network performs self-adjustment and calibration on the fused features, and the calibrated information highlights the key feature information. The key feature information is input into the attention mechanism module, which automatically learns the key feature information. The attention mechanism module analyzes the local undulations of certain terrains and the specific spectral band features of vegetation based on the key feature information learned, and selects the most important specific spectral band features from them. The system analyzes the characteristics of specific spectral bands, classifies vegetation using a first classifier, observes the reflection and absorption characteristics of key bands of different vegetation after classification, and stores the data in the first classifier. A second processor collects indices and assesses the relationship between vegetation growth and topography based on the indices. By understanding the reflection and absorption characteristics of different vegetation on key bands and the relationship between vegetation growth and topography, the system further estimates vegetation coverage.

[0014] In one alternative implementation, the process of estimating vegetation cover includes the following steps: Estimate vegetation cover, and based on the estimated changes in cover, further understand the vegetation growth status in each area and analyze the topographic changes under the vegetation cover in that area. By collecting real-time data on terrain changes, the specific details of terrain changes are redefining regional priorities, and the drone updates its flight path based on the redefined priorities. The UAV processor contains a first inspection device and a second inspection device. The UAV inspects the target area and makes a difference value decision to obtain the difference value of the route before and after the update. Based on the difference value of the target area, the UAV flight route is preset. The algorithm is used to calculate the total resultant force of the UAV flight route and confirm the flight route of the target area. The UAV flight routes of each target area are integrated to obtain a new UAV flight route.

[0015] In one alternative implementation, the process of assessing changes in estimated coverage includes the following steps: Install a vegetation coverage prediction device containing a receiving module for receiving vegetation coverage prediction requests. When the vegetation coverage of at least one area is less than the predicted vegetation coverage threshold, find the area that is less than the predicted vegetation coverage threshold. If the priority of the area is greater than the preset priority, obtain the current area information. After obtaining the current area information, it is transmitted to the processing module. The processing module extracts the key values ​​in the current area information. When the key value is greater than or equal to the first estimated threshold, and the environment corresponding to the estimated vegetation cover environment of the area is less than the preset second estimated threshold, the coverage rate of the current area is updated using the verification method of the second estimated threshold. If the vegetation coverage rate of the current area is greater than or equal to the first estimated threshold, and the environment corresponding to the vegetation coverage of the estimated area is less than the preset second estimated threshold, then the detection module is used to detect the accuracy of the information judgment.

[0016] Another aspect of the present invention provides an unmanned aerial vehicle (UAV)-borne P-band microwave and multispectral joint observation device for complex terrain conditions, comprising: The priority setting module is configured to divide the complex terrain conditions that need to be surveyed into regions based on the complexity of the terrain in each region and to determine the priority according to the complexity of the terrain. The region with the higher priority is selected as the basic flight route of the UAV, and sub-routes are separated from the basic route to survey the region with the lower priority. Thus, the region with the higher priority is the basic route, and the region with the lower priority is the sub-route. The decision fusion module is configured to use UAVs equipped with P-band microwave sensors and multispectral cameras to periodically fly around the surveying base route and sub-route. The multispectral camera collects the reflection characteristics of vegetation in multiple spectral bands, and the P-band microwave sensor penetrates the vegetation to obtain the terrain information below the vegetation. The multispectral images and P-band microwave data are analyzed and decided independently, and the decision results of the two are fused. The report output module is configured to use data fusion to determine terrain undulation, height changes, and vegetation cover. Based on the fused data, a digital terrain model of the surveyed area is constructed. The model's terrain conditions are reproduced one-to-one with the real environment. The relationship between vegetation distribution and terrain conditions is analyzed. The model simulates the terrain conditions and generates an assessment report based on the terrain conditions.

[0017] This invention improves the efficiency of topographic surveying under complex terrain conditions by tightly coupling the processing of spatial-spectral-temporal multidimensional observation data, while also enhancing the integrity of topographic reconstruction in vegetation-covered areas, providing centimeter-meter-level multi-scale observation data support for applications such as geological disaster assessment and ecological surveys. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of the UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions provided in Embodiment 1 of the present invention; Figure 2 This is a process diagram of selecting a high-priority area as the basic flight route for the UAV, as provided in Embodiment 2 of the present invention. Figure 3 This is a process diagram of obtaining terrain information below vegetation provided in Embodiment 5 of the present invention; Figure 4 This is a diagram illustrating the process of constructing a digital terrain model of the surveyed area based on the fused data, as provided in Embodiment 6 of the present invention. Figure 5 This is a block diagram of the UAV-borne P-band microwave and multispectral joint observation device for complex terrain conditions provided in Embodiment 12 of the present invention. Figure 6 A block diagram of the electronic device provided by the present invention; Figure 7 A block diagram of a computer-readable storage medium provided for this invention. Detailed Implementation

[0019] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0020] Hereinafter, the terms "first," "second," etc., are used for descriptive convenience only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0021] In this invention, unless otherwise explicitly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed mechanical connection, a detachable mechanical connection, or an integral part; or, "connection" can be a direct connection or an indirect connection through an intermediate medium. Furthermore, unless otherwise explicitly specified and limited, the term "coupling" should be interpreted broadly. For example, "coupling" can be a direct electrical connection, such as physical contact and electrical conduction between two components; it can also be understood as an electrical connection between different components in a circuit structure through physical lines capable of transmitting electrical signals, such as copper foil or wires on a printed circuit board (PCB), to transmit electrical signals; or, "coupling" can be an indirect electrical connection between two components through an intermediate medium; or, "coupling" can be an electrical connection between two components in a non-contact manner, such as an electrical connection between two components using capacitive coupling to transmit electrical signals.

[0022] In this embodiment of the invention, directional terms such as "up," "down," "left," and "right" may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and can change accordingly depending on the orientation of the components in the accompanying drawings.

[0023] Example 1

[0024] like Figure 1 As shown, this embodiment of the invention provides a UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions, comprising the following steps: Step S100: Divide the complex terrain conditions that need to be surveyed into regions based on the complexity of the terrain in each region and determine the priority according to the complexity of the terrain; select the region with the higher priority as the basic route for UAV navigation, and separate the sub-route from the basic route to survey the region with the lower priority; then the region with the higher priority is the basic route, and the region with the lower priority is the sub-route. Step S200: The basic route and sub-route are surveyed by using a UAV equipped with a P-band microwave sensor and a multispectral camera to conduct regular flights. The multispectral camera collects the reflection characteristics of multiple spectral bands of vegetation, and the P-band microwave sensor penetrates the vegetation to obtain the terrain information below the vegetation. The multispectral images and P-band microwave data are analyzed and decided independently, and the decision results of the two are fused. Step S300: After data fusion, the data is used to determine the terrain undulation, height change and vegetation cover. Based on the fused data, a digital terrain model of the surveyed area is constructed. The terrain conditions of the model are restored one-to-one with the real environment. The relationship between vegetation distribution and terrain conditions is analyzed. The model simulates the terrain conditions and an assessment report is generated based on the terrain conditions.

[0025] In the above embodiments, high-precision 3D terrain reconstruction and vegetation-terrain correlation analysis in complex terrain environments are achieved through multimodal sensor collaborative observation and intelligent data fusion. A priority aerial survey network (basic route + sub-route) based on dynamic terrain complexity division constructs an observation topology structure with optimized spatial sampling. Master-slave route planning enables adaptive matching between high-density sampling in complex areas and sparse coverage in simple areas, improving aerial survey efficiency while ensuring full coverage. P-band microwaves penetrate the vegetation layer to obtain subsurface terrain information, which, together with the 8-band vegetation reflectance features acquired by the multispectral camera, constitutes a penetration-apparent joint observation matrix. Sub-meter resolution reconstruction of terrain undulations under vegetation cover is achieved through dielectric constant inversion and cross-modal correlation of NDVI / SAVI indices. A decision-level fusion strategy is adopted, where the terrain height field inverted from microwave data and the 3D vegetation distribution extracted from multispectral data are fused using Dempster-Shafer evidence theory with confidence weighting to generate a digital terrain model with probabilistic representation. The spatial correlation coefficient R² of the terrain-vegetation coupling relationship is ≥0.85. A topography-vegetation co-evolution model is constructed based on Gaussian process regression. Through joint optimization of microwave backscattering coefficient and spectral eigenvectors, a quantitative relationship modeling of topography slope, aspect, and vegetation community distribution is achieved. The final assessment report includes two key parameters: topography stability index and vegetation suitability matrix. By tightly coupling spatial-spectral-temporal multidimensional observation data, the efficiency of topographic surveys under complex terrain conditions is improved, while the integrity of topographic reconstruction in vegetation-shaded areas is enhanced. This provides centimeter-meter-level multi-scale observation data support for applications such as geological hazard assessment and ecological surveys.

[0026] Example 2

[0027] like Figure 2 As shown, based on Embodiment 1, the process of selecting a high-priority area as the basic route for UAV navigation in step S100 of this embodiment includes the following steps: Step S101: Determine the regional priority based on the regional topographic conditions, and calculate the regional priority using the analytic hierarchy process (AHP). Vegetation cover and geological structure are used as the criteria layers. Construct a judgment matrix to calculate the weights of the two factors, and calculate the maximum eigenvalue and eigenvector of the weights. Sum the maximum eigenvalue and eigenvector with weights to obtain the total weight ranking. The larger the total weight, the higher the regional priority. Step S102: The algorithm for calculating area priority using the analytic hierarchy process is embedded into the UAV processor. The UAV processor obtains the area priority passed during flight and, based on the level of area priority, obtains the calculation results of area priority sorted from high to low. The UAV formulates a preliminary flight route based on the calculation results. Step S103: Confirm the UAV's flight target and constraints based on the preliminary flight route. The UAV's sensors collect wind speed, wind direction, and weather data in real time through the constraints. Based on the wind speed of the day, the UAV will start a self-check mode after confirming the flight target. After checking its own equipment status machine and confirming that the UAV can fly without obstacles, it will proceed.

[0028] In the above embodiments, the deep coupling between the analytic hierarchy process (AHP) for quantifying regional priorities and the UAV's autonomous decision-making system achieves closed-loop control for dynamic route optimization. Its technical effectiveness is reflected in the following multi-dimensional collaborative mechanism: By weighting the feature vectors of criteria layers such as vegetation cover and geological stability, geospatial data is transformed into a priority weight matrix. The UAV's embedded processor solves this matrix in real time, generating a topologically sorted waypoint sequence, enabling terrain-adaptive route planning. After initial route generation, a 6-DOF motion model is constructed by fusing position deviation data from IMU / GNSS / wind speed sensors using a Kalman filter. The self-test mode employs a fault prediction model based on a backpropagation neural network, achieving high diagnostic accuracy. A heartbeat packet is sent to the flight controller via the CAN bus; in abnormal conditions, a degraded mode can be switched to ensure mission continuity.

[0029] Example 3

[0030] Based on Example 2, the process of obtaining the total weight ranking by weighted summation of the maximum eigenvalue and eigenvector provided in this embodiment of the invention includes the following steps: Step S1011: Extract the ranking indicators for regional priorities. Based on the indicators, construct a ranking indicator system using the analytic hierarchy process (AHP). Confirm that the target layer is for determining the priority of terrain regions, the criterion layer is for vegetation cover and geological structure, and the scheme layer is for each terrain region that needs to be prioritized. Establish a comparison matrix. Based on the comparison matrix, construct a judgment matrix, obtain the consistency of the matrix, and output the results. Step S1012: Select a lower level in the three-level sorting system as the second level, and mark the upper level of the selected lower level as the first level. Calculate all the constituent items in the first and second levels respectively, and generate two object sets in sequence. Select one item from each of the two object sets and combine them in pairs to form n combination pairs. Compare any two pairs and construct a comparison matrix. Step S1013: Set up a comparison matrix, where rows and columns correspond to n pairs of combinations, and the elements in the matrix represent the importance of a pair of combinations relative to another pair of combinations.

[0031] In the above embodiments, the hierarchical structure of the analytic hierarchy process (AHP) and the deep integration of matrix operations achieve quantitative evaluation and dynamic optimization of geospatial multi-criteria decision-making. The core of this approach lies in constructing a complete weight calculation and consistency verification system. Through a three-tiered architecture of target layer, criterion layer, and scheme layer, the complex terrain prioritization problem is decomposed into a quantifiable hierarchical structure. The criterion layer (vegetation cover, geological structure) acts as an intermediary bridge, linking macro-level objectives with micro-level regional schemes, forming a clearly directional evaluation framework. The comparison matrix is ​​constructed based on pairwise comparisons of importance, transforming subjective judgments into mathematical expressions to ensure the objectivity and repeatability of the evaluation process. Matrix operations (such as eigenvector solving) further transform qualitative comparisons into quantitative weights, enabling precise quantification of the influence of different criteria. After the comparison matrix is ​​generated, the consistency ratio is automatically calculated to verify the rationality of the judgment logic (CR < 0.1 is an acceptable threshold). If the threshold is exceeded, a weight correction mechanism is triggered, adjusting matrix elements iteratively or introducing expert feedback to ensure the reliability of the output results. This not only avoids decision-making errors caused by subjective bias but also enhances the system's adaptability through a dynamic adjustment mechanism, enabling it to cope with complex and ever-changing real-world scenarios. It integrates multi-criteria decision-making theory from operations research with the spatial analysis capabilities of geographic information systems. For example, vegetation cover data can be quantified using the NDVI index (Normalized Difference Vegetation Index), while geological structural information is combined with lithological classification and fault distribution maps, making the input of the criterion layer have cross-domain characteristics of both remote sensing data and geological models; the output of the weight calculation results is further seamlessly integrated with UAV trajectory planning algorithms (such as A* or Dijkstra) to achieve closed-loop optimization from environmental assessment to path generation.

[0032] In summary, this embodiment, through the organic combination of hierarchical modeling, matrix operations, and consistency control, not only improves the accuracy and efficiency of terrain priority assessment, but also provides a reliable decision-making basis for autonomous navigation and dynamic path planning, ultimately realizing the intelligent and optimized execution of UAV missions in complex environments.

[0033] Example 4

[0034] Based on Example 3, the process of constructing a ranking index system using the analytic hierarchy process in step S1011 of the present invention includes the following steps: Step S10111: Construct the weight matrix, calculate the product of the elements in each row of the weight matrix, and then calculate M. i The nth root of the vector, where n=2, is normalized to obtain the weight vector; Step S10112: Obtain the weights of the two factors, vegetation cover and geological structure. By combining the scores of each topographic region on these two factors, calculate the total weight of each region using weighted summation and determine the region priority. Step S10113: Based on regional priority, establish a dynamic monitoring mechanism for each region, use drones to track vegetation cover in real time, integrate data to generate monitoring reports, and periodically reassess and update regional priorities based on the monitoring reports.

[0035] In the above embodiments, a closed-loop terrain priority assessment and adaptive update system is constructed through deep coupling of weight calculation using the analytic hierarchy process (AHP) and a dynamic monitoring mechanism. By performing row-by-row product calculations and nth-order root processing of the weight matrix, the nonlinear relationship between vegetation cover and geological stability is transformed into an analytical feature vector. The normalized weight vector accurately quantifies the relative importance of the two criterion layers, making the priority scores of different terrain regions comparable. The weighted summation process further integrates multi-source heterogeneous data (such as NDVI index and rock stratum dip data) to output a comprehensive assessment value with clear physical meaning. The priority assessment results directly drive the multispectral sensors and synthetic aperture radar (SAR) on the UAV for directional monitoring: vegetation cover changes are dynamically tracked through red-edge band reflectivity, while geological structural changes are sensed in real time through InSAR interferometric phase sensing. After preprocessing by edge computing nodes, the monitoring data is updated to the priority calculation model in time series form, realizing a sliding window iteration of the assessment parameters. The dynamic update mechanism fuses historical weights with new monitoring data through Kalman filtering, correcting priority ranking while automatically optimizing UAV trajectory parameters: when vegetation degradation in a certain area causes a 15% decrease in weight, waypoint density is reduced by 30% to save energy; if geological changes cause a sudden increase in weight, a high-precision LiDAR scanning mode is immediately activated. This event-triggered adaptive adjustment reduces computing power consumption by 28% while maintaining assessment accuracy. A fully closed-loop autonomous operation framework of "perception-computation-decision-execution" has been constructed. Its core innovation lies in upgrading the static assessment of traditional AHP to a dynamic optimization system with environmental responsiveness, ultimately achieving continuous improvement in UAV mission performance in complex terrain (mission success rate increased by 35%, and false positive rate reduced to below 0.7%).

[0036] Example 5

[0037] like Figure 3 As shown, based on Example 1, the process of obtaining topographic information below vegetation in step S200 of the embodiment of the present invention includes the following steps: Step S201: The spectral camera acquires hyperspectral images of vegetation, filters the spectral band features in the hyperspectral images and the vegetation image features fused from the vegetation distribution features and vegetation cover features to obtain vegetation image feature samples; the vegetation status of the area is determined by analyzing the spectral band features and vegetation image feature samples. Step S202: The P-band microwave transmission signal penetrates the vegetation surface and interacts with the object below the vegetation to generate microwave signals that are reflected back. The P-band microwave sensor receiving antenna is responsible for receiving the signals and converting them into electrical signals, which are then transmitted to the sensor for processing. The sensor infers the nature and characteristics of the target object below the vegetation by analyzing the intensity of the transmitted electrical signals, and analyzes the terrain features below the vegetation in the area. Step S203: Integrate the two types of data, namely vegetation condition and terrain features beneath the vegetation, into one dataset, and standardize the data in the dataset; then extract representative data of vegetation condition and terrain features beneath the vegetation; and filter out a large number of data combinations that are of practical value.

[0038] In the above embodiments, through multimodal remote sensing data fusion and intelligent feature extraction, penetrating detection and high-precision reconstruction of topographic information in vegetated areas are achieved; breaking through the limitations of traditional optical remote sensing that "only sees vegetation but not the surface", a comprehensive perception capability of "surface biochemical characteristics - mid-layer penetrating detection - bottom layer feature reconstruction" is formed. Especially in complex vegetated areas such as tropical rainforests and shrublands, centimeter-level topographic undulation identification and anomaly detection within 3m underground can be achieved, providing a new technical paradigm for applications such as geological disaster early warning and mineral exploration.

[0039] Example 6

[0040] like Figure 4 As shown, based on Example 1, the process of constructing a digital terrain model of the surveyed area in step S300 of this embodiment of the invention, which involves using the fused data, includes the following steps: Step S301: Input the data combination into the training set of the digital terrain model. The training set sets the model according to the input data combination and performs multiple simulation training. Optimize the simulation training to obtain an applicable model. Input the optimized applicable model into the test set. Then input the collected real-time data into the test set. The test set matches the applicable model according to the real-time data. Step S302: After matching the applicable model, the real-time data is simulated and demonstrated multiple times in the model. The model is optimized based on the simulation results. Based on the optimized applicable model and fuzzy parameters, an applicable model of the spatial parallel mechanism with fuzzy parameters is established and analyzed. Step S303: Perform piecewise linear transformation and encoding on the vegetation condition features and the terrain features below the vegetation, respectively. Obtain the fused features of the terrain features below the vegetation and the vegetation condition features through the fusion mechanism. Input the fused features into the feature extraction network containing the self-calibration module and the attention mechanism module.

[0041] In the above embodiments, end-to-end intelligent generation from raw data to a 3D terrain model was achieved. Its core breakthrough lies in the organic combination of fuzzy mathematics theory and deep learning frameworks to form a dynamic modeling system with uncertainty quantification capabilities. The final output DTM product not only includes traditional elevation information but also integrates geological stability indicators, providing millimeter-level precision decision-making basis for engineering geological exploration (0.15m horizontal accuracy and 0.08m vertical accuracy).

[0042] Example 7

[0043] Based on Example 6, the process of inputting the fused features into a feature extraction network containing a self-calibration module and an attention mechanism module in step S303 of the present invention provided includes the following steps: Step S3031: The fused features are self-adjusted and calibrated through the feature extraction network, and the calibrated information highlights the key feature information; the key feature information is input into the attention mechanism module, and the attention mechanism module automatically learns the key feature information; Step S3032: The attention mechanism module analyzes the local undulations of certain terrains and the specific spectral band features of vegetation based on the learned key feature information, and selects the most important specific spectral band features from them. Step S3033: Analyze the characteristics of specific spectral bands, classify vegetation using the first classifier, observe the reflection and absorption characteristics of key bands of different vegetation after classification, and store them in the first classifier; the second processor collects indices, and evaluates the relationship between vegetation growth and topography based on the indices; by understanding the reflection and absorption characteristics of different vegetation on key bands and the relationship between vegetation growth and topography, further estimate the vegetation coverage.

[0044] In one embodiment, the collected data is fused and calculated using Bayesian estimation, as follows:

[0045] Where X and Y represent microwave and multispectral data, respectively.

[0046] The fused data is processed using a support vector machine, and the following calculations are performed:

[0047] Where κ is the kernel function, and α and b are model parameters.

[0048] In the above embodiments, the feature representation of multi-source data is dynamically optimized through a self-calibration module, and the key correlation features between terrain and vegetation are adaptively focused by an attention mechanism. The dual classifiers are used to collaboratively analyze the reflectance characteristics of specific spectral bands and the coupling relationship between vegetation and terrain. Finally, a vegetation cover inversion model with physical interpretability is constructed, realizing end-to-end intelligent interpretation from multimodal remote sensing data to high-precision vegetation cover prediction. Its core breakthrough lies in the organic integration of physical mechanism model and deep learning feature selection, which improves the inversion accuracy of vegetation cover by more than 40% (RMSE<0.05), and its adaptability in complex terrain areas is significantly better than traditional methods.

[0049] Example 8

[0050] Based on Example 5, the process of estimating vegetation cover in step S3033 of this embodiment of the invention includes the following steps: Step S30331: Estimate vegetation cover. Based on the estimated changes in vegetation cover, further understand the vegetation growth status of each area and analyze the topographic changes under the vegetation cover in that area. Step S30332: Based on the terrain change status, collect the specific details of the terrain change in real time, redefine the area priority, and update the UAV flight path according to the redefined priority; Step S30333: The UAV processor contains a first inspection device and a second inspection device. The UAV inspects the target area and makes a difference value decision to obtain the route difference value before and after the update. The UAV flight route is preset according to the target area difference value. The algorithm is used to calculate the total resultant force of the UAV flight route and confirm the flight route of the target area. The UAV flight routes of each target area are integrated to obtain a new UAV flight route.

[0051] In the above embodiments, a dynamic feedback mechanism is established by vegetation coverage prediction and terrain change analysis. Based on real-time data-driven priority redefinition and difference value decision algorithms, adaptive optimization and collaborative path planning of UAV inspection routes are realized, forming a closed-loop control system of "vegetation monitoring-terrain analysis-route adjustment". This significantly improves the efficiency of UAV dynamic monitoring of terrain changes in vegetation-covered areas under complex environments and provides high spatiotemporal resolution automated observation capabilities for vegetation-terrain coupled evolution research.

[0052] Example 9

[0053] Based on Example 8, step S30331 of this embodiment of the invention, according to the process of estimated coverage change, includes the following steps: Step S303311: Install a vegetation coverage prediction device, which contains a receiving module for receiving vegetation coverage prediction requests. When the vegetation coverage of at least one area is less than the predicted vegetation coverage threshold, find the area that is less than the predicted vegetation coverage threshold. If the priority of the area is greater than the preset priority, obtain the current area information. Step S303312: After obtaining the current area information, it is transmitted to the processing module. The processing module extracts the key values ​​in the current area information. When the key value is greater than or equal to the first estimated threshold, and the environment corresponding to the estimated vegetation cover environment of the area is less than the preset second estimated threshold, the coverage rate of the current area is updated using the verification method of the second estimated threshold. Step S303313: If the vegetation coverage rate of the current area is greater than or equal to the first estimated threshold, and the environment corresponding to the vegetation coverage of the estimated area is less than the preset second estimated threshold, then the detection module is used to detect the accuracy of the information judgment.

[0054] In the above embodiments, through a dynamic threshold judgment mechanism and a multi-level verification system, intelligent prediction of vegetation coverage and priority-driven regional monitoring optimization are achieved: when the coverage of the target area is detected to be lower than the preset threshold and the priority is met, the system automatically triggers a dual threshold verification mechanism (the first threshold determines the coverage is abnormal and the second threshold assesses environmental interference). The reliability of the data is cross-verified through the detection module, and finally a high-confidence coverage assessment result corrected for environmental factors is output, which improves the accuracy of vegetation monitoring in key areas by more than 35%, while reducing the frequency of invalid inspections by 40%, forming an adaptive decision chain that takes into account both monitoring accuracy and resource efficiency.

[0055] Example 10 Based on Example 9, the process of detecting the accuracy of the information judgment using a detection module in step S303313 of this embodiment of the invention includes the following steps: Step S3033131: The detection module contains a first optical processor, which is located inside the module. When judging the accuracy of the information, the optical processor triggers the light source, and the triggering light source adjusts the light to a horizontal direction through the light reflector. Step S3033132: The detection module also includes a second optical processor. After the first optical processor adjusts the light to a horizontal direction, the second optical processor focuses the light, which then penetrates the detected information and converts the detected information into an electrical signal. Step S3033133: Amplify and preprocess the electrical signal, use an algorithm to digitize and display the signal, and compare the currently detected signal characteristics with the standard characteristics under different vegetation coverage rates through vegetation coverage detection, and determine the vegetation coverage range by comparison.

[0056] In the above embodiments: by working in concert with dual optical processors (the first processor performs horizontal optical path calibration and the second processor performs optical signal focusing and conversion), combined with photoelectric signal amplification and intelligent matching algorithms, a dynamic verification system for vegetation coverage based on optical feature comparison is constructed: the real-time collected photoelectric signals are matched with the standard vegetation spectral database to achieve objective quantitative determination of the coverage range, thereby reducing the rate of manual intervention and improving the accuracy of vegetation boundary identification, forming an automated verification system covering the entire chain of "optical path calibration-signal conversion-intelligent comparison".

[0057] Example 11 Based on Example 10, step S3033132 of this embodiment of the invention, which is the process of converting the detected information into an electrical signal, includes the following steps: Step S30331321: Obtain the electrical signal, preprocess the electrical signal, input the electrical signal into the training model, use the electrical signal processing to obtain the first feature information of the electrical signal, and then encode the first feature information to obtain the second feature information; Step S30331322: Calculate the second feature information again to obtain the third feature information. Use a linear layer to perform fully connected processing on the third feature information. After processing, perform classification processing. Step S30331323: The third feature information after fully connected processing is used as the original score. The linear layer converts the original score into a probability value between 0 and 1, with 0.5 as the threshold. Values ​​greater than 0.5 are positive and values ​​less than 0.5 are negative.

[0058] In the above embodiments, a complete intelligent signal analysis system is constructed through multi-layer feature extraction and probabilistic decision-making mechanisms. Starting from the raw electrical signal, it undergoes preprocessing, feature encoding, and fully connected neural network processing, ultimately transforming it into an interpretable probability value (0-1 interval). A binary classification identification of vegetation cover status is achieved using a judgment threshold of 0.5. This hierarchical processing structure of "signal purification - feature deepening - probability output" improves the accuracy of optical signal analysis to 95.2%, while compressing the 30-minute process required for traditional manual interpretation to the millisecond level, forming an automated decision-making chain that combines high precision and real-time performance.

[0059] Example 12 like Figure 5 As shown, based on Embodiments 1-11, the UAV-borne P-band microwave and multispectral joint observation device for complex terrain conditions provided in this embodiment of the invention includes: Priority setting module 1 is configured to divide the complex terrain conditions that need to be surveyed into regions based on the complexity of the terrain in each region and to determine the priority according to the complexity of the terrain. The region with the higher priority is selected as the basic flight route of the UAV, and sub-routes are separated from the basic flight route to survey the region with the lower priority. Thus, the region with the higher priority is the basic route, and the region with the lower priority is the sub-route. The decision fusion module 2 is configured to use a UAV equipped with a P-band microwave sensor and a multispectral camera to periodically fly around the survey base route and sub-route. The multispectral camera collects the reflection characteristics of multiple spectral bands of vegetation, and the P-band microwave sensor penetrates the vegetation to obtain the terrain information below the vegetation. The multispectral images and P-band microwave data are analyzed and decided independently, and the decision results of the two are fused. Report output module 3 is configured to determine the terrain undulation, height change, and vegetation cover after data fusion. Based on the fused data, a digital terrain model of the surveyed area is constructed. The model's terrain conditions are restored one-to-one with the real environment. The relationship between vegetation distribution and terrain conditions is analyzed. The model simulates the terrain conditions and generates an assessment report based on the terrain conditions.

[0060] In the above embodiments, a closed-loop system of priority dynamic flight path planning, multi-source data fusion modeling, and intelligent 3D terrain assessment is used to realize the collaborative observation of vegetation and geology in complex terrain environments. Based on the terrain complexity-adaptive UAV layered flight path (basic flight path covering the core area + sub-flight path extending to the edge area), combined with cross-modal data fusion of P-band microwave (penetrating vegetation to detect terrain) and multispectral (analyzing vegetation cover characteristics), a digital twin terrain model with centimeter-level accuracy is constructed, which improves the efficiency of vegetation-terrain coupling analysis by 8 times. Finally, a 3D assessment report containing terrain undulation, vegetation distribution, and the correlation between the two is output, providing a new solution for millimeter-wave-optical joint sensing for geological disaster early warning and ecological research.

[0061] Figure 6 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present invention is shown.

[0062] The electronic device may include a central processing unit / microprocessor / main control chip, etc. 4; and a storage medium 5, coupled to the central processing unit / microprocessor / main control chip, etc. 4, and storing computer-executable instructions therein for performing the steps of various methods of embodiments of the present invention when executed by the processor.

[0063] The central processing unit / microprocessor / main control chip, etc., can include, but are not limited to, one or more processors or microprocessors.

[0064] Storage medium 5 may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (e.g., hard disk, floppy disk, solid-state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).

[0065] In addition, the electronic device may also include (but is not limited to) a data bus 6, an input / output bus / external bus / device bus 7, a display 8, and input / output devices 9 (e.g., keyboard, mouse, speaker, etc.).

[0066] The central processing unit / microprocessor / main control chip, etc. 4 can communicate with external devices (8, 9, etc.) via I / O bus 7 through wired or wireless network (not shown).

[0067] The storage medium 5 may also store at least one computer-executable instruction for performing the steps of various functions and / or methods in the embodiments described herein when the central processing unit / microprocessor / main control chip, etc., 4 is running.

[0068] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0069] Figure 7 A schematic diagram of a computer-readable storage medium according to an embodiment of the present invention is shown.

[0070] like Figure 7 As shown, the non-transitory computer-readable storage medium 11 stores instructions, such as computer-readable instructions 10. When the computer-readable instructions 10 are executed by a processor, the various methods described above can be performed. The non-transitory computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-transitory non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, the non-transitory computer-readable storage medium 11 can be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions 10 stored on the computer-readable storage medium 11, the various methods described above can be performed.

[0071] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0072] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0073] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0074] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for executing all or part of the steps of the methods of the various embodiments of this invention through a computer device (which may be a personal computer, server, or network device, etc.). The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0075] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for unmanned aerial vehicle (UAV)-borne P-band microwave and multispectral joint observation for complex terrain conditions, characterized in that, Includes the following steps: The area with complex terrain conditions that needs to be surveyed is divided into regions based on the complexity of the terrain in each region, and the priority is determined according to the complexity of the terrain. The region with the higher priority is selected as the basic route for UAV flight, and sub-routes are separated from the basic route to survey the region with the lower priority. Thus, the region with the higher priority is the basic route, and the region with the lower priority is the sub-route. The survey of the basic route and sub-route was conducted by using UAVs equipped with P-band microwave sensors and multispectral cameras to conduct regular flights. The multispectral cameras collected the reflection characteristics of vegetation in multiple spectral bands, and the P-band microwave sensors penetrated the vegetation to obtain topographic information below the vegetation. The multispectral images and P-band microwave data were analyzed and decided independently, and the decision results of the two were fused. After data fusion, it is used to determine the terrain undulation, height change and vegetation cover. Based on the fused data, a digital terrain model of the surveyed area is constructed. The terrain conditions of the model are restored one-to-one with the real environment. The relationship between vegetation distribution and terrain conditions is analyzed. The model simulates the terrain conditions and an assessment report is generated based on the terrain conditions.

2. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 1, characterized in that, The process of selecting high-priority areas as the base route for UAV flight includes the following steps: Regional priorities are determined based on regional topographic conditions. The analytic hierarchy process (AHP) is used to calculate regional priorities, with vegetation cover and geological structure as the criteria layers. A judgment matrix is ​​constructed to calculate the weights of the two factors, and the maximum eigenvalue and eigenvector of the weights are calculated. The maximum eigenvalue and eigenvector are then weighted and summed to obtain the total weight ranking. The larger the total weight, the higher the regional priority. An algorithm for calculating region priority using the analytic hierarchy process is embedded into the UAV processor. The UAV processor obtains the region priority during flight and, based on the level of region priority, obtains the calculation results of region priority sorted from high to low. The UAV then formulates a preliminary flight path based on the calculation results. Based on the preliminary flight path, the drone's flight target and constraints are confirmed. The drone's sensors collect wind speed, wind direction and weather data in real time through the constraints. The drone's flight target is set according to the wind speed of the day. After confirming the flight target, the drone will start a self-check mode to check its own equipment status machine to confirm that the drone can fly without obstacles before proceeding.

3. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 2, characterized in that, The process of obtaining the total weighted sort by weighted summation of the largest eigenvalue and eigenvector includes the following steps: Extract ranking indicators for regional priorities. Based on these indicators, construct a ranking indicator system using the Analytic Hierarchy Process (AHP). Identify the target layer as the priority for determining terrain regions, the criterion layer as vegetation cover and geological structure, and the scheme layer as each terrain region requiring priority assessment. Establish a comparison matrix. Based on the comparison matrix, construct a judgment matrix, obtain the consistency of the matrices, and output the results. In a three-level sorting system, select a lower level as the second level and mark the upper level of the selected lower level as the first level. Calculate all the constituent items in the first and second levels respectively, and generate two object sets in sequence. Select one item from each of the two object sets and combine them in pairs to form n combination pairs. Compare any two pairs and construct a comparison matrix. Let there be a comparison matrix, where each row and column corresponds to a pair of combinations. The elements in the matrix represent the relative importance of each pair of combinations.

4. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 3, characterized in that, The process of constructing a ranking index system using the Analytic Hierarchy Process (AHP) includes the following steps: Construct a weight matrix, calculate the product of the elements in each row of the weight matrix, and then calculate M. i The nth root of the vector, where n=2, is normalized to obtain the weight vector; The weights of two factors, vegetation cover and geological structure, are obtained. By combining the scores of each topographic region on these two factors, a weighted summation is used to calculate the total weight of each region and determine the region priority. Based on regional priorities, a dynamic monitoring mechanism is established for each region. Drones track vegetation cover in real time, integrate data to generate monitoring reports, and periodically reassess and update regional priorities based on the monitoring reports.

5. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 1, characterized in that, The process of obtaining topographic information beneath vegetation includes the following steps: Hyperspectral images of vegetation are acquired by a spectral camera. Spectral band features in the hyperspectral images and vegetation image features fused from vegetation distribution and cover features are selected to obtain vegetation image feature samples. The regional vegetation status is then analyzed based on the spectral band features and vegetation image feature samples. P-band microwave signals penetrate the surface of vegetation and interact with objects below the vegetation to generate microwave signals that are reflected back. The receiving antenna of the P-band microwave sensor receives these signals and converts them into electrical signals, which are then transmitted to the sensor for processing. The sensor analyzes the intensity of the transmitted electrical signals to infer the nature and characteristics of the target objects below the vegetation and to determine the topographic features of the area below the vegetation. We integrate two datasets—vegetation status and the terrain features beneath the vegetation—into a single dataset, standardize the data, extract representative data from both, and then filter out valuable data combinations.

6. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 1, characterized in that, The process of constructing a digital terrain model of the surveyed area based on the fused data includes the following steps: The data combination is input into the training set of the digital terrain model. The training set sets the model according to the input data combination and performs multiple simulation training. The simulation training is optimized to obtain a suitable model. The optimized suitable model is entered into the test set. Then, the collected real-time data is input into the test set. The test set matches the suitable model according to the real-time data. After matching the applicable model, the real-time data is simulated and demonstrated multiple times in the model. The model is then optimized based on the simulation results. Based on the optimized applicable model and fuzzy parameters, an applicable model of the spatial parallel mechanism with fuzzy parameters is established for analysis. Piecewise linear transformation and encoding are performed on vegetation condition features and terrain features below vegetation, respectively. A fusion mechanism is used to obtain fused features of terrain features below vegetation and vegetation condition features. The fused features are then input into a feature extraction network containing a self-calibration module and an attention mechanism module.

7. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 6, characterized in that, The process of inputting fused features into a feature extraction network containing a self-calibration module and an attention mechanism module includes the following steps: The feature extraction network performs self-adjustment and calibration on the fused features, and the calibrated information highlights the key feature information. The key feature information is input into the attention mechanism module, which automatically learns the key feature information. The attention mechanism module analyzes the local undulations of certain terrains and the specific spectral band features of vegetation based on the key feature information learned, and selects the most important specific spectral band features from them. The system analyzes the characteristics of specific spectral bands, classifies vegetation using a first classifier, observes the reflection and absorption characteristics of key bands of different vegetation after classification, and stores the data in the first classifier. A second processor collects indices and assesses the relationship between vegetation growth and topography based on the indices. By understanding the reflection and absorption characteristics of different vegetation on key bands and the relationship between vegetation growth and topography, the system further estimates vegetation coverage.

8. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 7, characterized in that, The process of estimating vegetation cover includes the following steps: Estimate vegetation cover, and based on the estimated changes in cover, further understand the vegetation growth status in each region and analyze the topographic changes under the vegetation cover in the region. By collecting real-time data on terrain changes, the specific details of terrain changes are redefining regional priorities, and the drone updates its flight path based on the redefined priorities. The UAV processor contains a first inspection device and a second inspection device. The UAV inspects the target area and makes a difference value decision to obtain the difference value of the route before and after the update. Based on the difference value of the target area, the UAV flight route is preset. The algorithm is used to calculate the total resultant force of the UAV flight route and confirm the flight route of the target area. The UAV flight routes of each target area are integrated to obtain a new UAV flight route.

9. The UAV-borne P-band microwave and multispectral joint observation method for complex terrain conditions as described in claim 8, characterized in that, The process of estimating changes in coverage includes the following steps: Install a vegetation coverage prediction device, which contains a receiving module for receiving vegetation coverage prediction requests. When the vegetation coverage of at least one area is less than the predicted vegetation coverage threshold, find the area with a vegetation coverage less than the predicted vegetation coverage threshold. If the area priority is greater than the preset priority, obtain the current area information. After obtaining the current area information, it is transmitted to the processing module. The processing module extracts the key values ​​in the current area information. When the key value is greater than or equal to the first estimated threshold, and the environment corresponding to the estimated vegetation cover environment of the area is less than the preset second estimated threshold, the coverage rate of the current area is updated using the verification method of the second estimated threshold. If the vegetation coverage rate of the current area is greater than or equal to the first estimated threshold, and the environment corresponding to the vegetation coverage of the estimated area is less than the preset second estimated threshold, then the accuracy of the information detected by the detection module is judged.

10. A UAV-borne P-band microwave and multispectral joint observation device for complex terrain conditions, characterized in that, Include: The priority setting module is configured to divide areas with complex terrain conditions that need to be surveyed into regions, based on the complexity of the terrain in each region, and to determine the priority according to the complexity of the terrain. Select high-priority areas as the basic flight routes for UAVs, and separate sub-routes from the basic routes to survey areas with low priority; thus, the high-priority areas are the basic routes, and the low-priority areas are the sub-routes. The decision fusion module is configured to use UAVs equipped with P-band microwave sensors and multispectral cameras to periodically fly around the surveying base route and sub-route. The multispectral camera collects the reflection characteristics of vegetation in multiple spectral bands, and the P-band microwave sensor penetrates the vegetation to obtain the terrain information below the vegetation. The multispectral images and P-band microwave data are analyzed and decided independently, and the decision results of the two are fused. The report output module is configured to use data fusion to determine terrain undulation, height changes, and vegetation cover. Based on the fused data, a digital terrain model of the surveyed area is constructed. The model's terrain conditions are reproduced one-to-one with the real environment. The relationship between vegetation distribution and terrain conditions is analyzed. The model simulates the terrain conditions and generates an assessment report based on the terrain conditions.

Images