Three-dimensional terrain reconstruction method and system based on unmanned aerial vehicle image

Abstract

The invention relates to the field of photogrammetry, in particular to a three-dimensional terrain reconstruction method and system based on an unmanned aerial vehicle image. A three-dimensional terrain reconstruction system based on an unmanned aerial vehicle image comprises an information acquisition module, a terrain difference judgment module, a flight mission planning module and a terrain reconstruction module. According to the method, key parameters such as the flight height, the overlapping rate, the speed and the exposure interval can be dynamically updated in the flight process through a period adjustment mechanism driven by a difference index and a neural network task parameter solving model; when the wind speed suddenly increases, the illumination changes or the landform drastically fluctuates in the measured area, the system automatically shortens the refresh cycle and executes global or local adjustment, so that potential holes or fuzzy images are compensated in time, high-coverage and high-definition acquisition of the complex landform is realized, and manual intervention or expensive hardware addition is not needed.

Description

Technical Field

[0001] The present invention relates to the field of photogrammetry, and in particular to a three-dimensional terrain reconstruction method and system based on unmanned aerial vehicle (UAV) images. Background Art

[0002] UAV aerial photogrammetry, relying on lightweight flight platforms, digital cameras, and post-processing software, has become the primary method for acquiring high-resolution terrain data and orthophotos. Existing workflows typically complete route planning once before takeoff: a fixed flight altitude, track spacing, heading-to-side overlap, speed, and exposure interval are given based on factors such as desired resolution, average regional relief, and conventional overlap. The drone then performs photography according to these parameters, and upon completion of the mission, image preprocessing, feature matching, structure-from-motion (SfM), multi-view stereo (MVS), and orthophotoization are completed at a ground workstation. This "static planning-offline reconstruction" model can meet accuracy requirements in small, flat, and stable environments, but exhibits significant shortcomings in complex terrain and dynamic environments.

[0003] First, environmental factors (such as wind speed, visibility, and illumination) often experience short-term and dramatic fluctuations during flight. Light drones have limited wind resistance. If the wind speed suddenly increases, the body's attitude jitter will cause stripe distortion. If visibility suddenly drops or the light changes rapidly, it is very easy to cause underexposure, overexposure, or motion blur. Static routes cannot respond to these changes in real time, and imaging quality is difficult to guarantee. Secondly, large-scale survey areas often include plains, hills or canyons with dramatic elevation differences, and forest areas with severe line of sight obstruction. A unified flight altitude and overlap setting makes it difficult to take into account all sub-areas: when the altitude is too low, image holes will appear in the undulating areas; when the altitude is too high, redundant shooting will occur in the flat areas, which not only wastes electricity but also increases the amount of post-processing data. In addition, traditional parameter settings are highly dependent on the experience of the operators, and it is difficult to maintain consistency between different teams or different batches of tasks. In the face of sudden weather or complex terrain, manual real-time intervention is both time-consuming and error-prone.

[0004] To address the above pain points, the present invention proposes a three-dimensional terrain reconstruction method and system based on drone images. Summary of the Invention

[0005] The present invention uses a periodic adjustment mechanism driven by difference indicators and a neural network task parameter solution model to dynamically update key parameters such as flight altitude, overlap rate, speed and exposure interval during flight. When a sudden increase in wind speed, changes in lighting or drastic terrain fluctuations occur in the survey area, the system automatically shortens the refresh cycle and performs global or local adjustments to promptly compensate for potential holes or blurred images, achieving high coverage and high-definition acquisition of complex terrain without the need for manual intervention or the addition of expensive hardware.

[0006] A three-dimensional terrain reconstruction method based on UAV images, comprising: Obtain environmental factors that can affect flight mission planning and target area terrain features, and organize them into environmental factor vectors and terrain feature vectors respectively; Calculating a terrain difference index of the target area based on the terrain feature vector, comparing the index with a preset difference threshold, and classifying the target area into a high consistency scene or a low consistency scene according to the comparison result; If the target area is a high-consistency scene, the trained mission parameter solution model is used, taking the environmental factor vector and terrain feature vector as input, and outputting a flight mission planning parameter set, including flight altitude, track spacing, heading overlap rate, side overlap rate, flight speed, and exposure interval. The UAV then uses the flight mission planning parameter set to execute the flight mission. During the UAV flight, the environmental factor vector is regularly updated at a preset basic cycle, and local adjustments to the flight mission planning parameter set are made based on changes in the environmental factor vector, and the flight mission planning parameter set is issued for execution. If the target area is a low consistency scenario, the duration of the basic cycle is first adjusted according to the terrain difference index of the target area to obtain the adaptive update cycle; then the terrain feature vector and environmental factor vector of the drone's takeoff sub-area are input into the mission parameter solution model to obtain the initial flight mission planning parameter set, and then the drone uses the initial flight mission planning parameter set to perform the flight mission; during the drone's flight, the environmental factor vector is regularly updated according to the adaptive update cycle, and the terrain feature vector of the sub-area where the drone is currently located is obtained. Based on the changes in the environmental factor vector and the current terrain feature vector, the initial flight mission planning parameter set is globally adjusted and issued for execution; After image acquisition is completed, all images are subjected to image preprocessing, feature matching, sparse 3D reconstruction, dense matching, surface reconstruction and orthophotoization in sequence, and finally the digital elevation model and orthophoto results are output.

[0007] Preferably, a terrain difference index of the target area is calculated based on the terrain feature vector, the index is compared with a preset difference threshold, and the target area is divided into a high consistency scene or a low consistency scene according to the comparison result. The specific steps include: The grid side length is selected based on the expected ground resolution, ensuring that the grid diagonal distance is no greater than half the planned heading spacing. The target area is divided into several sub-areas. For each sub-area, pilot images acquired before the flight mission are used to extract the following four types of terrain indicators that can be obtained from the pilot images: texture complexity, edge density, brightness variance, and parallax amplitude. The four types of terrain indicators mentioned above are normalized by their ranges. Then, they are linearly superimposed according to the preset weight coefficients to obtain a single scalar as the terrain complexity score of the sub-region. The terrain complexity scores of all sub-regions are combined into a sample set. The set mean is first calculated, and then the absolute value of each score's deviation from the mean is calculated and averaged to obtain the average absolute deviation. At the same time, the interquartile range of the sample set is calculated, and the larger value between the average absolute deviation and the interquartile range is selected as the final terrain difference index. Set a difference threshold. If the terrain difference index is not greater than the difference threshold, the target area is judged to be a high consistency scene; if the terrain difference index is greater than the difference threshold, it is judged to be a low consistency scene.

[0008] Preferably, the local adjustment operation for high consistency scenarios specifically includes: At the end of each basic cycle, the latest environmental factor vector is collected and subtracted from the environmental factor vector of the previous cycle to obtain the environmental change. If the change does not exceed the preset tolerance threshold, the current flight mission planning parameter set remains unchanged and the flight mission continues. When the environmental change is between the tolerance threshold and the preset warning threshold, only the flight speed and exposure interval parameters are fine-tuned synchronously: when wind speed increases or visibility decreases, the flight speed is proportionally reduced and the exposure interval is shortened; when wind speed decreases or light intensity increases, the flight speed is proportionally restored and the exposure interval is lengthened. The adjustment range is limited to the pre-set maximum adjustment limit for a single cycle. When the environmental change exceeds the warning threshold, the safety strategy is triggered: pause shooting or enter the hovering state, while maintaining the current posture and waiting for the environment to recover.

[0009] Preferably, the duration of the basic cycle is adjusted according to the terrain difference index of the target area to obtain the adaptive update cycle. The specific operations are as follows: Set a base period and two fixed proportional boundary values: the first proportional boundary value is equal to 1, which is used for the case of minimum difference, and the second proportional boundary value is between 0 and 1, which is used for the case of maximum difference; Read the calculated terrain difference index and map it to a periodic adjustment ratio k using linear interpolation, so that k decreases monotonically with the terrain difference index and satisfies the following conditions: when the terrain difference index is equal to 0, k is equal to 1 (i.e., the first ratio boundary value); when the terrain difference index is equal to 1, k is equal to the second ratio boundary value; The basic period is multiplied by the period adjustment ratio k to obtain the adaptation update period.

[0010] Preferably, the global adjustment operation for low consistency scenarios specifically includes: At the end of each adaptation update cycle, the current environmental factor vector and the terrain feature vector of the sub-area where the drone is located are first synchronously collected. The environmental factor vector and the terrain feature vector are combined and used as input to the mission parameter solution model to output a new flight mission planning parameter set. The new flight mission planning parameter set is then compared item by item with the parameter set executed in the previous cycle. If the change in any parameter exceeds the preset global change safety limit, the parameter is interpolated and smoothed in a linear decreasing manner until the change falls within the global change safety limit. The model output values ​​of other parameters remain unchanged. In subsequent adaptation update cycles, the drone continues to perform the shooting mission according to the new flight altitude, track spacing, heading overlap rate, side overlap rate, flight speed and exposure interval.

[0011] Preferably, the task parameter solving model is established based on a feedforward neural network, including: an input layer, several hidden layers and an output layer.

[0012] A three-dimensional terrain reconstruction system based on drone images, comprising: The information acquisition module is used to obtain environmental factors that can affect flight mission planning and the terrain characteristics of the target area, and organize them into environmental factor vectors and terrain feature vectors respectively; a terrain difference judgment module, configured to calculate a terrain difference index of a target area based on the terrain feature vector, compare the index with a preset difference threshold, and classify the target area into a high consistency scene or a low consistency scene based on the comparison result; The flight mission planning module is used to use the trained mission parameter solution model when the target area is a high-consistency scenario, take the environmental factor vector and the terrain feature vector as input, and output the flight mission planning parameter set. The UAV then uses the flight mission planning parameter set to perform the flight mission; during the flight of the UAV, the environmental factor vector is regularly updated at a preset basic cycle, and the flight mission planning parameter set is locally adjusted based on the changes in the environmental factor vector, and then sent for execution; when the target area is a low-consistency scenario, the duration of the basic cycle is first adjusted according to the terrain difference index of the target area to obtain an adaptive update cycle; then the terrain feature vector and the environmental factor vector of the sub-area where the UAV takes off are input into the mission parameter solution model to obtain the initial flight mission planning parameter set, and then the UAV uses the initial flight mission planning parameter set to perform the flight mission; during the flight of the UAV, the environmental factor vector is regularly updated at an adaptive update cycle, and the terrain feature vector of the sub-area where the UAV is currently located is obtained. Based on the changes in the environmental factor vector and the current terrain feature vector, the initial flight mission planning parameter set is globally adjusted and sent for execution; The terrain reconstruction module is used to perform image preprocessing, feature matching, sparse 3D reconstruction, dense matching, surface reconstruction and orthophotoization on all images in sequence after completing image acquisition, and finally output digital elevation models and orthophoto results.

[0013] The present invention has the following advantages: 1. Through a periodic adjustment mechanism driven by differential indicators and a neural network task parameter solution model, the present invention can dynamically update key parameters such as flight altitude, overlap rate, speed, and exposure interval during flight. When a sudden increase in wind speed, changes in lighting, or dramatic terrain fluctuations occur in the measurement area, the system automatically shortens the refresh cycle and performs global or local adjustments to promptly compensate for potential holes or blurred images, achieving high coverage and high-definition acquisition of complex terrain without the need for manual intervention or the addition of expensive hardware.

[0014] 2. The present invention maintains a longer basic update cycle for areas with low variability and relatively consistent terrain, and only performs limited fine-tuning when the environment fluctuates slightly, avoiding unnecessary multiple reshoots and parameter recalculations. At the same time, the model is frequently called for global optimization in areas with high variability, so that the amount of collected data matches the complexity of the scene. This strategy significantly reduces the size of redundant images and point clouds, alleviates the storage and computing pressure of the back-end 3D reconstruction, and improves overall operational efficiency and drone endurance utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a schematic structural diagram of a three-dimensional terrain reconstruction system based on drone images used in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0017] Example 1, a method for 3D terrain reconstruction based on drone images, comprising: Obtain environmental factors that can affect flight mission planning and target area terrain characteristics, and organize them into environmental factor vectors and terrain feature vectors respectively; environmental factor vectors include wind speed, visibility, light intensity, air pressure, temperature and relative humidity; terrain feature vectors include texture complexity, edge density, brightness variance and parallax amplitude; Calculating a terrain difference index of the target area based on the terrain feature vector, comparing the index with a preset difference threshold, and classifying the target area into a high consistency scene or a low consistency scene according to the comparison result; If the target area is a high-consistency scene, the trained mission parameter solution model is used, taking the environmental factor vector and terrain feature vector as input, and outputting a flight mission planning parameter set, including flight altitude, track spacing, heading overlap rate, side overlap rate, flight speed, and exposure interval. The UAV then uses the flight mission planning parameter set to execute the flight mission. During the UAV flight, the environmental factor vector is regularly updated at a preset basic cycle, and local adjustments to the flight mission planning parameter set are made based on changes in the environmental factor vector, and the flight mission planning parameter set is issued for execution. If the target area is a low consistency scenario, the duration of the basic cycle is first adjusted according to the terrain difference index of the target area to obtain the adaptive update cycle; then the terrain feature vector and environmental factor vector of the drone's takeoff sub-area are input into the mission parameter solution model to obtain the initial flight mission planning parameter set, and then the drone uses the initial flight mission planning parameter set to perform the flight mission; during the drone's flight, the environmental factor vector is regularly updated according to the adaptive update cycle, and the terrain feature vector of the sub-area where the drone is currently located is obtained. Based on the changes in the environmental factor vector and the current terrain feature vector, the initial flight mission planning parameter set is globally adjusted and issued for execution; After image acquisition is completed, all images are subjected to image preprocessing, feature matching, sparse 3D reconstruction, dense matching, surface reconstruction and orthophotoization in sequence, and finally the digital elevation model and orthophoto results are output.

[0018] Calculating a terrain difference index of the target area based on the terrain feature vector, comparing the index with a preset difference threshold, and classifying the target area into a high consistency scene or a low consistency scene according to the comparison result, specifically comprising: The grid side length is selected based on the expected ground resolution, so that the grid diagonal distance is no greater than half the planned heading spacing; the target area is divided into several sub-areas along the survey area projection coordinate system, and each sub-area is assigned a unique number and center coordinates to facilitate rapid positioning and data update during the flight process; For each sub-area, the pilot imagery acquired before the flight mission is executed is used to extract the following four types of terrain indicators that can be obtained based on the pilot imagery: Texture complexity index: construct a gray-level co-occurrence matrix and take the average value of the energy term and entropy term to represent the surface texture details; Edge density index: After extracting edges using an edge operator, calculate the ratio of the number of edge pixels to the total number of pixels; Brightness variance index: Calculates the square deviation of the gray value relative to the sub-region mean and takes the average to measure the brightness change; Disparity amplitude index: Run a fast optical flow algorithm between two overlapping pilot images, record the pixel displacement of the tracked feature points, and take the average absolute value of all pixel displacements as the disparity amplitude index; The four types of terrain indicators are normalized for their ranges. They are then linearly superimposed based on preset weight coefficients to obtain a single scalar as the terrain complexity score of the sub-region. The four-dimensional original indicator vector is saved for subsequent adjustments and quality traceability. The weights are determined by historical samples or expert experience. The terrain complexity scores of all sub-regions are grouped into a sample set. The set mean is first calculated, and then the absolute value of each score's deviation from the mean is calculated and averaged to obtain the average absolute deviation. The interquartile range of the sample set is also calculated, and the larger value between the average absolute deviation and the interquartile range is selected as the final difference indicator to suppress the influence of extreme values ​​on the judgment results. A difference threshold is set. If the terrain difference index is not greater than the difference threshold, the target area is judged to be a high-consistency scenario; if the terrain difference index is greater than the difference threshold, it is judged to be a low-consistency scenario. The threshold can be adaptively adjusted by professionals in related fields according to historical statistical data or operational experience using a rolling average strategy to take into account both universality and judgment stability under multi-regional conditions.

[0019] Local adjustments for high-consistency scenarios include: At the end of each basic cycle, the latest environmental factor vector is collected and subtracted from the environmental factor vector of the previous cycle to obtain the environmental change. If the change does not exceed a preset tolerance threshold, the current flight mission planning parameter set remains unchanged and the flight mission continues. The tolerance threshold can be determined by historical statistical data of similar operations or expert experience, reflecting the acceptable range of image quality caused by small environmental fluctuations. When the environmental change is between the tolerance threshold and the preset warning threshold, only the flight speed and exposure interval are fine-tuned synchronously. When wind speed increases or visibility decreases, the flight speed is proportionally reduced and the exposure interval is shortened. When wind speed decreases or light intensity increases, the flight speed is proportionally restored and the exposure interval is extended. The adjustment range is limited to the pre-set single-cycle maximum adjustment limit, and the flight altitude, track spacing, and overlap rate are not changed. The single-cycle maximum adjustment limit is determined by professionals in the relevant field during the system design phase, taking into account the dynamic response performance of the UAV platform, the gimbal stability margin, and the image clarity tolerance, and is written into the parameter area. When the limit is high, the system allows larger speed and exposure adjustments within a single basic cycle, which can enhance the system's adaptability to rapidly deteriorating environments, but the track tracking error increases accordingly. When the limit is low, the system only allows smaller adjustments to maximize track stability and track geometry consistency, but may cause a short-term decrease in image clarity in scenarios with sudden environmental changes. When the environmental change exceeds the warning threshold, the safety strategy is triggered: pause shooting or enter the hovering state, while maintaining the current posture and waiting for the environment to recover.

[0020] Adjust the duration of the basic cycle according to the terrain difference index of the target area to obtain the adaptive update cycle. The specific operations are as follows: A base period is set, and two fixed proportional boundary values ​​are set simultaneously: the first proportional boundary value is equal to 1 for the case of minimum variability, and the second proportional boundary value is between 0 and 1 for the case of maximum variability. The second proportional boundary value is determined by professionals in the relevant field during the system design phase and remains unchanged during a single flight mission; Read the calculated terrain difference index and map it to a periodic adjustment ratio k using linear interpolation, so that k decreases monotonically with the terrain difference index and satisfies the following conditions: when the terrain difference index is equal to 0, k is equal to 1 (i.e., the first ratio boundary value); when the terrain difference index is equal to 1, k is equal to the second ratio boundary value; The basic period is multiplied by the period adjustment ratio k to obtain the adaptation update period.

[0021] Global adjustment operations for low consistency scenarios include: At the end of each adaptation update cycle, the current environmental factor vector and the terrain feature vector of the sub-area where the drone is located are first synchronously collected. The environmental factor vector and the terrain feature vector are combined and used as input to the mission parameter solution model to output a new flight mission planning parameter set. The new flight mission planning parameter set is then compared item by item with the parameter set executed in the previous cycle. If the change in any parameter exceeds the preset global change safety limit, the parameter is interpolated and smoothed in a linear decreasing manner until the change falls within the global change safety limit. The model output values ​​of other parameters remain unchanged. In subsequent adaptation update cycles, the drone continues to perform the shooting mission according to the new flight altitude, track spacing, heading overlap rate, side overlap rate, flight speed and exposure interval.

[0022] The mission parameter solution model is established based on a feedforward neural network and includes: an input layer, several hidden layers, and an output layer. The input layer is used to simultaneously receive environmental factor vectors and terrain feature vectors and splice the two types of vectors into a unified feature sequence in a predetermined order. The first hidden layer performs a fully connected mapping on the feature sequence and completes the initial fusion of environmental information and terrain information. At least one subsequent hidden layer continues to perform nonlinear mapping and feature abstraction to learn the complex correspondence between the flight mission planning parameter set and the input features. The output layer is a fully connected structure with a one-to-one correspondence between the number of nodes and the number of parameters in the flight mission planning parameter set. It is used to directly output the flight altitude, track spacing, heading overlap rate, side overlap rate, flight speed, and exposure interval. The resulting output serves as the flight mission planning parameter set. The specific operations for training the task parameter solution model are as follows: Obtain several model training samples with labeled flight mission planning parameter sets, each model training sample contains the environmental factor vector and terrain feature vector of a certain historical moment in the historical flight record of the UAV; divide all model training samples into a training set and a validation set, use the training set to train the parameter-initialized mission parameter solution model, and then input the validation set into the mission parameter solution model for verification to obtain the verification result; set the training conditions, and determine whether the obtained verification result meets the training conditions. If so, output the trained mission parameter solution model; if not, continue to use the training set to train the mission parameter solution model.

[0023] Example 2, a 3D terrain reconstruction system based on drone images, such as Figure 1 Shown, including: The information acquisition module is used to obtain environmental factors that can affect flight mission planning and the terrain characteristics of the target area, and organize them into environmental factor vectors and terrain feature vectors respectively; a terrain difference judgment module, configured to calculate a terrain difference index of a target area based on the terrain feature vector, compare the index with a preset difference threshold, and classify the target area into a high consistency scene or a low consistency scene based on the comparison result; The flight mission planning module is used to use the trained mission parameter solution model when the target area is a high-consistency scenario, take the environmental factor vector and the terrain feature vector as input, and output the flight mission planning parameter set. The UAV then uses the flight mission planning parameter set to perform the flight mission; during the flight of the UAV, the environmental factor vector is regularly updated at a preset basic cycle, and the flight mission planning parameter set is locally adjusted based on the changes in the environmental factor vector, and then sent for execution; when the target area is a low-consistency scenario, the duration of the basic cycle is first adjusted according to the terrain difference index of the target area to obtain an adaptive update cycle; then the terrain feature vector and the environmental factor vector of the sub-area where the UAV takes off are input into the mission parameter solution model to obtain the initial flight mission planning parameter set, and then the UAV uses the initial flight mission planning parameter set to perform the flight mission; during the flight of the UAV, the environmental factor vector is regularly updated at an adaptive update cycle, and the terrain feature vector of the sub-area where the UAV is currently located is obtained. Based on the changes in the environmental factor vector and the current terrain feature vector, the initial flight mission planning parameter set is globally adjusted and sent for execution; The terrain reconstruction module is used to perform image preprocessing, feature matching, sparse 3D reconstruction, dense matching, surface reconstruction and orthophotoization on all images in sequence after completing image acquisition, and finally output digital elevation models and orthophoto results.

[0024] It should be understood that those skilled in the art may make improvements or modifications based on the above description, and all such improvements and modifications shall fall within the scope of protection of the appended claims. Any portion of this specification not described in detail is prior art known to those skilled in the art.

Claims

1. A three-dimensional terrain reconstruction method based on drone images, characterized in that: include: Obtain environmental factors that can affect flight mission planning and target area terrain features, and organize them into environmental factor vectors and terrain feature vectors respectively; Calculating a terrain difference index of the target area based on the terrain feature vector, comparing the index with a preset difference threshold, and classifying the target area into a high consistency scene or a low consistency scene according to the comparison result; If the target area is a high-consistency scene, the trained mission parameter solution model is used, taking the environmental factor vector and terrain feature vector as input, and outputting a flight mission planning parameter set, including flight altitude, track spacing, heading overlap rate, side overlap rate, flight speed, and exposure interval. The UAV then uses the flight mission planning parameter set to execute the flight mission. During the UAV flight, the environmental factor vector is regularly updated at a preset basic cycle, and local adjustments to the flight mission planning parameter set are made based on changes in the environmental factor vector, and the flight mission planning parameter set is issued for execution. If the target area is a low consistency scenario, the duration of the basic cycle is first adjusted according to the terrain difference index of the target area to obtain the adaptive update cycle; then the terrain feature vector and environmental factor vector of the drone's takeoff sub-area are input into the mission parameter solution model to obtain the initial flight mission planning parameter set, and then the drone uses the initial flight mission planning parameter set to perform the flight mission; during the drone's flight, the environmental factor vector is regularly updated according to the adaptive update cycle, and the terrain feature vector of the sub-area where the drone is currently located is obtained. Based on the changes in the environmental factor vector and the current terrain feature vector, the initial flight mission planning parameter set is globally adjusted and issued for execution; After image acquisition is completed, all images are subjected to image preprocessing, feature matching, sparse 3D reconstruction, dense matching, surface reconstruction and orthophotoization in sequence, and finally the digital elevation model and orthophoto results are output.

2. The three-dimensional terrain reconstruction method based on drone images according to claim 1, characterized in that: Calculating a terrain difference index of the target area based on the terrain feature vector, comparing the index with a preset difference threshold, and classifying the target area into a high consistency scene or a low consistency scene according to the comparison result, specifically comprising: The grid side length is selected based on the expected ground resolution, ensuring that the grid diagonal distance is no greater than half the planned heading spacing. The target area is divided into several sub-areas. For each sub-area, pilot images acquired before the flight mission are used to extract the following four types of terrain indicators that can be obtained from the pilot images: texture complexity, edge density, brightness variance, and parallax amplitude. The four types of terrain indicators mentioned above are normalized by their ranges. Then, they are linearly superimposed according to the preset weight coefficients to obtain a single scalar as the terrain complexity score of the sub-region. The terrain complexity scores of all sub-regions are combined into a sample set. The set mean is first calculated, and then the absolute value of each score's deviation from the mean is calculated and averaged to obtain the average absolute deviation. At the same time, the interquartile range of the sample set is calculated, and the larger value between the average absolute deviation and the interquartile range is selected as the final terrain difference index. Set a difference threshold. If the terrain difference index is not greater than the difference threshold, the target area is judged to be a high consistency scene; if the terrain difference index is greater than the difference threshold, it is judged to be a low consistency scene.

3. The method for 3D terrain reconstruction based on drone images according to claim 2, characterized in that: Local adjustments for high-consistency scenarios include: At the end of each basic cycle, the latest environmental factor vector is collected and subtracted from the environmental factor vector of the previous cycle to obtain the environmental change. If the change does not exceed the preset tolerance threshold, the current flight mission planning parameter set remains unchanged and the flight mission continues. When the environmental change is between the tolerance threshold and the preset warning threshold, only the flight speed and exposure interval parameters are fine-tuned synchronously: when wind speed increases or visibility decreases, the flight speed is proportionally reduced and the exposure interval is shortened; when wind speed decreases or light intensity increases, the flight speed is proportionally restored and the exposure interval is lengthened. The adjustment range is limited to the pre-set maximum adjustment limit for a single cycle. When the environmental change exceeds the warning threshold, the safety strategy is triggered: pause shooting or enter the hovering state, while maintaining the current posture and waiting for the environment to recover.

4. The method for 3D terrain reconstruction based on drone images according to claim 3, characterized in that: Adjust the duration of the basic cycle according to the terrain difference index of the target area to obtain the adaptive update cycle. The specific operations are as follows: Set a base period and two fixed proportional boundary values: the first proportional boundary value is equal to 1, which is used for the case of minimum difference, and the second proportional boundary value is between 0 and 1, which is used for the case of maximum difference; Read the calculated terrain difference index and map it to a periodic adjustment ratio k using linear interpolation, so that k decreases monotonically with the terrain difference index and satisfies the following conditions: when the terrain difference index is equal to 0, k is equal to 1 (i.e., the first ratio boundary value); when the terrain difference index is equal to 1, k is equal to the second ratio boundary value; The basic period is multiplied by the period adjustment ratio k to obtain the adaptation update period.

5. The method for 3D terrain reconstruction based on drone images according to claim 4, characterized in that: Global adjustment operations for low consistency scenarios include: At the end of each adaptation update cycle, the current environmental factor vector and the terrain feature vector of the sub-area where the drone is located are first synchronously collected. The environmental factor vector and the terrain feature vector are combined and used as input to the mission parameter solution model to output a new flight mission planning parameter set. The new flight mission planning parameter set is then compared item by item with the parameter set executed in the previous cycle. If the change in any parameter exceeds the preset global change safety limit, the parameter is interpolated and smoothed in a linear decreasing manner until the change falls within the global change safety limit. The model output values ​​of other parameters remain unchanged. In subsequent adaptation update cycles, the drone continues to perform the shooting mission according to the new flight altitude, track spacing, heading overlap rate, side overlap rate, flight speed and exposure interval.

6. The method for 3D terrain reconstruction based on drone images according to claim 5, characterized in that: The task parameter solution model is established based on a feedforward neural network, including: an input layer, several hidden layers and an output layer.

7. A three-dimensional terrain reconstruction system based on drone images, characterized in that: The system is applied to a three-dimensional terrain reconstruction method based on drone images as described in any one of claims 1 to 6, comprising: The information acquisition module is used to obtain environmental factors that can affect flight mission planning and the terrain characteristics of the target area, and organize them into environmental factor vectors and terrain feature vectors respectively; a terrain difference judgment module, configured to calculate a terrain difference index of a target area based on the terrain feature vector, compare the index with a preset difference threshold, and classify the target area into a high consistency scene or a low consistency scene based on the comparison result; The flight mission planning module is used to use the trained mission parameter solution model when the target area is a high-consistency scenario, take the environmental factor vector and the terrain feature vector as input, and output the flight mission planning parameter set. The UAV then uses the flight mission planning parameter set to perform the flight mission; during the flight of the UAV, the environmental factor vector is regularly updated at a preset basic cycle, and the flight mission planning parameter set is locally adjusted based on the changes in the environmental factor vector, and then sent for execution; when the target area is a low-consistency scenario, the duration of the basic cycle is first adjusted according to the terrain difference index of the target area to obtain an adaptive update cycle; then the terrain feature vector and the environmental factor vector of the sub-area where the UAV takes off are input into the mission parameter solution model to obtain the initial flight mission planning parameter set, and then the UAV uses the initial flight mission planning parameter set to perform the flight mission; during the flight of the UAV, the environmental factor vector is regularly updated at an adaptive update cycle, and the terrain feature vector of the sub-area where the UAV is currently located is obtained. Based on the changes in the environmental factor vector and the current terrain feature vector, the initial flight mission planning parameter set is globally adjusted and sent for execution; The terrain reconstruction module is used to perform image preprocessing, feature matching, sparse 3D reconstruction, dense matching, surface reconstruction and orthophotoization on all images in sequence after completing image acquisition, and finally output digital elevation models and orthophoto results.

Images