A control point-free low-altitude photogrammetry method for coastal terrain

Abstract

This invention discloses a low-altitude photogrammetry method for coastal topography without control points, comprising the following steps: 1) back-optimizing the optimal flight parameters with a preset accuracy index as the constraint target; 2) acquiring high-precision position and three-axis attitude at each exposure time; 3) determining the coastline by combining a 3D model with tidal level calculation. This invention provides an effective solution to the problems of difficulty in establishing coastal control points and the infeasibility of traditional point-laying methods due to frequent tidal fluctuations. By optimizing flight design, considering exposure delay, and employing high-precision positioning and attitude determination methods, this invention significantly improves the reliability of aerial photogrammetry results under control point-free conditions. By combining aerial photogrammetry and tidal level calculation to determine the coastline position, this method quickly identifies the coastline determination problem in current coastal topography surveys, solving the problem of inconsistent measurements when coastline traces are not obvious.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of surveying science and technology, and particularly relates to a low-altitude photogrammetry method for coast terrain without control points. BACKGROUND

[0002] The low-altitude photogrammetry of unmanned aerial vehicles (UAVs) is gradually applied to coast terrain measurement due to its comprehensive advantages of high efficiency, high precision, high flexibility and low cost. The traditional coast terrain measurement basically adopts manual field measurement mode, which is of high labor intensity, low operation efficiency and high safety risk. However, due to its high precision and easy implementation, the traditional coast terrain measurement is still the main measurement scheme for current coast terrain measurement. The determination of coastlines, the survey and measurement of port facilities and buildings, navigation aid markers, large-area dry beach and navigation obstacles in coast terrain measurement also have higher requirements for aerial photogrammetry technology. The current low-altitude photogrammetry method for coast terrain still has the following deficiencies:

[0003] (1) In order to ensure the high precision of the results, the traditional low-altitude aerial photogrammetry of UAVs needs to set up image control points. In the process of coast terrain measurement, large-area beach and coral reefs are submerged by water at high tide, so it is difficult to set up image control points and the measurement efficiency is extremely low.

[0004] (2) The instantaneous speed and instantaneous direction of the UAV at the exposure moment are not the same due to the fact that the UAV cannot maintain a uniform speed, which leads to inconsistent exposure delay errors of exposure points. The conventional GNSS-assisted bundle adjustment does not consider the error influence of exposure delay. SUMMARY

[0005] The purpose of the present application is to overcome the deficiencies of the prior art, provide a low-altitude photogrammetry method for coast terrain without control points, and use the low-altitude aerial photography mode of UAVs to collect element information in coast terrain measurement, so as to ensure the acquisition of high-precision coast terrain data under the condition of no control points, improve the operation efficiency, and meet the requirements of coast terrain measurement.

[0006] A low-altitude photogrammetry method for coast terrain without control points, comprising the following steps,

[0007] 1) reverse optimization of a preset precision index as a constraint target to solve optimal flight parameters, the optimal flight parameters including an exposure mode, a base-height ratio and a heading overlap, wherein the exposure mode adopts equal-distance interval exposure, the final heading overlap is 65%-80%, and the corresponding base-height ratio range is 0.35-0.60, wherein the initial heading overlap is designed to be 80-85%, and the aerial photographs are screened to obtain the final heading overlap according to a preset deletion rule;

[0008] 2) Obtain high-precision position and three-axis attitude at each exposure moment;

[0009] 3) Based on the assumption of uniform flight in the exposure delay time interval, calculate the position offset correction exposure delay deviation according to the actual exposure moment speed vector, and solve the exposure delay deviation and system error in the adjustment model to obtain high-precision image exterior orientation elements and encrypted point three-dimensional coordinates;

[0010] 4) Generate dense point cloud and generate digital surface model (DSM), combine digital surface model (DSM) to generate digital orthophoto map (DOM) by digital differential correction, generate 3D mesh model in OSGB format by poisson reconstruction or triangulation and texture mapping of dense point cloud, superimpose digital orthophoto map (DOM) and digital surface model (DSM) to construct a solid model, combine the mesh model to collect coastal terrain elements, and determine the coastline by combining the stereo model and the tide level.

[0011] As a specific embodiment of the present application, the preset deletion rule includes ensuring that the overlap of adjacent two images is 60%-80% and there is no image coverage blind area, rejecting aerial photographs with excessive inclination and blurred images, and removing aerial photographs with inaccurate positioning and attitude data.

[0012] As a specific embodiment of the present application, high-precision spatial positioning is realized by using differential GNSS technology combined with Beidou satellite navigation system receiver carrier phase observation, an unmanned aerial vehicle attitude accurate link is constructed by using high-performance IMU measurement, sensor fusion algorithm and flight control calculation mechanism, and the three-axis attitude of the unmanned aerial vehicle at each exposure moment is obtained by sequentially performing original attitude calculation based on IMU, sensor fusion attitude correction based on error state Kalman filter, linkage with Beidou satellite navigation system auxiliary information and final attitude output based on flight control calculation results.

[0013] As a specific embodiment of the present application, for the navigation obstacles and navigation aids in the coastal terrain, a rough three-dimensional model of the target area is first obtained, and then high-resolution images are collected based on the model to plan a flight route close to the flight route.

[0014] As a specific embodiment of the present application, the flight route close to the flight route is a surrounding flight, the base height ratio is 0.5-0.8, and the final heading overlap is 60%-70% and the lateral overlap is 70%-80%.

[0015] The advantages and beneficial effects of the present application are:

[0016] The non-control point coastal terrain low-altitude photogrammetry method considering exposure delay provided by the application can solve the difficulty of not being able to arrange image control points in a coastal terrain area, reduces the image control measurement link, greatly improves the operation efficiency, and realizes high-precision acquisition of element data in a complex terrain area. The technology provides an effective solution to the difficulty of not being able to arrange control points in a coastal zone and the problem of traditional point arrangement methods being infeasible due to frequent rise and fall of tidal water. The application can greatly improve the reliability of aerial photogrammetry results under non-control point conditions by optimizing flight design, considering exposure delay, and high-precision positioning and attitude determination. The combination of aerial photogrammetry and tidal level calculation to determine the position of a coastline can quickly determine the coastline in the current coastal terrain measurement, and can solve the problem of inconsistent measurement when the coastline trace is not obvious. DETAILED DESCRIPTION

[0017] In order for those skilled in the art to better understand the application scheme, the technical scheme of the application will be further described below in combination with specific embodiments.

[0018] The non-control point coastal terrain low-altitude photogrammetry method of the application comprises the following steps,

[0019] 1) Optimize the optimal flight parameters in reverse with preset accuracy indicators such as ground sampling distance (GSD) and model relative accuracy as constraint targets, seek the flight design parameter setting rules for realizing optimal positioning accuracy based on the photogrammetry principle and the stereo imaging model, so as to ensure the reliability of the results under non-control point conditions. The optimal flight parameters include exposure mode, base-height ratio, and heading overlap,

[0020] At present, the exposure modes of digital cameras are generally fixed-point exposure, equidistance exposure, and equal time interval exposure. When fixed-point exposure is adopted, the unmanned aerial vehicle will deviate from the flight path and miss the exposure point when affected by the outside world, which is easy to miss and results in insufficient overlap. When equal time interval exposure is adopted, the unmanned aerial vehicle will not use the same time to fly two flight strips under the influence of wind, and the overlap of adjacent two flight strips is different. The application adopts equal distance interval exposure to control the overlap of aerial photographs.

[0021] Based on the inverse relationship between elevation accuracy and base-height ratio, the heading overlap directly affects the baseline length under the condition of keeping the resolution constant (i.e. fixed flight height), the greater the overlap, the shorter the baseline, and the smaller the base-height ratio. Under the condition of a certain image resolution, the accuracy of unmanned aerial low-altitude photogrammetry can be improved by increasing the base-height ratio, so the overlap and base-height ratio need to be balanced during design.

[0022] The forward overlap should generally be between 60% and 80%, with a minimum of 53%. The lateral overlap should generally be between 15% and 60%, with a minimum of 8%. Even if the flight results meet these requirements, elevation accuracy often exceeds the limit under conditions without control points. Therefore, the forward overlap is designed to be 85%. Considering the influence of the base-to-elevation ratio on elevation, some aerial photos are deleted according to certain rules. Under the premise of ensuring the final overlap is met, aerial photos with ideal attitude are obtained.

[0023] Taking the DJI Mavic 3E drone as an example, with a flight altitude of 120m, a forward overlap rate of 80%, and a lateral overlap rate of 80%, the forward baseline length is 25.85m and the lateral baseline length is 34.60m. The base-to-height ratio is approximately 0.22, which can be increased. Reducing the forward overlap rate to 70% results in a forward baseline length of 38.78m and a base-to-height ratio of approximately 0.32. Further reducing the forward overlap rate to 65% results in a forward baseline length of 45.24m and a base-to-height ratio of 0.38. The design of the forward overlap rate needs to be based on changes in coastal topography. For densely built-up port terrain, both the forward overlap and flight altitude need to be increased, with the forward overlap rate not less than 75%. For dry beaches or ordinary coastlines with relatively flat terrain, the overlap rate can be reduced to increase the base-to-height ratio. Therefore, a forward overlap rate of 65%-70% can be designed, ensuring sufficient matching points for the images while also increasing the base-to-height ratio and guaranteeing measurement accuracy.

[0024] The aerial photograph deletion rules are as follows:

[0025] 1) Ensure that the overlap between two adjacent images is between 60% and 80%, and that there are no areas not covered by the images. Adjust the flight path by deleting aerial photos to ensure that the area is fully covered without omissions, while also ensuring that the base-to-height ratio is reasonable, thereby ensuring the accuracy of aerial photography.

[0026] 2) While ensuring the degree of overlap, remove aerial photographs with excessive tilt angles or those that are blurry.

[0027] 3) Remove aerial photographs with inaccurate positioning and attitude determination.

[0028] Among the rules for excluding aerial photographs with tilt angles exceeding the limit, the number of photographs exceeding 8° should not exceed 10% of the total. In particularly difficult areas: the tilt angle should not exceed 8°, the maximum should not exceed 15°, and the number of photographs exceeding 10° should not exceed 10% of the total.

[0029] When removing inaccurate positioning and attitude determination, the POS position obtained from the output data is mainly based on the results. If the positioning accuracy is not high, there will be log markings. For example, high accuracy can be obtained when RTK is fixed, but the accuracy will decrease when floating. Therefore, aerial photos taken when RTK is fixed should be selected.

[0030] Set the base-to-height ratio as , B is the baseline length, H is the flight altitude, and the heading overlap is set to... If the image size is L, then While maintaining a constant resolution (i.e., a fixed flight altitude), forward overlap directly affects the baseline length. Greater overlap results in a shorter baseline and a smaller base-to-elevation ratio. An excessively large base-to-elevation ratio increases the visual difference between adjacent images, leading to increased occlusion areas, difficulty in image matching, and reduced reliability. This invention strikes a balance between forward overlap and base-to-elevation ratio, keeping them within an optimized range. In particular, the strategy of reducing forward overlap through image deletion coordinates the conflict between shooting and data processing, improving overall shooting quality.

[0031] 2) Obtain high-precision position and three-axis attitude at each exposure moment;

[0032] Specifically, to achieve centimeter-level positioning accuracy to meet the requirements of surveying and mapping applications, this invention employs differential GNSS technology (RTK / PPK) combined with carrier phase observations from the BeiDou Navigation Satellite System receiver (BDS) for high-precision spatial positioning. Traditional GNSS single-point positioning (SPP) is based solely on pseudorange observations and is affected by various error sources such as multipath propagation, ionospheric and tropospheric delays, satellite clock errors, and receiver clock errors. Its typical positioning accuracy is usually at the meter level, which is insufficient to meet the high-precision POS calculation requirements of image-free UAVs. This invention utilizes carrier phase measurement and reference station differential technology, employing two centimeter-level positioning strategies: RTK (Real-Time Kinematic) and PPK (Post-Processed Kinematic), to ensure the high-precision controllability of the flight platform's trajectory. RTK (Real-Time Kinematic) real-time dynamic differential technology relies on a reference station with precise coordinates. It receives satellite observations in real time and calculates carrier phase corrections, transmitting the correction information to the UAV rover station in real time via radio or mobile network. The rover station uses these corrections to refine its observations, thereby calculating centimeter-level coordinates in real time. In areas with good network conditions or base station coverage, RTK can achieve continuous and stable real-time centimeter-level positioning, and is therefore commonly used for tasks such as automated flight path planning and high-precision attitude control. However, this method is highly dependent on communication links, and its availability may be limited in real-world scenarios such as mountainous areas, power lines, or remote locations. In contrast, PPK (Post-Processed Kinematic) dynamic post-processing differential technology does not rely on any real-time data transmission. Its core idea is to simultaneously record the raw carrier phase observation data from the base station and the airborne receiver during flight, and then perform offline joint calculations after the mission is completed. PPK effectively eliminates system errors through double-difference processing and can improve the success rate of ambiguity fixation through post-processing.

[0033] This invention obtains high-precision POS data that meets the requirements of UAV (Unmanned Aerial Vehicle) operation without control points through this differential enhancement BeiDou positioning strategy, providing reliable support for subsequent trajectory optimization and 3D reconstruction.

[0034] Furthermore, in low-altitude photogrammetry using unmanned aerial vehicles (UAVs) without control points, accurately acquiring the platform's attitude parameters (roll, pitch, yaw) at each exposure moment is crucial for ensuring that aerial survey results achieve mapping-grade accuracy. To this end, this invention employs a high-performance IMU measurement and sensor fusion algorithm combined with a flight control calculation mechanism to construct a precise attitude link for the UAV, achieving accurate quantification of dynamic attitude changes.

[0035] Through the entire process described above—obtaining high-precision position based on the Beidou Navigation Satellite System (BDS) and high-precision attitude based on IMU and multi-sensor fusion algorithms—the UAV platform can ultimately obtain high-precision position solutions and stable, reliable three-axis attitude solutions at every exposure moment. This high-precision, time-synchronized, and continuous pose sequence allows the accuracy of exterior orientation elements to be obtained entirely autonomously by the onboard sensors, without relying on ground-based image control points to complete the geometric constraints during aerial photography, further meeting the application requirements of high-precision mapping, 3D reconstruction, and topographic surveying.

[0036] 3) Based on the assumption of uniform flight within the exposure delay time interval, the position offset is calculated to correct the exposure delay deviation by using the velocity vector at the actual exposure time through Lagrange interpolation; the exposure delay deviation and system error are substituted into the adjustment model for joint solution to obtain the high-precision image exterior orientation elements and the three-dimensional coordinates of densification points.

[0037] Specifically, due to the impact of camera exposure delay, the recording by the BeiDou Navigation Satellite System (BDS) and the aerial camera's troubleshooting operations are out of sync. The instantaneous coordinates recorded by the BDS at the moment of camera exposure during the BDS recording operation are related to the coordinates at which the BDS begins recording after receiving the signal from the flight control system, as follows:

[0038]

[0039] This indicates the coordinates received by the BeiDou Navigation Satellite System (BDS) at the actual exposure time of the camera; This indicates the coordinates measured when the BeiDou Navigation Satellite System (BDS) receives the signal from the flight control system and begins recording. This represents the position offset vector between the recorded exposure time and the actual exposure time. Due to the small mass and low inertia of the drone, it is easily affected by airflow during flight operations, making it difficult to maintain a constant speed. Therefore, the offset vector is different at each camera exposure time. However, the exposure delay time from signal reception to exposure is relatively short each time, and the aircraft can be considered to be in a constant-speed flight state during the exposure delay time. Therefore, the position correction is:

[0040]

[0041] In the formula To record the flight velocity vector at the exposure moment, the flight velocity vector is obtained from the velocity vector of continuous moments recorded by the BeiDou Navigation Satellite System (BDS) receiver software, and then the instantaneous velocity vector at the actual exposure moment is obtained by interpolating the instantaneous exposure moment recorded by the BeiDou Navigation Satellite System (BDS) using the Lagrange polynomial interpolation method. This is the exposure delay time.

[0042] Establish an exposure delay model:

[0043]

[0044] Indicates the location of the center of the photograph at the moment of true exposure. Indicates the position of the phase center of the BeiDou antenna; This represents the coordinates of the phase center of the BDS antenna in the image space coordinate system of the BeiDou Navigation Satellite System. This represents the orthogonal transformation matrix constructed using the angular elements of the image's exterior orientation elements. Combining the geometric relationship between the center of the actual exposure time and the phase center of the BeiDou Navigation Satellite System (BDS) antenna at the actual exposure time, the exposure delay model for the shooting time is established as follows:

[0045]

[0046] This represents the velocity vector of the unmanned aerial platform in three directions at the moment of exposure; This is the exposure delay time.

[0047] An adjustment model that takes exposure delay into account is established. The coordinates of the camera station recording the exposure time are used as BDS observations to establish error equations. These error equations are then added to the error equation system of the adjustment mathematical model.

[0048]

[0049] in Representing image points Coordinate values ​​in a spatial coordinate system These represent projection coefficients that cancel each other out during the solution process. It is the rotation matrix transpose of the camera's exterior orientation elements. It is the camera's principal distance. It is the auxiliary coordinate system of image space corresponding to the image point.

[0050]

[0051] Considering the characteristics of low-altitude photogrammetry operations using UAVs, this method takes into account the camera's exposure delay time and positional offset at each exposure moment, and performs point-by-point exposure compensation for the exposure delay error at each exposure point. Compared to traditional methods, this method focuses on point-by-point exposure compensation for the exposure delay error at each exposure point. It extends and expands upon the theoretical basis of the traditional BDS-assisted bundle adjustment model for the BeiDou Navigation Satellite System, making it more suitable for UAV-based BDS-assisted bundle adjustment. During the model solution process, the exposure delay is used as a solution for positional errors and substituted into the equations along with other system errors for unified solution, eliminating the error influence caused by different exposures, thereby improving the accuracy of BDS-assisted aerial triangulation adjustment for the BeiDou Navigation Satellite System.

[0052] 4) Based on the above-mentioned calculated data, a dense point cloud is generated, and a Digital Surface Model (DSM) is created. The DSM is then combined with digital differential correction to generate a Digital Orthophoto Map (DOM). The point cloud is reconstructed or triangulated using Poisson regression and texture mapping to generate a 3D Mesh Model in OSGB format. The DOM and DSM are overlaid to construct a stereo model. Coastal topographic features are collected using the Mesh model, and the coastline is determined by combining the stereo model with tidal level calculations. The coastline is the trace line formed during the ebb tide of average spring tides. Generally, the coastline location is determined by observing the ebb tide traces on-site, such as vegetation, shellfish remains, and rock soaking marks. Based on the acquired coastal topographic elevation features of the area, the location of the coastline can be determined by estimating the water level at spring tide.

[0053] Low-altitude photogrammetry products from UAVs used for coastal topographic feature acquisition require the construction of DOM, DSM, and Mesh models. A stereoscopic measurement model is built based on the DOM and DSM, and coastal topographic feature acquisition is then carried out on both the stereoscopic and Mesh models. The specific technical solution is as follows:

[0054] Firstly, based on existing technology, the data obtained above is used to produce DSM, DOM, and Mesh model products, allowing users to perform stereoscopic observation and 3D measurement on a regular screen without the need for professional stereoscopic glasses. Coastal topographic feature acquisition includes coastline measurement, landforms above and below the coastline, coastal port facilities, navigational aids, and navigational obstacles. Coastline measurement utilizes aerial photogrammetry to obtain a stereoscopic measurement model and tide level calculations to determine the coastline location. Landform measurement uses a stereoscopic measurement model to collect all landform features, including coastal port facilities, rocky beaches, sandy beaches, and seawalls. Navigational aid measurement uses a 3D Mesh model to collect attributes such as height, position, and texture of navigational aids. Navigational obstacle measurement uses a 3D Mesh model to measure the position and elevation of navigational obstacles such as exposed reefs and dry reefs.

[0055] This invention provides a low-altitude photogrammetry method for coastal topography without control points, taking into account exposure delay. This method overcomes the difficulty of establishing control points in coastal areas, reducing the number of control measurement steps and significantly improving operational efficiency while achieving high-precision acquisition of feature data in complex terrain areas. This technology offers an effective solution to the challenges of establishing control points along the coast and the impracticality of traditional point placement methods due to frequent tidal fluctuations. By optimizing flight design, considering exposure delay, and employing high-precision positioning and attitude determination methods, this invention significantly improves the reliability of aerial photogrammetry results under control-point-free conditions. By combining aerial photogrammetry and tidal level estimation to quickly determine the coastline location in current coastal topographic surveys, it solves the problem of inconsistent measurements when coastline traces are not obvious.

[0056] Furthermore, considering the measurement requirements for navigational obstacles and navigational aids in coastal topographic surveying, the flight design for UAV low-altitude photogrammetry needs to adopt a close-range photogrammetry scheme. For some unconventional ground surfaces or complex man-made object surfaces requiring detailed 3D reconstruction, vertical aerial photogrammetry and oblique photogrammetry suffer from problems such as insufficient detail or redundant image data. Close-range photogrammetry is an object-oriented photogrammetry method. The photographed object is the target's "surface," such as any slope surface in 3D space or the facade of a building, thereby extracting detailed target surface information for subsequent 3D reconstruction. Close-range photogrammetry requires the UAV's photographic path to follow the surface of the object being photographed, and the UAV's flight angle or the camera's shooting angle needs to be adjusted according to the target surface shape to obtain a high-resolution image of the target surface. The specific process of close-range photogrammetry, from coarse to fine, is as follows: first, a rough 3D model of the target area is obtained using conventional UAV photography; then, based on the initial model information, a close-range UAV flight path is planned and high-resolution images are acquired; finally, the images are processed for 3D reconstruction to generate high-quality, detailed geographic information. The essence of close-range photogrammetry lies in its "object-oriented" nature, namely, the flight of UAVs along the target surface and the photographing of that surface. Therefore, close-range photogrammetry flight path planning refers to designing continuous flight paths and photographic directions that meet the requirements of fine 3D reconstruction in photogrammetry, based on the surface distribution and flight space constraints of the reconstructed object, through photogrammetric geometric calculations and spatial analysis. This provides flight control parameters for the UAV, enabling automated flight and photographic maneuvers. Close-range photogrammetry flight path planning is a planning method based on initial target model information. It requires obtaining the shape information of the target scene beforehand and using this as the basis for subsequent flight path planning. This initial scene information can be a basic rough model in a geographic information database, or an initial rough model generated by conventional photogrammetry methods or manually controlled UAV flight. Conventional flight paths are horizontal and ground-based, while close-range flight paths are non-horizontal and circle the target object. The specific planning is as follows: Based on preliminary surveys, a rough 3D model is obtained to determine the size and extent of the target object; then, risks are identified, and all potential obstacles and danger zones are marked in the rough 3D model; finally, the base-to-height ratio and overlap are designed. Due to the extremely small height H, the base-to-height ratio B / H significantly affects elevation accuracy. A high base-to-height ratio, such as 0.5-0.8, is typically required, but this necessitates a reduced forward overlap, for example, 60%-70%, to increase the baseline length B. A careful trade-off must be struck between overlap and base-to-height ratio; initially, a high forward overlap, such as 85%, can be used, then reduced to 65% by deleting images to ensure successful matching. Simultaneously, a high lateral overlap, such as 70%-80%, is maintained to increase redundancy and handle more complex terrain features at low altitudes. Finally, a circumnavigation flight is executed, where waypoints are set in the flight path planning software to construct the circumnavigation mission.

[0057] This technology enables efficient, rapid, and clear 3D reconstruction of isolated landmarks such as islands, reefs, and navigational aids located far from the shore, providing information on their height, location, and structure. It addresses the challenges of stable modeling and inaccurate positioning of isolated islands and reefs far from land and where control points are difficult to establish, a problem inherent in traditional methods. For special coastal areas such as large-scale subsidence mudflats and coral reefs, this method allows for rapid acquisition of geographic information data and micro-topographic features, offering convenience and efficiency. Furthermore, it solves problems such as weak surface texture in mudflat areas, instability in conventional aerial triangulation, and difficulties in feature point extraction due to blurred boundary lines in coral reef areas.

[0058] To verify the method of this invention, a port area of ​​1.5 square kilometers and a large dry beach area were selected as test areas. This area includes various typical features such as coastline, obstacles, navigation aids, residential areas, ports, green belts, and water bodies. A DJI Matrice 4E drone was used for data collection. The flight altitude was set to 120 meters, and the forward overlap and lateral overlap were pre-set to 85% and 75%, respectively. After data collection, images with excessive water area that could not be matched were deleted according to image deletion rules, as well as images with RTK positioning lock-out. While ensuring full coverage of the survey area, the forward overlap rate was adjusted to 60-70% to ensure sufficient model connection points and matching accuracy. The BeiDou receiver on the drone recorded differential POS data in real time.

[0059] To verify the effectiveness of the method of this invention, approximately 20 evenly distributed high-precision GNSS checkpoints were set up in two survey areas. The coastal topographic feature data collected using this invention were compared with those collected at the checkpoints. The mean square errors in horizontal position were ±7cm and ±7cm, respectively, and the mean square errors in elevation were ±8cm and ±16cm, respectively, fully meeting the accuracy requirements of hydrographic surveying specifications. This fully demonstrates the reliability, efficiency, and portability of the method of this invention in measuring coastal topography.

[0060] The present invention has been described above by way of example. It should be noted that any simple modifications, alterations or other equivalent substitutions that can be made by those skilled in the art without creative effort without departing from the core of the present invention fall within the protection scope of the present invention.

Claims

1. A method for low-altitude photogrammetry of coastal topography without control points, characterized in that, Includes the following steps, 1) The optimal flight parameters are solved by reverse optimization with the preset accuracy index as the constraint target. The optimal flight parameters include exposure mode, base height ratio and forward overlap. The exposure mode adopts equal interval exposure, and the final forward overlap is 65%-80%. The corresponding base height ratio range is 0.35~0.

60. The initial forward overlap is designed to be 80-85%. The final forward overlap is obtained by filtering the aerial images according to the preset deletion rules. 2) Obtain high-precision position and three-axis attitude at each exposure moment; 3) Based on the assumption of uniform flight within the exposure delay time interval, the position offset is calculated according to the velocity vector at the actual exposure time to correct the exposure delay deviation. The exposure delay deviation and system error are substituted into the adjustment model for joint solution to obtain the high-precision image exterior orientation elements and the three-dimensional coordinates of densification points. 4) Generate a dense point cloud and a DSM. Combine the DSM with digital differential correction to generate a DOM. The dense point cloud is reconstructed or triangulated and textured to generate an OSGB format 3D Mesh model. The DOM and DSM are superimposed to construct a stereo model. Coastal topographic features are collected in combination with the Mesh model. The coastline is determined by combining the stereo model with tide level calculation.

2. The method for low-altitude photogrammetry of coastal topography without control points according to claim 1, characterized in that, The preset deletion rules include ensuring that the overlap between two adjacent images is between 60% and 80% and there are no blind spots in the image coverage, removing aerial images with excessive tilt angles and blurry images, and removing aerial images with inaccurate positioning and attitude data.

3. The method for low-altitude photogrammetry of coastal topography without control points according to claim 1, characterized in that, High-precision spatial positioning is achieved by using differential GNSS technology combined with carrier phase observation from the BeiDou satellite navigation system receiver. A precise attitude link for the UAV is constructed using high-performance IMU measurement, sensor fusion algorithms, and flight control calculation mechanisms. The process involves sequentially performing raw attitude estimation based on the IMU, sensor fusion attitude correction based on error state Kalman filtering, linkage with auxiliary information from the BeiDou satellite navigation system, and final attitude output based on the flight control calculation results, thus obtaining the three-axis attitude of the UAV at each exposure moment.

4. The method for low-altitude photogrammetry of coastal topography without control points according to claim 1, characterized in that, To address navigational obstacles and navigational aids in the coastal terrain, a rough 3D model of the target area is first obtained. Then, based on this model, a flight path is planned to closely follow the target area, and high-resolution images are acquired.

5. The method for low-altitude photogrammetry of coastal topography without control points according to claim 4, characterized in that, The planned close approach flight route is a circular flight, with a base-to-altitude ratio of 0.5-0.8, and the final forward overlap is 60%-70%, while the lateral overlap is 70%-80%.

Images