A coastal zone measurement method and system based on aerial image and laser point data fusion

Abstract

The present application provides a kind of based on aerial image and laser point cloud data fusion's coastal zone measurement method and system, comprising the following steps: using unmanned aerial vehicle to carry integrated system obtains LiDAR point cloud data and aerial remote sensing image;Remote sensing image is pretreated;The machine-mounted laser sounding data is pretreated;Depth inversion model is constructed;The model is applied to remote sensing image depth inversion;Output target water area depth data and high-precision three-dimensional seabed topographic map.The present application can improve the precision and efficiency of coastal zone detection, while obtaining high-precision seabed topographic data, solve the problem of multi-system fusion light and small, can realize the depth of island reef and other shallow water area Rapid measurement and can improve the sounding precision.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of ocean technology, and particularly relates to a coastal zone measurement method and system based on fusion of aerial images and laser point cloud data. BACKGROUND

[0002] The coastal zone is the area where land and sea meet, and has important ecological, economic and strategic value. The detection and monitoring of the coastal zone are important contents of marine scientific research and marine resource development and utilization. At present, the commonly used methods for detecting the coastal zone mainly include the following:

[0003] The method based on remote sensing satellites: the remote sensing satellites are used to obtain image data of the coastal zone, and then image processing and analysis are performed to extract feature information and perform water depth inversion. This method has the advantages of wide coverage, good periodicity, low cost, etc., but also has the disadvantages of low resolution, great influence of weather, poor real-time performance, etc.

[0004] The method based on unmanned aerial vehicles: various sensors such as visible light cameras, infrared cameras, radars, etc. are carried by unmanned aerial vehicles to detect and photograph the coastal zone, and then image processing and analysis are performed to extract feature information and perform water depth inversion. This method has the advantages of high resolution, good real-time performance, strong flexibility, etc., but also has the disadvantages of low flight height, great interference, short endurance time, etc.

[0005] The method based on sonar: the sonar emits sound waves, receives the reflected sound wave signals, and detects the coastal zone according to the intensity, frequency, phase, etc. of the sound wave signals. This method has the advantages of being suitable for underwater environment, strong anti-interference capability, etc., but also has the disadvantages of low resolution, great influence of water flow, large amount of calculation, etc.

[0006] The above methods each have advantages and disadvantages, but all have a common problem, i.e. it is difficult to accurately and quickly identify unknown or complex sea targets.

[0007] Therefore, it is very meaningful to propose a coastal zone measurement method and system based on fusion of aerial images and laser point cloud data. SUMMARY

[0008] The present application provides a coastal zone measurement method and system based on fusion of aerial images and laser point cloud data to solve the above technical defects.

[0009] In a first aspect, the present application proposes a coastal zone measurement method based on fusion of aerial images and laser point cloud data, which comprises the following steps:

[0010] S1, using an unmanned aerial vehicle to carry an integrated system to detect and collect the coastal zone of a specified area to obtain LiDAR point cloud data and aerial remote sensing images;

[0011] S2, preprocessing the acquired remote sensing image, including radiation correction and geometric correction, the radiation correction eliminating the influence of sensor characteristics and atmospheric conditions on the image by using a correction algorithm based on a physical model, and the geometric correction accurately aligning the image data with a known coordinate system using high-precision ground control points and image matching technology to ensure spatial consistency of the data;

[0012] S3, preprocessing the acquired LiDAR point cloud data, including data preprocessing, waveform data processing, error correction and point cloud data processing, to obtain high-precision water depth data on the measurement route; further using the measured high-precision water depth points as control points, combining the gray values of the aerial photogrammetry image to construct a water depth inversion model;

[0013] S4, using the processed laser data and image data, based on the constructed water depth inversion model, performing water depth inversion, optimizing the model by introducing a machine learning algorithm, and generating high-precision seabed topographic data using three-dimensional reconstruction technology; and

[0014] S5, further outputting the water depth data and high-precision three-dimensional seabed topographic map of the target water area, outputting the generated topographic data as a high-precision three-dimensional seabed topographic map in a graphical or textual manner and displaying it on a terminal device for user viewing and analysis.

[0015] Preferably, in step S3, further comprising:

[0016] An algorithm is used which uses the difference between the radiance brightness of the shallow water area pixels and the radiance brightness of the deep water area pixels in the same image as a logarithmic linear transformation, only using a single band to explain the exponential decay of visible light in water, and then further extending the linear model to dual-band and multi-band, the algorithm formula is: where a0 and a i are constant coefficients, N is the number of bands participating in inversion, R i is the gray value of the band, (R ∞ ) i is the gray value of the optical deep water area pixel adjacent to the shallow water area.

[0017] Further preferably, X i = ln(R i -(R ∞ ) i ), combined with the panchromatic band image of aerial photogrammetry, the water depth inversion model is constructed as: z = a0 + a1ln(R 全色 -(R ∞ ) 全色), input the high-precision water depth extracted from the laser radar echo waveform or point cloud into the water depth inversion model, calculate the parameters a0 and a1 of the model, and thus complete the construction of the water depth inversion model.

[0018] Preferably, in step S1, the integrated system comprises a visible light camera and an airborne laser radar, and the integrated system can synchronously acquire remote sensing images and laser sounding data of the coastal zone of a specified area, and the unmanned aerial vehicle has the function of adjusting flight parameters to adapt to different coastal zone measurement environments.

[0019] Preferably, in step S3, the data preprocessing removes outliers and noises through a preset data cleaning algorithm; the waveform data processing uses a preset waveform decomposition and feature extraction technology to improve the accuracy of waveform recognition and the precision of depth calculation; the error correction is accurately corrected based on real-time environmental monitoring data and a sensor error model; and the point cloud data processing uses an optimized point cloud filtering and classification algorithm to automatically adjust filtering parameters according to the topographic features of the coastal zone, to obtain high-precision water depth data on the measurement route, and the precision of the point cloud data processed is in the centimeter level.

[0020] Preferably, in step S4, the machine learning algorithm comprises a neural network algorithm, a linear regression algorithm and a logistic regression algorithm.

[0021] Preferably, in step S5, the high-precision three-dimensional seabed topographic map output also supports multiple formats for export and interactive operation.

[0022] In a second aspect, an embodiment of the present application provides a coastal zone measurement system based on fusion of aerial images and laser point cloud data, comprising:

[0023] A data acquisition module configured to detect and collect the coastal zone of a specified area by using an unmanned aerial vehicle carrying an integrated system, to obtain LiDAR point cloud data and aerial remote sensing images;

[0024] An image preprocessing module configured to preprocess the obtained remote sensing images, including radiation correction and geometric correction, the radiation correction eliminates the influence of sensor characteristics and atmospheric conditions on the images by using a correction algorithm based on a physical model, and the geometric correction uses high-precision ground control points and image matching technology to accurately align the image data with a known coordinate system, to ensure the spatial consistency of the data;

[0025] A laser point cloud preprocessing module configured to preprocess the obtained LiDAR point cloud data, including data preprocessing, waveform data processing, error correction and point cloud data processing, to obtain high-precision water depth data on the measurement route;

[0026] The water depth inversion construction module is configured to take the measured high-precision water depth points as control points, combine the gray values of the aerial photogrammetry images, and construct a water depth inversion model;

[0027] The model application module is configured to use the processed laser data and image data, perform water depth inversion based on the constructed water depth inversion model, optimize the model by introducing a machine learning algorithm, and generate high-precision seabed topographic data by using a three-dimensional reconstruction technology.

[0028] The display result module is configured to output the water depth data of the target water area and the high-precision three-dimensional seabed topographic map, output the generated topographic data as a high-precision three-dimensional seabed topographic map in a graphical or textual manner, and display the topographic map on a terminal device for user viewing and analysis.

[0029] Compared with the prior art, the beneficial results of the present application are as follows:

[0030] The present application can improve the accuracy and efficiency of coastal zone detection, obtain high-precision seabed topographic data, solve the problem of multi-system fusion miniaturization, realize rapid water depth measurement in shallow water areas such as islands, and improve the depth measurement accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0031] The accompanying drawings are included to provide a further understanding of embodiments and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments and, together with the description, serve to explain the principles of the present application. Other embodiments and many of the intended advantages of the present application will be readily appreciated as the same becomes better understood by reference to the following detailed description. The elements of the drawings are not necessarily to scale relative to each other. Like reference numerals designate corresponding similar parts.

[0032] Figure 1 The flowchart of a coastal zone measurement method based on aerial image and laser point cloud data fusion according to an embodiment of the present application;

[0033] Figure 2 The laser point cloud and aerial image water depth inversion flowchart according to a specific embodiment of the present application;

[0034] Figure 3 The schematic diagram of the system architecture of a coastal zone measurement system based on aerial image and laser point cloud data fusion according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] The present application will be further described below in conjunction with the drawings and embodiments. It can be understood that the specific embodiments described herein are only used to explain the related application, and not to limit the application. In addition, it should be noted that only the parts related to the application are shown in the drawings for ease of description.

[0036] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0037] Figure 1 An embodiment of the present application discloses a coastal zone measurement method based on fusion of aerial images and laser point cloud data, as shown in Figure 1 The method comprises the following steps:

[0038] S1, using a UAV integrated system to detect and collect the coastal zone of a specified area to obtain LiDAR point cloud data and aerial remote sensing images;

[0039] Specifically, the integrated system includes a visible light camera and an airborne laser radar, and the integrated system can synchronously obtain remote sensing images and laser sounding data of the coastal zone of the specified area. The UAV has the function of adjustable flight parameters to adapt to different coastal zone measurement environments.

[0040] S2, preprocessing the obtained remote sensing images, including radiation correction and geometric correction, the radiation correction eliminates the influence of sensor characteristics and atmospheric conditions on the images by using a correction algorithm based on a physical model, and the geometric correction uses high-precision ground control points and image matching technology to accurately align the image data with the known coordinate system, ensuring the spatial consistency of the data;

[0041] S3, preprocessing the obtained LiDAR point cloud data, including data preprocessing, waveform data processing, error correction and point cloud data processing, to obtain high-precision water depth data on the measurement route; further using the measured high-precision water depth points as control points, combining the gray value of the aerial photogrammetry image, and constructing a water depth inversion model;

[0042] Further, step S3 further comprises: using an algorithm that uses the difference between the radiance brightness of the shallow water area pixels and the radiance brightness of the deep water area pixels in the same image as a logarithmic linear transformation, using only a single band to explain the exponential decay of visible light in water, and then further extending the linear model to dual-band and multi-band, the algorithm formula is: Wherein, a0 and a i are constant coefficients, N is the number of bands participating in inversion, R i is the gray value of the band, (R ∞ ) i is the gray value of the optical deep water area pixel adjacent to the shallow water area.

[0043] Let X i = ln(R i -(R ∞ )i ), combined with the panchromatic band image of aerial photogrammetry, the water depth inversion model is constructed as: z = a0 + a1ln(R 全色 -(R ∞ ) 全色 ), the high-precision water depth extracted from the laser radar echo waveform or point cloud is input into the water depth inversion model, and the parameters a0 and a1 of the model are calculated, that is, the construction of the water depth inversion model is completed.

[0044] Wherein, the data preprocessing removes outliers and noises through a preset data cleaning algorithm; the waveform data processing improves the accuracy of waveform recognition and the accuracy of depth calculation by using a preset waveform decomposition and feature extraction technology; the error correction is accurately corrected based on real-time environmental monitoring data and sensor error model; the point cloud data processing adopts an optimized point cloud filtering and classification algorithm, automatically adjusts the filtering parameters according to the coastal zone terrain characteristics, obtains high-precision water depth data on the measurement route, and the accuracy after the point cloud data processing is in centimeter level.

[0045] S4, using the processed laser data and image data, based on the constructed water depth inversion model, water depth inversion is carried out, the model is optimized by introducing a machine learning algorithm, and high-precision seabed topographic data is generated by using three-dimensional reconstruction technology; and

[0046] Specifically, the machine learning algorithm includes neural network algorithm, linear regression algorithm and logistic regression algorithm. Machine learning algorithm is a kind of algorithm that can let computer automatically learn pattern and rule from data, and use these learned knowledge to make prediction, classification, clustering or other tasks. Through the analysis and training of a large amount of data, the model parameters of these algorithms are constantly optimized to improve the processing ability of unknown data.

[0047] S5, further output the water depth data of the target water area and the high-precision three-dimensional seabed topographic map, and output the generated topographic data as a high-precision three-dimensional seabed topographic map in a graphical or textual manner and display it on the terminal device for users to view and analyze.

[0048] Further, the output high-precision three-dimensional seabed topographic map supports multiple formats for export and interactive operation. In addition, the output three-dimensional seabed topographic map supports virtual reality (VR) and augmented reality (AR) display, and users can immerse themselves in viewing and analyzing by wearing corresponding equipment.

[0049] Figure 2 The laser point cloud and aerial image water depth inversion flowchart of one specific embodiment of the present application is shown in Figure 2 , specifically comprising:

[0050] Step 1: Obtain LiDAR point cloud data and aerial remote sensing images using a UAV-mounted integrated system;

[0051] A small-sized water-land integrated measurement device is mounted on a UAV, which contains a visible light camera and an airborne laser radar. The device is used to detect and collect data of the coastal zone of a specified area, and obtain remote sensing images and laser bathymetry data.

[0052] Step 2: Preprocess the remote sensing images;

[0053] Radiometric correction is performed on the image data to eliminate the effects of sensor characteristics and atmospheric conditions on the images. Geometric correction is performed to align the image data with the known coordinate system, ensuring spatial consistency of the data.

[0054] Step 3: Preprocess the airborne laser bathymetry data;

[0055] After data preprocessing, waveform data processing, error correction, and point cloud data processing of the collected data, the water depth data along the measurement route is finally obtained.

[0056] Step 4: Construct a water depth inversion model;

[0057] The high-precision water depth points measured by the laser radar are used as control points, and the gray values of the aerial photogrammetry images are combined to construct a water depth inversion model.

[0058] A logarithmic linear transformation is used to explain the exponential attenuation of visible light in water using the difference between the radiance brightness of shallow water area pixels and the radiance brightness of deep water area pixels in the same image, using only a single band. The linear model is then further extended to dual-band and multi-band. The specific algorithm formula is as follows:

[0059]

[0060] where a0 and a i are constant coefficients, N is the number of bands participating in inversion, R i is the gray value of the band, (R ∞ ) i is the gray value of the optical deep water area pixel adjacent to the shallow water area. Usually, X i = ln(R i -(R ∞ ) i ), combined with the panchromatic band image of aerial photogrammetry, the water depth inversion model can be constructed as:

[0061] z = a0 + a1 ln(R 全色 -(R ∞ ) 全色 )

[0062] The high-precision water depth extracted from the laser radar echo waveform or point cloud is input into the water depth inversion model, and the model parameters a0 and a1 are calculated, that is, the water depth inversion model is constructed.

[0063] Step 5: applying the model to the water depth inversion of the remote sensing image;

[0064] The laser data and image data are used to perform water depth inversion based on the model, and high-precision seabed topographic data is generated by using three-dimensional reconstruction and other technologies.

[0065] Step 6: outputting the water depth data of the target water area and the high-precision three-dimensional seabed topographic map;

[0066] The generated topographic data is output in a graphical or textual manner to a high-precision three-dimensional seabed topographic map displayed on a terminal device for user viewing and analysis.

[0067] Further reference Figure 3 , as an implementation of the method shown in the above figures, the present application provides an embodiment of a system, which corresponds to the method embodiment shown in Figure 1 , and the system can be applied in various electronic devices.

[0068] In a second aspect, the present application also discloses a coastal zone measurement system based on fusion of aerial images and laser point cloud data, which comprises a data acquisition module 31, an image preprocessing module 32, a laser point cloud preprocessing module 33, a water depth inversion construction module 34, a model application module 35 and a display result module 36.

[0069] In one specific embodiment, the data acquisition module 31 is configured to use a UAV-mounted integrated system to detect and collect the coastal zone of a specified area to obtain LiDAR point cloud data and aerial remote sensing images; the image preprocessing module 32 is configured to preprocess the obtained remote sensing images, including radiation correction and geometric correction, the radiation correction eliminates the influence of sensor characteristics and atmospheric conditions on the image by using a correction algorithm based on a physical model, and the geometric correction uses high-precision ground control points and image matching technology to accurately align the image data with the known coordinate system, ensuring the spatial consistency of the data; the laser point cloud preprocessing module 33 is configured to preprocess the obtained LiDAR point cloud data, including data preprocessing, waveform data processing, error correction and point cloud data processing, to obtain high-precision water depth data on the measurement route;

[0070] The water depth inversion construction module 34 is configured to use the measured high-precision water depth points as control points, combine the gray values of the aerial photogrammetry images, and construct a water depth inversion model; the model application module 35 is configured to use the processed laser data and image data, perform water depth inversion based on the constructed water depth inversion model, optimize the model by introducing a machine learning algorithm, generate high-precision seabed topographic data by using a three-dimensional reconstruction technology; and the display result module 36 is configured to output the water depth data and high-precision three-dimensional seabed topographic map of the target water area, output the generated topographic data as a high-precision three-dimensional seabed topographic map in a graphical or textual manner, and display the high-precision three-dimensional seabed topographic map on a terminal device for a user to view and analyze.

[0071] The functions of the above modules correspond to the methods, and thus will not be described here.

[0072] Compared with the existing manufacturing technology, the present application can improve the accuracy and efficiency of the coastal belt detection, obtain high-precision seabed topographic data, solve the problem of multi-system fusion and miniaturization, realize rapid water depth measurement in shallow water areas such as islands and reefs, and improve the depth measurement accuracy.

[0073] The above description is only the preferred embodiments of the present application and the explanation of the applied technical principles. Those skilled in the art should understand that the scope of the present application is not limited to the technical solutions formed by the specific combinations of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or equivalent features without departing from the above inventive concept. For example, the above features can be replaced with the technical features disclosed in the present application (but not limited to) having similar functions to form technical solutions.

Claims

1. A method for measuring a coastal zone based on fusion of aerial images and laser point cloud data, characterized in that, The method comprises the following steps: S1, using an unmanned aerial vehicle integrated system to detect and collect the coastal zone of the specified area to obtain LiDAR point cloud data and aerial remote sensing images; S2, preprocessing the obtained remote sensing images, including radiation correction and geometric correction, the radiation correction eliminates the influence of sensor characteristics and atmospheric conditions on the image by using a correction algorithm based on a physical model, and the geometric correction uses high-precision ground control points and image matching technology to accurately align the image data with the known coordinate system, ensuring the spatial consistency of the data; S3, preprocessing the obtained LiDAR point cloud data, including data preprocessing, waveform data processing, error correction and point cloud data processing, to obtain high-precision water depth data on the measurement route; further, the measured high-precision water depth points are used as control points, combined with the gray value of the aerial photogrammetry image, to construct a water depth inversion model; S4, using the processed laser data and image data, based on the constructed water depth inversion model, to perform water depth inversion, introducing a machine learning algorithm to optimize the model, and using three-dimensional reconstruction technology to generate high-precision seabed topographic data; and S5, further outputting the water depth data and high-precision three-dimensional seabed topographic map of the target water area, and outputting the generated topographic data as a high-precision three-dimensional seabed topographic map in a graphical or textual manner and displaying it on a terminal device for users to view and analyze. 2.The method according to claim 1, wherein, In step S3, it also includes: An algorithm is used that uses the difference between the radiance brightness of the shallow water area pixels and the radiance brightness of the deep water area pixels in the same image as a logarithmic linear transformation, which only uses a single band to explain the exponential decay of visible light in water, and then further extends the linear model to double-band and multi-band, and the algorithm formula is: where a0 and a i are constant coefficients, N is the number of bands participating in the inversion, R i is the gray value of the band, (R ∞ ) i is the gray value of the optical deep-water pixel adjacent to the shallow-water region. 3.The method according to claim 2, wherein, Record X i = ln(R i -(R ∞ ) i ), combined with the panchromatic band image of aerial photogrammetry, the water depth inversion model is constructed as: z = a0+ a1ln(R 全色 - (R ∞ ) 全色 ) The high-precision water depth extracted from the laser radar echo waveform or point cloud is input into the water depth inversion model, and the parameters a0 and a1 of the model are calculated, that is, the construction of the water depth inversion model is completed. 4.The method according to claim 1, wherein, In step S1, the integrated system includes a visible light camera and an airborne laser radar, and the integrated system can synchronously acquire remote sensing images and laser sounding data of the coastal zone of the specified area, and the unmanned aerial vehicle has the function of adjustable flight parameters to adapt to different coastal zone measurement environments. 5.The method according to claim 1, wherein, In step S3, the data preprocessing removes outliers and noise through a pre-set data cleaning algorithm; the waveform data processing uses pre-set waveform decomposition and feature extraction techniques to improve the accuracy and depth of waveform recognition; The error correction is based on real-time environmental monitoring data and sensor error models for accurate correction; The point cloud data processing uses optimized point cloud filtering and classification algorithms to automatically adjust the filtering parameters according to the topographic features of the coastal zone to obtain high-precision water depth data on the measurement route, and the precision after the point cloud data processing is in the centimeter level. 6.The method according to claim 1, wherein, In step S4, the machine learning algorithm includes neural network algorithm, linear regression algorithm and logistic regression algorithm. 7.The method according to claim 1, wherein, In step S5, it also includes: the output high-precision three-dimensional seabed topographic map supports multiple formats export and interactive operation.

8. A coastal zone measurement system based on fusion of aerial imagery and laser point cloud data, characterized in that, It includes: The data acquisition module is configured to detect and collect the coastal zone of a designated area by using the UAV-carrying integrated system to obtain LiDAR point cloud data and aerial remote sensing images; The image preprocessing module is configured to preprocess the obtained remote sensing images, including radiation correction and geometric correction. The radiation correction eliminates the influence of sensor characteristics and atmospheric conditions on the images by using a correction algorithm based on a physical model. The geometric correction accurately aligns the image data with a known coordinate system by using high-precision ground control points and image matching technology, ensuring the spatial consistency of the data. The laser point cloud preprocessing module is configured to preprocess the obtained LiDAR point cloud data, including data preprocessing, waveform data processing, error correction, and point cloud data processing, to obtain high-precision water depth data along the measurement route. The water depth inversion construction module is configured to use the measured high-precision water depth points as control points and combine the gray values of the aerial photogrammetry images to construct a water depth inversion model. The model application module is configured to use the processed laser data and image data to perform water depth inversion based on the constructed water depth inversion model, optimize the model by introducing a machine learning algorithm, and generate high-precision seabed terrain data using three-dimensional reconstruction technology. The display result module is configured to output the water depth data and high-precision three-dimensional seabed terrain map of the target water area. The generated terrain data is output as a high-precision three-dimensional seabed terrain map in a graphical or textual manner and displayed on a terminal device for users to view and analyze.

Images