CN-121999393-A - Sea ice thickness extraction method and equipment based on unmanned aerial vehicle laser point cloud rapid classification

CN121999393ACN 121999393 ACN121999393 ACN 121999393ACN-121999393-A

Abstract

The invention provides sea ice thickness extraction method and equipment based on unmanned aerial vehicle laser point cloud rapid classification, which comprise the steps of obtaining laser point cloud data collected by an unmanned aerial vehicle platform, determining a local sea surface elevation reference based on the laser point cloud data, identifying a ship point cloud through relative height screening and spatial clustering based on the local sea surface elevation reference, performing geometric feature verification on a clustering result, removing sea water points and the ship point cloud in the laser point cloud data, performing spatial clustering on the rest point cloud to identify sea freezing point cloud, and extracting sea ice thickness based on the local sea surface elevation reference and the sea ice point cloud. The invention utilizes the advantages of high spatial resolution and high cost performance of the unmanned aerial vehicle, is matched with a refined point cloud processing technology, has rapid and accurate classification and high thickness extraction precision, is suitable for polar region complex observation environment, can provide reliable data support for polar region channel development, navigation safety investigation and climate change research in real time, and has remarkable practical value.

Inventors

XIAO FENG
ZHANG SHENGKAI
XIE ZHENG
SUN XIAOYU
LI JIAXING
GENG TONG
HU XICHENG
MA JIE

Assignees

武汉大学

Dates

Publication Date: 20260508
Application Date: 20260112

Claims (10)

1. A sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud rapid classification is characterized by comprising the following steps: acquiring laser point cloud data acquired by an unmanned aerial vehicle platform; determining a local sea surface elevation reference based on the laser point cloud data; Based on the local sea surface elevation reference, identifying a ship point cloud through relative height screening and spatial clustering, and performing geometric feature verification on a clustering result; After excluding sea water points and the ship point clouds from the laser point cloud data, performing spatial clustering on the rest point clouds to identify sea freezing point clouds; And extracting the sea ice thickness based on the local sea surface elevation reference and the sea ice point cloud.
2. The sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud rapid classification of claim 1, wherein when the local sea surface Gao Chengji is determined based on laser point cloud data, a part of points with the lowest elevation in the point cloud is selected, and the average elevation is calculated as the reference.
3. The sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud rapid classification of claim 1, further comprising the step of performing downsampling and denoising processing on the laser point cloud data before determining the local sea surface elevation reference.
4. The sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud rapid classification of claim 1, wherein the identification of the ship candidate point cloud through relative height screening refers to screening out points with heights exceeding a preset threshold value relative to the local sea surface elevation reference.
5. The sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud rapid classification of claim 1 is characterized in that the plane geometric feature verification of the clustering result comprises the steps of calculating the projection size of a candidate object on a horizontal plane and judging according to preset size conditions.
6. The method for extracting sea ice thickness based on unmanned aerial vehicle laser point cloud rapid classification of claim 1, wherein the residual point cloud is spatially clustered to identify sea ice point cloud, and the clustering distance threshold is smaller than that used when identifying ship point cloud.
7. The method for extracting sea ice thickness based on unmanned aerial vehicle laser point cloud rapid classification of claim 1, wherein the extracting sea ice thickness comprises calculating sea ice dry height according to sea ice point cloud, and converting the dry height into sea ice thickness based on hydrostatic balance principle.
8. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor implements a sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud fast classification as set forth in any one of claims 1 to 7 when executing the program.
9. A non-transitory computer readable storage medium having stored thereon a computer program, wherein the computer program when executed by a processor implements a sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud fast classification as set forth in any one of claims 1 to 7.
10. A computer program product comprising a computer program, characterized in that the computer program when executed by a processor implements a sea ice thickness extraction method based on unmanned aerial vehicle laser point cloud fast classification as claimed in any one of claims 1 to 7.

Description

Sea ice thickness extraction method and equipment based on unmanned aerial vehicle laser point cloud rapid classification Technical Field The invention belongs to the technical field of polar remote sensing monitoring and geodetic measurement, and particularly relates to a sea ice thickness extraction technical scheme for realizing quick point cloud classification of sea ice, sea water and icebreaker by utilizing unmanned plane laser point cloud data. Background Sea ice is an important component of the earth system, influences the formation of ocean deep water through fresh water output, regulates heat transmission to deep ocean, further influences global ocean circulation and climate change, and influences the middle-latitude weather process through influencing regional energy balance and large-scale atmospheric circulation. The thickness of sea ice is the third dimension of sea ice change, which dominates the thermodynamic and kinetic characteristics of sea ice, influences the movement, deformation, freezing and ablation processes of sea ice, and directly determines the exchange process and rate of sea-gas energy and substances. In addition, the sea ice thickness is an important parameter affecting the passage of the polar channel, and the accurate grasp of the sea ice thickness condition in the channel is important for the development of the polar channel. Sea ice thickness is also one of the most difficult sea ice parameters to invert. The commonly used sea ice thickness acquisition method mainly comprises ice surface drilling, navigation observation, upward sonar, aviation and satellite remote sensing and the like. The drilling observation precision is highest, but the drilling observation precision is limited by regional climate and environment, and only a very small amount of observation data can be acquired. The sailing observation is simple and easy to observe the sea ice thickness along the line by using a photogrammetry method in the sailing of the icebreaker, but the spatial distribution of the sea ice thickness is generally underestimated due to the thin ice tendency of the design of the icebreaker route. The bottom view sonar observes the draft of sea ice from the ice, and then reverses sea ice thickness, and submarine bottom view sonar is difficult to obtain the continuous ice thickness characteristic in specific region, moors bottom view sonar and then is difficult according to data transmission. Aviation and satellite remote sensing are main means for obtaining a large-scale sea ice thickness by carrying an altimeter or a microwave radiometer for ice thickness observation, but the time-space resolution is limited to a certain extent. In particular, satellite remote sensing has wide coverage range, but the spatial resolution is usually kilometer level, so that the fine sea ice structure of key areas such as a channel, an operation point and the like is difficult to capture, and meanwhile, satellite data acquisition has delay, so that the real-time or near-real-time navigation safety and operation decision requirement is difficult to meet. The aeronautical remote sensing (such as the laser radar carried by an aircraft) can acquire the observation data with higher resolution, but has the problems of high operation cost, long operation period, large constraint by airspace and climate conditions and the like, and is difficult to realize the intensive monitoring of normalization and maneuverability. The appearance and the rapid development of unmanned aerial vehicle remote sensing provide a new observation means for sea ice thickness observation. The unmanned aerial vehicle remote sensing method can obtain the sea ice thickness observation result with refined region, and has the characteristics of high spatial resolution, high frequency, high cost performance and the like. However, applying unmanned plane laser radar technology to actual sea ice thickness extraction still faces a series of technical challenges to be solved, namely firstly, the point cloud data in the polar environment contains complex noise such as sea wave droplets, crushed ice and the like, and secondly, in the typical application scene of manned platform operation such as icebreaker and the like, the collected point cloud data contains sea water, sea ice and the ship itself. The conventional general point cloud segmentation and classification algorithm (such as plane fitting based on random sampling consistency or semantic segmentation based on deep learning) has the limitation in processing such special mixed scenes that on one hand, sea ice surface relief and a ship deck structure are approximate in three-dimensional form and are easy to cause misjudgment only by means of geometric or reflectivity characteristics, and on the other hand, when a ship is used as a large-scale artificial object, the point cloud cannot be accurately identified and removed, calculation of sea ice surface elevation is seriously polluted, and furthe