Airborne LiDAR Seabed Echo Identification via Neural Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing airborne LiDAR bathymetry technologies face challenges in identifying seabed weak echoes due to low signal-to-noise ratios, especially in deeper water areas where refraction and scattering effects are more pronounced.
Innovation Solution
The method involves dividing sea area laser pulses into surface-seabed pulses or seabed-undefined pulses, using a two-stage local maximum method to determine potential seabed echo positions, extracting spatial features of a pre-generated point cloud using an ellipsoid neighborhood, and constructing a BP neural network to identify seabed points, optimizing the results based on laser pulse echo sequences and point cloud density.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional CFD method is used to identify target echoes, then reliability is improved, but seabed weak echoes are omitted reducing measurement precision
Solution Approach 1:
The patent segments the echo identification process into multiple stages: initial echo detection using CFD, followed by waveform stacking of neighboring echoes, and final identification through comparative analysis. This segmentation allows the system to maintain the reliability of CFD while adding precision through multi-stage processing of weak seabed echoes.
Solution Approach 2:
The patent merges multiple neighboring waveforms through stacking to enhance weak seabed echoes. By combining adjacent waveforms that contain similar seabed echo patterns, the signal-to-noise ratio is improved, allowing reliable detection of weak echoes that would be missed by conventional single-waveform analysis.
2Adaptability or versatility
If waveform processing methods such as peak detection and mathematical fitting are used, then applicability to waveforms with obvious double peaks is improved, but effectiveness for seabed weak echoes with low signal-to-noise ratio deteriorates
Solution Approach 1:
The patent applies preliminary waveform stacking before peak detection and fitting operations. By pre-processing the waveforms through stacking of neighboring echoes, the signal-to-noise ratio is enhanced in advance, making subsequent peak detection and mathematical fitting methods effective for both obvious double-peak waveforms and weak seabed echoes.
Solution Approach 2:
The patent introduces waveform stacking as an intermediary processing step between raw echo reception and final seabed position identification. This intermediary process enhances weak signals while preserving the characteristics needed for peak detection and fitting methods, bridging the gap between applicability to different waveform types and reliability for weak echo detection.
3Area of stationary object
If laser pulses are emitted in deeper water areas, then bathymetry coverage is improved, but signal intensity deteriorates due to refraction and scattering effects
Solution Approach 1:
The patent combines multiple neighboring waveforms through stacking to compensate for signal intensity loss in deeper water. By merging adjacent waveforms that all contain information about the same seabed location, the accumulated signal strength overcomes the attenuation caused by refraction and scattering in deeper water areas, maintaining bathymetry coverage effectiveness.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach significantly improves the identification of seabed weak echoes, increasing the number of identified points by 195.9% and the maximum available bathymetry depth by 60%, compared to conventional methods, thereby enhancing the reliability and effectiveness of airborne LiDAR bathymetry in deeper water areas.
Implementation Method 1
The propagation time of the laser pulse in deeper water areas (relative to the ALB) is longer, and it is more obviously influenced by a refraction effect and a scattering effect.
Implementation Method 2
The propagation time of the laser pulse in deeper water areas (relative to the ALB) is longer, and it is more obviously influenced by a refraction effect and a scattering effect.
Implementation Method 3
the receiver receives the backscattered echo waveforms of targets
Data Source
AI summary
An identification method of seabed weak echoes in airborne LiDAR bathymetry and a system thereof are provided, which belong to the technical field of marine surveying and mapping. The method includes: dividing sea area laser pulses into surface-seabed pulses or seabed-undefined pulses based on preprocessed data; using two-stage local maximum method to determine potential seabed echo positions of ALB waveforms and pre-generate seabed points; defining an ellipsoid neighborhood related to density and extracting spatial features of a pre-generated point cloud; constructing a BP neural network to identify seabed points in the pre-generated point cloud, and optimizing the identification result of the seabed points based on a laser pulse echo sequence and point cloud density.