机载激光雷达海洋测深信号精确探测技术研究

Ocean development in China starts to speed up, which requires better ability of land and underwater topography surveying for coast, reef and rocks. Airborne LiDAR bathymetry (ALB) is an efficient and high-precision method for underwater topography surveying, which has been used for shallow water monitering and underwater target detection. The research on ALB in China is not sufficient compared with USA, Canada, Sweden and Australia, especially on detection data processing, which needs more attention. Signal detection techniques of ALB will be studied in this project for solving some problems including difficulties in classification of water and land waveforms, low detection rate and poor reliability of bathymetry echo signal detection, and insufficiency of waveform decomposition model. The project will mainly focus on some key technologies, such as classification of sea and land waveforms based on multi-channel waveform feature analysis, enhancement of water waveforms based on feature fusion, and high-precision blue-green channel waveform decomposition based on the echo signal model. Some breakthroughs will be achieved on multi-channel waveform classification model, de-noising and enhancement of water waveforms and high-precision waveform decomposition. The aim of the project is the high-precision detection of the bathymetry signal and the reliable and accurate reconstruction of the underwater terrain. The achievements of the project will be beneficial to the development and application of ALB equipment and technology in China.

我国海洋战略的快速推进对海岸带、岛礁等区域的水上水下地形测量能力提出了更高的要求。机载激光海洋测深技术是一种高效的水陆一体化地形高精度测量手段，广泛适用于浅海、海岸带区域的海底地形测绘、浅层水域监测、水下目标探测等领域。我国在该技术领域上，特别是数据处理方面，与国外还存在明显的差距，迫切需要加强相关理论和方法的研究。本项目拟对机载激光海洋测深信号探测技术开展研究，针对目前存在的海陆波形混淆、测深信号检测正确率低、可靠性差、波形分解的模型不适应等问题，重点对基于多通道波形特征联合分析的海陆回波分类、基于特征融合的水域回波增强以及基于回波模型的蓝绿通道水域回波波形分解等关键技术进行研究，力求在多通道波形分类模型、水域回波去噪增强方法、高精度波形分解算法等方面实现突破，实现测深信号的高精度检测和水下地形的可靠、精确重建，为我国自主机载激光测深设备的研制以及在我国的应用开展核心理论研究。

机载激光海洋测深技术是一种高效的水陆一体化地形高精度测量手段，广泛适用于浅海、海岸带区域的海底地形测绘、浅层水域监测、水下目标探测等领域。我国在该技术领域上，特别是数据处理方面，与国外还存在明显的差距，迫切需要加强相关理论和方法的研究。本项目对机载激光海洋测深信号探测技术开展了研究，针对目前存在的海陆波形混淆、测深信号检测正确率低、可靠性差、波形分解的模型不适应等问题，构建了单频机载激光测深海陆波形自动分类模型、近水面渗透误差修正的半经验波形分解模型、基于二流法的漫衰减系数半分析反演模型，提出了基于高斯过程回归改进的GPR-QAA反演算法、基于先验波形水深分类的测深信号自适应检测方法以及基于卷积神经网络的测深信号复原方法，并在上海、海南、西沙群岛、山东、福建等地开展了有人机和无人机激光测深飞行试验，完成了算法模型的实验与精度分析，实现了水底和水面回波信号的高精度探测，以及水下地形的精确重建和水下目标的准确探测，为机载激光海洋测深技术在我国海岸带、岛礁地形测绘中的深入应用消除了技术瓶颈。