Classification of sea and land waveforms based on deep learning for airborne laser bathymetry

Hu Shanjiang; He Yan; Tao Bangyi; Yu Jiayong; Chen Weibiao

doi:10.3788/IRLA201948.1113004

Classification of sea and land returns in airborne lidar was essential for the research of coastal zones and their changing nature. A method for classification using deep learning on the original airborne lidar echo was proposed. A fully connected neural network, and a one-dimensional convolutional neural network (CNN), were used on a training dataset and test datasets from in-situ measurements, and a classification accuracy of 99.6% was obtained. The model was utilized on the datasets from different areas, a classification accuracy of 95.6% was achieved and the processing speed was increased by about 52% compared to support vector machine (SVM) method. The results denote that the deep learning method is very effective for classification of airborne lidar echo waveforms with high precision and speed. It may present further use as a candidate method for classifying species on the sea floor with airborne laser bathymetry.

Classification of sea and land waveforms based on deep learning for airborne laser bathymetry

1. Key Laboratory of Space Laser Communication and Detection Technology,Shanghai Institute of Fine Mechanics and Optics,Chinese Academy of Sciences,Shanghai 201800,China;

2. University of Chinese Academy of Sciences,Beijing 100049,China;

3. Second Institute of Oceanography,MNR,Hangzhou 310012,China;

4. Shandong University of Science and Technology,Qingdao 266590,China

Received Date: 2019-03-13

Rev Recd Date: 2019-05-10

Publish Date: 2019-11-25

Key words:

Abstract: Classification of sea and land returns in airborne lidar was essential for the research of coastal zones and their changing nature. A method for classification using deep learning on the original airborne lidar echo was proposed. A fully connected neural network, and a one-dimensional convolutional neural network (CNN), were used on a training dataset and test datasets from in-situ measurements, and a classification accuracy of 99.6% was obtained. The model was utilized on the datasets from different areas, a classification accuracy of 95.6% was achieved and the processing speed was increased by about 52% compared to support vector machine (SVM) method. The results denote that the deep learning method is very effective for classification of airborne lidar echo waveforms with high precision and speed. It may present further use as a candidate method for classifying species on the sea floor with airborne laser bathymetry.

HTML

Reference (18)

[1]	Baltsavias E P. Airborne laser scanning:existing systems and firms and other resources[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 1999, 54(2-3):164-198.
[2]	Tuell G, Barbor K, Wozencraft J. Overview of the coastal zone mapping and imaging lidar (CZMIL):A new multisensor airborne mapping system for the US army corps of engineers[C]//Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XVI. International Society for Optics and Photonics, 2010, 7695:76950R.
[3]	Pe'eri S, Morgan L V, Philpot W D, et al. Land-water interface resolved from airborne LiDAR bathymetry (ALB) waveforms[J]. Journal of Coastal Research, 2011,62:75-85.
[4]	Collin A, Long B, Archambault P. Merging land-marine realms:Spatial patterns of seamless coastal habitats using a multispectral LiDAR[J]. Remote Sensing of Environment, 2012, 123:390-399.
[5]	Allouis T, Bailly J S, Pastol Y, et al. Comparison of LiDAR waveform processing methods for very shallow water bathymetry using Raman, near-infrared and green signals[J]. Earth Surface Processes and Landforms, 2010, 35(6):640-650.
[6]	Huang Tiancheng, Tao Bangyi, Mao Zhihua, et al.Classification of sea and land waveform based on multi-channel ocean lidar[J]. Chinese Journal of Lasers, 2017, 44(6):0610002. (in Chinese)
[7]	Ma Yue, Zhang Wenhao, Zhang Zhiyu, et al. Sea and sea-ice waveform classification for the laser altimeter based on semi-analytic model[J]. Infrared and Laser Engineering, 2018, 47(5):0506005. (in Chinese)
[8]	Nahhas F H, Shafri H Z M, Sameen M I, et al. Deep learning approach for building detection using lidar-orthophoto fusion[J]. Journal of Sensors, 2018:7212307.
[9]	Hu X, Yuan Y. Deep-learning-based classification for DTM extraction from ALS point cloud[J]. Remote Sensing, 2016, 8(9):730.
[10]	Arief H, Strand G H, Tveite H, et al. Land cover segmentation of airborne LiDAR data using stochastic atrous network[J]. Remote Sensing, 2018, 10(6):973.
[11]	Maturana D, Scherer S. 3d convolutional neural networks for landing zone detection from lidar[C]//2015 IEEE International Conference on Robotics and Automation (ICRA), 2015:3471-3478.
[12]	Velas M, Spanel M, Hradis M, et al. Cnn for very fast ground segmentation in velodyne lidar data[C]//2018 IEEE International Conference on Autonomous Robot Systems and Competitions (ICARSC), 2018:97-103.
[13]	Matti D, Ekenel H K, Thiran J P. Combining lidar space clustering and convolutional neural networks for pedestrian detection[C]//2017 14th IEEE International Conference on Advanced Video and Signal Based Surveillance (AVSS), 2017:1-6.
[14]	Dewan A, Oliveira G L, Burgard W. Deep semantic classification for 3d lidar data[C]//2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2017:3544-3549.
[15]	Wang A, He X, Ghamisi P, et al. Lidar data classification using morphological profiles and convolutional neural networks[J]. IEEE Geoscience and Remote Sensing Letters, 2018, 15(5):774-778.
[16]	LeCun Y, Bottou L, Bengio Y, et al. Gradient-based learning applied to document recognition[J]. Proceedings of the IEEE, 1998, 86(11):2278-2324.
[17]	Dai W, Dai C, Qu S, et al. Very deep convolutional neural networks for raw waveforms[C]//2017 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2017:421-425.
[18]	Zhao Ming, Chen Shi, Yuen Dave. Waveform classification and seismic recognition by convolution neural network[J]. Chinese Journal of Geophysics, 2019, 62(1):374-382. (in Chinese)

Classification of sea and land waveforms based on deep learning for airborne laser bathymetry

doi: 10.3788/IRLA201948.1113004

Abstract

References

Proportional views

通讯作者: 陈斌, bchen63@163.com

Article Metrics

Related

Proportional views