Deep Learning for LiDAR Point Cloud Classification in Remote Sensing

Abstract

Point clouds are one of the most widely used data formats produced by depth sensors. There is a lot of research into feature extraction from unordered and irregular point cloud data. Deep learning in computer vision achieves great performance for data classification and segmentation of 3D data points as point clouds. Various research has been conducted on point clouds and remote sensing tasks using deep learning (DL) methods. However, there is a research gap in providing a road map of existing work, including limitations and challenges. This paper focuses on introducing the state-of-the-art DL models, categorized by the structure of the data they consume. The models’ performance is collected, and results are provided for benchmarking on the most used datasets. Additionally, we summarize the current benchmark 3D datasets publicly available for DL training and testing. In our comparative study, we can conclude that convolutional neural networks (CNNs) achieve the best performance in various remote-sensing applications while being light-weighted models, namely Dynamic Graph CNN (DGCNN) and ConvPoint.

Utilizing a Single-Stage 2D Detector in 3D LiDAR Point Cloud With Vertical Cylindrical Coordinate Projection for Human Identification

“…The issue of accuracy and data quality in object detection using 3D LiDAR must be acknowledged and resolved [33]. Additionally, it is important to recognize that the efficiency could be diminished as a result of limited processing capabilities and storage capacity [34]. A commonly used approach to overcome these challenges is using bird's eye view (BEV) geometry projection, which provides a horizontal perspective from an elevated position.…”

The Use of Green Laser in LiDAR Bathymetry: State of the Art and Recent Advancements

“…Mismatch of the training series may result in under- or over-fitting of the network. In the works [ 129 , 130 ] summarizing the current applications of deep learning in laser scanning, further dynamic development of these techniques in the classification and detection of objects is predicted.…”

Artificial Neural Networks for IoT-Enabled Smart Applications: Recent Trends

“…In the age of neural networks and the Internet of Things (IoT), the search for new neural network architectures capable of operating on devices with limited computing power and small memory size is becoming an urgent agenda. Trends in the development of artificial intelligence (AI) applications in the field of the Internet of Things include smart healthcare services [ 1 , 2 , 3 , 4 , 5 ], smart object-recognition [ 6 , 7 , 8 ], smart environment monitoring [ 9 , 10 ], and smart disaster rescue [ 11 ]. Traditionally, such applications operate in real time.…”

“…The realm of smart object-recognition applications of AI systems is presented in the subsequent articles [ 6 , 7 , 8 ]. Driver assistants have become a more and more popular class of smart IoT-enabled smart applications, as illustrated in study [ 6 ] that detects distracting actions in driver activities.…”

“…Research on feature extraction from disordered and irregular point cloud data has advanced recently. The overview [ 8 ] of the different types of models is presented, and studies of point clouds and remote sensing problems have been carried out using deep learning methods. It is concluded that convolutional neural networks achieve the best performance in various remote sensing applications that operate directly with raw cloud data.…”