Resilient UAV Autonomy and Remote Sensing
A special issue of Drones (ISSN 2504-446X).
Deadline for manuscript submissions: 31 May 2024 | Viewed by 17096
Special Issue Editors
Interests: image/LiDAR point clouds processing; sensor fusion; SLAM; unmanned systems; remote sensing methods for the power industry
2. Department of Land Surveying and Geo-Informatics, The Hong Kong Polytechnic University, Hong Kong 999077, China
Interests: image retrieval; image matching; structure from motion; multi-view stereo; deep learning
Special Issues, Collections and Topics in MDPI journals
Interests: laser scanning; point cloud segmentation; object recognition; semantic segmentation; image classification; instance segmentation
Interests: multi-view stereo; LiDAR data processing; deep learning; computer vision
Interests: UAV; mobile mapping; laser scanning; point cloud; inertial navigation
Special Issue Information
Dear Colleagues,
With the development of aerial imaging, oblique photogrammetry, laser scanning techniques and unmanned aircraft systems (UAVs), accurate and efficient perception, the reconstruction and recognition of large-scale 3D scenes have become popular topics in the fields of photogrammetry and computer vision. However, there are still several problems that need to be urgently solved, such as a low processing efficiency, difficulty to render the details of objects, and poor robustness of dense 3D reconstructions for poor-textured and occluded areas. Motivated by this rapid development, we are excited to invite you to submit a research paper to this Special Issue of Drones titled” UAV Image and LiDAR Processing”. The UAV data, including primarily UAV image and LiDAR data, has been widely used in aerial surveillance, 3D reconstruction and visualization, autonomous driving, and smart cities. This Special Issue aims to promote the further application of the UAV data, specifically in the fields of instance segmentation, object detection/tracking, SLAM, SFM, MVS, 3D mesh surface reconstruction, etc. Original submissions aligned with the above-mentioned research areas are highly welcomed.
Papers are welcomed from all fields directly related to these topics, including but not limited to the following:
- Trajectory planning for UAV data acquisition;
- The fusion of UAV sensor data (image/point clouds/GNSS/IMU);
- The registration of UAV image/point clouds;
- Real-time AI in motion planning and control, data gathering and analysis of UAVs;
- Image/LiDAR feature extraction, matching and bundle adjustment between UAV and UGV;
- Semantic/instance segmentation, classification, object detection and tracking with UAV data using the deep learning method;
- 3D reconstructions from UAV image/point clouds;
- SfM and SLAM using UAVs image/LiDAR data;
- Cooperative perception and mapping utilizing multiple UAVs and UGVs;
- Mobile edge computing (MEC) in UAVs;
- UAV image/point clouds processing in inspection, surveillance, GNSS-denied environment (underground/indoor spaces), etc.;
- UAV image/point clouds processing in power/oil/ industry, hydraulics, agriculture, ecology, emergency response and smart cities;
Dr. Chi Chen
Dr. San Jiang
Dr. Xijiang Chen
Dr. Mao Tian
Dr. Jianping Li
Dr. Jian Zhou
Guest Editors
Manuscript Submission Information
Manuscripts should be submitted online at www.mdpi.com by registering and logging in to this website. Once you are registered, click here to go to the submission form. Manuscripts can be submitted until the deadline. All submissions that pass pre-check are peer-reviewed. Accepted papers will be published continuously in the journal (as soon as accepted) and will be listed together on the special issue website. Research articles, review articles as well as short communications are invited. For planned papers, a title and short abstract (about 100 words) can be sent to the Editorial Office for announcement on this website.
Submitted manuscripts should not have been published previously, nor be under consideration for publication elsewhere (except conference proceedings papers). All manuscripts are thoroughly refereed through a single-blind peer-review process. A guide for authors and other relevant information for submission of manuscripts is available on the Instructions for Authors page. Drones is an international peer-reviewed open access monthly journal published by MDPI.
Please visit the Instructions for Authors page before submitting a manuscript. The Article Processing Charge (APC) for publication in this open access journal is 2600 CHF (Swiss Francs). Submitted papers should be well formatted and use good English. Authors may use MDPI's English editing service prior to publication or during author revisions.
Keywords
- drones
- UAV
- UAV swarm
- computer vision
- photogrammetry
- remote sensing
- LiDAR
- aerial imagery
- image and point cloud fusion
- detection and tracking
- segmentation
- SLAM
- path planning
- 3D reconstruction
- 3D visualization
Planned Papers
The below list represents only planned manuscripts. Some of these manuscripts have not been received by the Editorial Office yet. Papers submitted to MDPI journals are subject to peer-review.
Title: Semantic segmentation of indoor UAV point cloud
Authors: Xijiang Chen; Peng Li; Hui Deng
Affiliation: School of Safety Science and Emergency Management, Wuhan University of Technology, Wuhan, China
Abstract: The extraction of objects in an indoor environment is the key for many applications including object identification for UAV indoor navigation, facility management and reconstruction of construction. In view of this, this paper uses deep learning to conduct the point semantic segmentation. First, we apply exponential function to construct density clustering model according to the local density within cut-off distance. Second, the local density model of different objects is constructed and the constraint distance is determined according to the size of local density. Simultaneously, we find the cluster centres are recognized as points for which the product value of the local density and the constraint distance is anomalously large. Third, the cluster belonging of each point to the cluster centre is obtained according to the distance between point and cluster centre. Finally, we use the deep learning to conduct the fine point cloud semantic segmentation. Experiments show that the proposed method can conduct the point cloud segmentation and it is not affected by the point cloud distance resolution.
Title: Superpixel Hierarchical Stereo Matching Using Tree Dynamic Programming
Authors: Mao Tiana; Jiajin Fana; Qiaosheng Lia
Affiliation: College of Computer Science and Technology, Chongqing University of Posts and Telecommunications
Abstract: With the rapid development of sensor hardware and oblique photogrammetry technologies, stereo matching technology has been widely used for urban scene 3D modelling. Aiming at the low efficiency and poor robustness of disparity map reconstruction of traditional stereo matching algorithms. This approach converts the disparity map reconstruction problem into a slanted-plane-based continuous global energy optimization model, the PatchMatch and tree dynamic programming strategies are utilized to solve the super-pixel-based energy model optimization, and experimental data is used to verify the effectiveness and robustness of the proposed algorithm. The experimental results show that the proposed method can quickly and accurately reconstruct the geometric structure of 3D scene.
Title: Automatically extract the DBH and height of individual tree using TLS and UAV LiDAR points
Authors: Huang Xia; ZHU Ningning; LIU Rundong
Affiliation: State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, Hubei, China
Abstract: Forest biomass is an important biophysical parameter to describe the function and productivity of forest ecosystem. Biomass model is the main method to estimate forest biomass. The DBH and tree height are the two most important parameters in the construction of biomass model. In this paper, the method of automatic extraction of diameter at breast height (DBH) and tree height from TLS and LiDAR points is studied. First, based on the fact that individual tree is approximately perpendicular to the ground, the TLS and UAV LiDAR points are divided into different sizes with horizontal grids; Second, the number of points, the lowest, highest and height difference .et al. in each grid are calculated, and the combination of these features is used to quickly separate the individual tree from TLS and UAV LiDAR points; Third, the DBH were fitted by RANSAC algorithm with different thickness TLS LiDAR points, and the tree height were calculated by the grid height; Finally, 5 TLS and UAV LiDAR points are used to experiment the effectiveness of our proposed method.
Title: : Automatic registration of UAV-based photogrametry and lidar data using individual trees in forestry area
Authors: Xin Zhao; Rundong Liu; Ruibo Chen; Jianping Li; ChiChen
Affiliation: 1. State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, Hubei, China
2. Guangxi Zhuang Autonomous Region Remote Sensing Institute of Natural Resources, Nanning 530000, Guangxi, China
Abstract: With the development of digital forestry and precision forestry, the demand for digital visualization of forest regions is becoming more and more prominent. In the field of 3d modeling of forest area, UAV-based photogrammetry and lidar technology can provide reliable image data and lidar data respectively. Image can contribute rich spectral information and texture information, while lidar can directly provide three-dimensional coordinates and reflection intensity of ground objects. In forestry areas, ground object information is more scarce than in urban areas, and it is difficult to obtain the overall information by using a single data source. Registration of the image and lidar data can make up for the deficiency of single data and improve the property information of ground objects. Therefore, it is of great significance to integrate the two kinds of data. Due to the lack of artificial features in forest areas, features such as house corners and building facades in urban areas cannot be adopted, while individual trees in forest contain obvious geometric features. Therefore, this paper proposes to successfully realize the fusion of point cloud and image by extracting the features of individual trees as the basis of registration. For each tree, the linear structure is extracted from the trunk part and the spherical structure is extracted from the crown part. Meanwhile, height information is recorded and the plane distance between the tree and other trees within the set radius is measured to form a unique feature. The relative registration of image and image data is realized by matching the features of the image.
Title: : Automatic individual tree estimation using a low-cost helmet-based laser scanning system
Authors: Weitong Wu; Rundong Liu; Ruibo Chen; Jianping Li1
Affiliation: 1. State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, Hubei, China
2. Guangxi Zhuang Autonomous Region Remote Sensing Institute of Natural Resources, Nanning 530000, Guangxi, China
Abstract: Automated individual tree estimation using point clouds plays an increasingly significant role in efficiently, accurately, and completely monitoring forests. This paper presents an automatic tree detection with an estimation of two dendrometric variables: diameter at breast height (DBH) and total tree height (TH) using a low-cost helmet-based laser scanning system(HLS). Operative processes for data collection and automatic forest inventory are described in detail. The approach used is based on the clustering of points belonging to each individual tree, the isolation of the trunks, the iterative fitting of circles for the DBH calculation and the computation of the TH of each tree. TLS and HLS point clouds were compared by the statistical analysis of both estimated forest dendrometric parameters and the possible presence of bias. Results show that the apparent differences in point density and relative precision between both 3D forest models do not affect tree detection, and helmet-based laser scanning system can meet the needs of forest survey.