Abstract:With the rapid development of drone technology, the application of drones equipped with visual sensors for building inspection and surveillance has attracted much attention. View planning aims to find a set of near-optimal viewpoints for vision-related tasks to achieve the vision coverage goal. This paper proposes a new clustering-based two-step computational method using spectral clustering, local potential field method, and hyper-heuristic algorithm to find near-optimal views to cover the target building surface. In the first step, the proposed method generates candidate viewpoints based on spectral clustering and corrects the positions of candidate viewpoints based on our newly proposed local potential field method. In the second step, the optimization problem is converted into a Set Covering Problem (SCP), and the optimal viewpoint subset is solved using our proposed hyper-heuristic algorithm. Experimental results show that the proposed method is able to obtain better solutions with fewer viewpoints and higher coverage.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to plan a set of approximately optimal viewpoints through the visual sensors carried by drones for building surface detection and monitoring tasks, so as to achieve efficient target coverage. Specifically, the research aims to use a model - based method to improve the coverage rate of the target building surface by generating a small number of but efficient viewpoints, while reducing the number of required viewpoints. ### Problem Background With the rapid development of drone technology, the application of using drones equipped with visual sensors for building detection and monitoring has received extensive attention. The goal of view planning is to find a set of viewpoints (including position and direction) from the visual sensors carried by the robot to cover the required surface area of the target object. Common optimization goals include increasing the surface coverage rate of the target area and reducing the number of viewpoints. ### Research Methods To achieve this goal, the paper proposes a new two - step calculation method based on clustering. The specific steps are as follows: 1. **Generate Candidate Viewpoints**: - Use the spectral clustering algorithm to cluster the triangular mesh model of the target object according to the weighted distance and normal vector. - Generate high - quality candidate viewpoints based on the clustering results, and use the newly proposed local potential field method to correct the positions of the candidate viewpoints. 2. **Combinatorial Optimization**: - Transform the problem into a set covering problem (SCP), and use the newly proposed hyper - heuristic algorithm to search for the optimal subset from the candidate viewpoints. ### Main Contributions - A new method for generating candidate viewpoints based on spectral clustering is proposed. This method is robust to the clustering of non - convex shapes, can better consider complex geometric features, and generate high - quality candidate viewpoints. - A new local potential field method is proposed to correct the positions of candidate viewpoints, which solves the problem of unreasonable viewpoint positions generated by the clustering algorithm. - A new hyper - heuristic algorithm is proposed to solve the combinatorial optimization problem, which is more effective and robust than traditional heuristic methods. ### Experimental Results The experimental results show that this method can obtain a higher coverage rate with fewer viewpoints, verifying its effectiveness in practical applications. Compared with the random sampling method, this method reduces the number of viewpoints by about 20.65% and does not significantly increase the computational cost. Through these improvements, this method performs well in drone building detection and monitoring tasks and has high practical value.

A New Clustering-based View Planning Method for Building Inspection with Drone

Offsite aerial path planning for efficient urban scene reconstruction

UAV Trajectory Planning via Viewpoint Resampling for Autonomous Remote Inspection of Industrial Facilities

Three-dimensional coverage path planning via viewpoint resampling and tour optimization for aerial robots

A co-optimal coverage path planning method for aerial scanning of complex structures

UAS-based Automated Structural Inspection Path Planning via Visual Data Analytics and Optimization

A Clustering-Based Coverage Path Planning Method for Autonomous Heterogeneous UAVs

Multi-UAV Coverage Path Planning for the Inspection of Large and Complex Structures

A hierarchical multi-UAV cooperative framework for infrastructure inspection and reconstruction

Near-Optimal 3-D Visual Coverage for Quadrotor Unmanned Aerial Vehicles Under Photogrammetric Constraints

Clustering-Based Multi-Region Coverage-Path Planning of Heterogeneous UAVs

Multi-Drone Collaborative Trajectory Optimization for Large-Scale Aerial 3D Scanning

UAV Cluster Mission Planning Strategy for Area Coverage Tasks

Coverage Path Planning and Point-of-Interest Detection Using Autonomous Drone Swarms

A Route Planning Method for UAV Swarm Inspection of Roads Fusing Distributed Droneport Site Selection

ECHO: An Efficient Heuristic Viewpoint Determination Method on Frontier-based Autonomous Exploration for Quadrotors

Aerial path planning for online real-time exploration and offline high-quality reconstruction of large-scale urban scenes

Offshore Oilfield Inspection Planning With Drone Routing Optimization

Optimal Trajectory Planning of Drones for 3D Mobile Sensing

Density-Based Optimal UAV Path Planning for Photovoltaic Farm Inspection in Complex Topography

Aircraft Inspection Using Unmanned Aerial Vehicles