Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CPP-DIP: Multi-objective Coverage Path Planning for MAVs in Dispersed and Irregular Plantations

Weijie Kuang, Hann Woei Ho, Ye Zhou

TL;DR

CPP-DIP addresses efficient coverage path planning for MAVs in dispersed and irregular plantations by turning CPP into a TSP and avoiding GPS-based environmental models. It combines HOG-based oil palm detection with ForaNav for real-time tree localization, density-aware KDE/DBSCAN waypoint generation, and three TSP solvers (GHI, ACO, MCRL) plus object-optimized replanning to balance travel distance, turning angles, and path intersections. The approach yields substantial improvements in path smoothness and intersection elimination, with MCRL achieving a favorable trade-off between distance and maneuverability, and real-world flight experiments validate accurate MAV positioning above trees. Overall, CPP-DIP offers a robust, GPS-free framework suitable for scalable, precision-agriculture tasks such as targeted spraying and monitoring, with potential for multi-agent extensions and onboard planning.

Abstract

Coverage Path Planning (CPP) is vital in precision agriculture to improve efficiency and resource utilization. In irregular and dispersed plantations, traditional grid-based CPP often causes redundant coverage over non-vegetated areas, leading to waste and pollution. To overcome these limitations, we propose CPP-DIP, a multi-objective CPP framework designed for Micro Air Vehicles (MAVs). The framework transforms the CPP task into a Traveling Salesman Problem (TSP) and optimizes flight paths by minimizing travel distance, turning angles, and intersection counts. Unlike conventional approaches, our method does not rely on GPS-based environmental modeling. Instead, it uses aerial imagery and a Histogram of Oriented Gradients (HOG)-based approach to detect trees and extract image coordinates. A density-aware waypoint strategy is applied: Kernel Density Estimation (KDE) is used to reduce redundant waypoints in dense regions, while a greedy algorithm ensures complete coverage in sparse areas. To verify the generality of the framework, we solve the resulting TSP using three different methods: Greedy Heuristic Insertion (GHI), Ant Colony Optimization (ACO), and Monte Carlo Reinforcement Learning (MCRL). Then an object-based optimization is applied to further refine the resulting path. Additionally, CPP-DIP integrates ForaNav, our insect-inspired navigation method, for accurate tree localization and tracking. The experimental results show that MCRL offers a balanced solution, reducing the travel distance by 16.9 % compared to ACO while maintaining a similar performance to GHI. It also improves path smoothness by reducing turning angles by 28.3 % and 59.9 % relative to ACO and GHI, respectively, and effectively eliminates intersections. These results confirm the robustness and effectiveness of CPP-DIP in different TSP solvers.

CPP-DIP: Multi-objective Coverage Path Planning for MAVs in Dispersed and Irregular Plantations

TL;DR

CPP-DIP addresses efficient coverage path planning for MAVs in dispersed and irregular plantations by turning CPP into a TSP and avoiding GPS-based environmental models. It combines HOG-based oil palm detection with ForaNav for real-time tree localization, density-aware KDE/DBSCAN waypoint generation, and three TSP solvers (GHI, ACO, MCRL) plus object-optimized replanning to balance travel distance, turning angles, and path intersections. The approach yields substantial improvements in path smoothness and intersection elimination, with MCRL achieving a favorable trade-off between distance and maneuverability, and real-world flight experiments validate accurate MAV positioning above trees. Overall, CPP-DIP offers a robust, GPS-free framework suitable for scalable, precision-agriculture tasks such as targeted spraying and monitoring, with potential for multi-agent extensions and onboard planning.

Abstract

Coverage Path Planning (CPP) is vital in precision agriculture to improve efficiency and resource utilization. In irregular and dispersed plantations, traditional grid-based CPP often causes redundant coverage over non-vegetated areas, leading to waste and pollution. To overcome these limitations, we propose CPP-DIP, a multi-objective CPP framework designed for Micro Air Vehicles (MAVs). The framework transforms the CPP task into a Traveling Salesman Problem (TSP) and optimizes flight paths by minimizing travel distance, turning angles, and intersection counts. Unlike conventional approaches, our method does not rely on GPS-based environmental modeling. Instead, it uses aerial imagery and a Histogram of Oriented Gradients (HOG)-based approach to detect trees and extract image coordinates. A density-aware waypoint strategy is applied: Kernel Density Estimation (KDE) is used to reduce redundant waypoints in dense regions, while a greedy algorithm ensures complete coverage in sparse areas. To verify the generality of the framework, we solve the resulting TSP using three different methods: Greedy Heuristic Insertion (GHI), Ant Colony Optimization (ACO), and Monte Carlo Reinforcement Learning (MCRL). Then an object-based optimization is applied to further refine the resulting path. Additionally, CPP-DIP integrates ForaNav, our insect-inspired navigation method, for accurate tree localization and tracking. The experimental results show that MCRL offers a balanced solution, reducing the travel distance by 16.9 % compared to ACO while maintaining a similar performance to GHI. It also improves path smoothness by reducing turning angles by 28.3 % and 59.9 % relative to ACO and GHI, respectively, and effectively eliminates intersections. These results confirm the robustness and effectiveness of CPP-DIP in different TSP solvers.
Paper Structure (15 sections, 12 equations, 9 figures, 2 tables, 6 algorithms)

This paper contains 15 sections, 12 equations, 9 figures, 2 tables, 6 algorithms.

Table of Contents

  1. Introduction
  2. Related works
  3. Methodology
  4. Oil palm tree detection
  5. Waypoint and dense region generation
  6. Target-based path planning
  7. Greedy Heuristic Insertion
  8. Ant Colony Optimization
  9. Monte Carlo Reinforcement Learning
  10. Object-optimized path replanning
  11. Experiments and discussion
  12. Performance evaluation of optimization steps
  13. Comparison of Algorithm Performance
  14. Flight experiments
  15. Conclusions

Figures (9)

  • Figure 1: The framework of CPP-DIP. Tree image coordinates are derived from detection results obtained through a HOG-based method. Next, high-density regions are segmented using a KDE approach, while cover circles are produced in low-density areas using a greedy algorithm to generate waypoint coordinates. A path is then planned through these waypoints by solving the TSP. Finally, this path is further refined through object-based optimization.
  • Figure 2: Coverage-based waypoint generation. (a) Search in potential region; (b) First waypoint generation; (c) Sequential waypoint generation; (d) Path plan based on waypoints.
  • Figure 3: Dense region generation. (a) Tree detection results; (b) Kernel Density Estimation; (c) Dense region generation; (d) Boustrophedon path plan.
  • Figure 4: Object-optimized path replanning. (a) Initial waypoints and planned path; (b) Feasible radius calculate; (c) Feasible region generate; (d) Path replan.
  • Figure 5: Path planning results using the MCRL Method in different optimization steps: (a) Path optimized for distance; (b) Path optimized for both distance and turning angle; (c) Path optimized for distance, angle, and intersection count; (d) Object-based optimization.
  • ...and 4 more figures