Hardware
- 10-inch FPV drone platform
- PX4 flight controller
- Laser distance sensors
- Optical-flow sensing systems
- Vision-processing hardware
- Pipe inspection camera systems
Robotics
Autonomous Pipe Inspection FPV Drone System
Developed an autonomous confined-space FPV drone system for pipe inspection using PX4 firmware modification, optical-flow fusion, laser-based centering, and YOLO-based defect segmentation.
Summary
Engineering context
Developed an autonomous confined-space FPV drone system for pipe inspection using PX4 firmware modification, optical-flow fusion, laser-based centering, and YOLO-based defect segmentation.
- Category
- Robotics
- Year
- Jan 2025 - Present
- Status
- Research and Prototype Deployment
- Context
- University of Regina MSc (Jan 2025 - Present)
My Role
Autonomous Systems and Computer Vision Engineer
Technical Stack
System Architecture
- Custom 10-inch FPV platform provided confined-space mobility
- Modified PX4 firmware integrated custom optical-flow fusion behavior
- Six laser sensors enabled autonomous pipe centering
- Vision system detected pipe entrances and inspection targets
- YOLO segmentation system identified pipe defects during inspection
- Autonomous navigation logic stabilized flight inside confined cylindrical structures
Engineering Challenges
- Stabilizing autonomous flight in confined cylindrical environments
- Modifying PX4 sensor-fusion behavior
- Maintaining accurate pipe centering during motion
- Detecting pipe defects under variable lighting conditions
- Integrating computer vision with autonomous flight systems
Hardware / Firmware / Software
Firmware
- Modified PX4 firmware
- Sensor-fusion integration
- Autonomous stabilization logic
- Embedded flight-control modifications
Software
- Computer vision inspection systems
- YOLO segmentation pipeline
- Pipe-entry detection algorithms
- Autonomous navigation systems
- Defect-detection systems
Sensors
- Optical-flow sensors
- Laser distance sensors
- Vision-based inspection systems
Protocols
- MAVLink
- Embedded serial communication
Results / Outcomes
- Successfully demonstrated autonomous pipe traversal
- Implemented laser-based autonomous centering system
- Integrated YOLO-based defect segmentation
- Developed confined-space autonomous inspection architecture
- Demonstrated proof-of-concept autonomous inspection flight inside pipe structures
Gallery
Engineering Notes
Autonomous Pipe Navigation
The drone platform used:
- six laser distance sensors
- optical-flow sensing
- autonomous stabilization logic
to maintain centered flight inside pipes and confined cylindrical structures.
The autonomous navigation system continuously estimated relative position within the pipe to reduce drift and maintain stable traversal.
PX4 Firmware Modification
The PX4 firmware was modified to support:
- custom optical-flow fusion behavior
- autonomous centering logic
- sensor-integration modifications
- confined-space flight behavior
Additional PX4 features were selectively disabled or adapted to improve stability and performance for indoor pipe-inspection scenarios.
Computer Vision and Defect Detection
A computer-vision subsystem was developed to:
- detect pipe entrances
- analyze inspection imagery
- identify defects inside the pipe
YOLO segmentation models were integrated for defect detection and localization during inspection operations.
Proof-of-Concept Demonstration
A proof-of-concept autonomous inspection demonstration was conducted using a pipe-like barrel structure.
The system demonstrated:
- stable confined-space flight
- autonomous centering
- sensor-fusion operation
- visual inspection capability
inside cylindrical structures representative of pipe-inspection environments.
Related projects
More engineering case studies in this domain
Autonomous Drone System for Copper Cable Recovery and Infrastructure Inspection
Developed an autonomous PX4-based drone platform for utility-infrastructure inspection and copper-cable recovery using Mask R-CNN segmentation, monocular metric depth estimation, and vision-guided autonomous positioning.
Autonomous Solar Panel Cleaning Robot
Developed an autonomous robotic solar-panel cleaning system using computer vision, embedded control, and environmental sensing for industrial solar-farm deployment.
AJNA Autonomous ICU Assistance and Telepresence Robot
Developed an autonomous healthcare robotics platform during the COVID-19 pandemic for ICU assistance, telepresence communication, patient monitoring, and autonomous medicine delivery.