- Lead Organisation
Australian National UniversityLocation
Acton, Australian Capital TerritoryARENA Program
11 April 2016
30 August 2020
- Project PartnersWhite Knight Unmanned Arial Systems PTY LTD, Vast Solar Pty Ltd, NICTA, FRV Services Australia Pty LtdThis solar project was completed on 30 August 2020.
The Robotic vision system project will develop a cost-effective robotic inspection system for automated monitoring of concentrating solar power and solar photovoltaic power plants using cameras mounted on ground and aerial drones.
After initial construction, the primary costs associated with a solar power plant are cleaning and maintenance. Accumulation of dust, dirt, or other contaminants, and defective panels can be a significant source of lost power production.
This is the first time in the world that intelligent drones and innovative computer vision solutions will be used for automated inspection of solar power plant facilities.
This project will develop a system that measures the accumulated dust on the panels and diagnose possible defects to determine efficient maintenance and cleaning procedures.
Advanced data driven and deep learning based techniques will aim to dramatically improve the accuracy of analysis. Novel and inexpensive sensor configurations will be used to deliver automatic methods for mapping, visualising, and optimal route planning, which could ultimately enable cost-effective cleaning procedures.
Report: A Robotic Vision System for Rapid Inspection and Evaluation of Solar Plant Infrastructure
This project investigated the feasibility of automated collection and analysis of visual data for predictive inspection and evaluation of the condition of concentrating solar power (CSP) and solar photovoltaic (PV) plants.
Drones will be used to identify defects and measure dust concentration in a timely manner. This information is essential for developers, owners, and operators of solar power plants that wish to quantify and minimise power losses due to dust accumulation, and confirm the underlying performance of the power plant to determine if and when cleaning provides an economic payback.
The Robotic vision system project has a performance objective of a 90 per cent correct estimation of dust accumulation and 90 per cent detection rate of visible defects. With these objectives, the efficiency of the solar power systems can be improved up to 8 per cent.
Read more about solar PV.