Scientists from China’s Zhejiang University have developed a multi-scale method to assess vegetation conditions inside PV power plants.
The new approach combines field data, unmanned aerial vehicle (UAV) images, and Sentinel-2 satellite data. It is intended to correct the biased normalized difference vegetation index (NDVI) results yielded by an analysis of only satellite images.
“For quantitative assessments of the impacts of PV panel installations on vegetation at the regional scale, it is crucial to accurately retrieve the true conditions of vegetation within PV power plants, from satellite-recorded signals,” said the academics. “In this study, we propose a solution focused on NDVI from Sentinel-2 data by utilizing UAV images to mediate field and satellite data integration. Our objectives are to: quantitatively evaluate the relationship of vegetation under and between the panels and develop a framework to use UAV data to correct the Sentinel-2 NDVI data for evaluating the true conditions of vegetation in solar plants.”
The research focused on nine PV facilities in mainland China, across diverse climate zones and land use types.
Overall, 76 visits to the different parks were conducted to establish ground truth, during which 3,295 pairs of under-panel and between-panel images were collected to determine whether a correlation exists. Additionally, the DJI Phantom 4 Multispectral quadcopter and the DJI Mavic 3 Multispectral quadcopter were used to collect ultra-high-resolution aerial remote sensing data. They flow on days with good visibility at a height of 200 m. These were matched with Level-2 Sentinel-2 data of high spatial and temporal resolution.
“We modified a U-Net model to calibrate the Sentinel-2 NDVI to better capture the complex relationship between UAV and satellite data,” explained the researchers. “We selected 5,054 patches of 64 × 64 pixels as inputs for model training, with 80% used for training and 20% for validation. To further evaluate the model’s generalization capability, we tested it on six solar plants across China. We then trained the modified U-net model using all available data, which was subsequently used to calibrate the Sentinel-2 NDVI and mosaicked the predicted NDVI for each plant for further analyses.”
Their analysis showed that the corrected outputs significantly reduced the discrepancy between Sentinel-2 NDVI and ground truth, which was measured at the field days. However, perfect agreement was not achieved. Overall, they found that before the correction, Sentinel-2 NDVI tended to overestimate values in low-NDVI plants and underestimate them in high-NDVI ones.
“Our results show that vegetation beneath and between the photovoltaic panels is strongly correlated, and the inclusion of under-panel vegetation raises the average normalized difference vegetation index from 0.248 ± 0.158 to 0.298 ± 0.193,” the results showed. “Compared to satellite estimates alone, our method reduces bias by 16.98%. At the regional scale, approximately 61.59% of the power plants did not suppress vegetation growth. This approach enables more accurate environmental assessments of the development of photovoltaic power plants and supports better-informed planning and management of solar energy infrastructure.”
Their findings were presented in “Leveraging unmanned aerial vehicle images improves vegetation mapping in photovoltaic power plants,” published in Communications Earth & Environment. Scientists from China’s Zhejiang University, the Chinese Academy of Sciences, and the Beijing Institute of Space Mechanics and Electricity have participated in the study.
This content is protected by copyright and may not be reused. If you want to cooperate with us and would like to reuse some of our content, please contact: editors@pv-magazine.com.
By submitting this form you agree to pv magazine using your data for the purposes of publishing your comment.
Your personal data will only be disclosed or otherwise transmitted to third parties for the purposes of spam filtering or if this is necessary for technical maintenance of the website. Any other transfer to third parties will not take place unless this is justified on the basis of applicable data protection regulations or if pv magazine is legally obliged to do so.
You may revoke this consent at any time with effect for the future, in which case your personal data will be deleted immediately. Otherwise, your data will be deleted if pv magazine has processed your request or the purpose of data storage is fulfilled.
Further information on data privacy can be found in our Data Protection Policy.