MERIDIAN

data & geospatial intelligence
  • Italiano
  • English

Monitoring Large-Scale Photovoltaic Assets

CASE STUDY Energy Infrastructure

32oE meridian

The Scenario

For energy asset management companies and investment funds, verifying the actual progress of remote, globally distributed photovoltaic construction sites represents a significant logistical and economic challenge. Traditional on-site inspection methods are costly, intermittent, and lack an objective historical perspective.

The Technical Challenge

The objective is to automate the detection of the ground footprint of photovoltaic modules using multi-temporal and multi-sensor satellite data. The core difficulty lies in accurately distinguishing panels from other surfaces and managing environmental variables (such as cloud cover) and imaging parameters (spatial and temporal resolution) to obtain reliable statistical data.

A solar farm from space imaged by high resolution satellite sensors. Left hand side: false colour RGB image from Sentinel 2 Multi-Spectral Instrument (MSI) data, using band B12 for the red channel, B09 for the green and B03 for the blue one. Right hand side: Sentinel 1 synthetic aperture radar (SAR) image, VH polarization. Time lag between the two images is one day, while the pixel spacing is 20m. Source: contains modified Copernicus ESA Sentinel data (2026), processed via Copernicus Data Space Ecosystem (CDSE) by studio MERIDIAN-Data & Geospatial Intelligence.

The Solution

Development of an automated data pipeline integrating:


• Multi-Source Data Ingestion: Utilizing satellite time series for frequent and historical coverage.
• Computer Vision & Machine Learning: Implementing semantic segmentation algorithms specifically trained for photovoltaic module recognition.
• Time Series Analysis: Processing the growth curve of the occupied area, allowing for the precise identification of plant expansion phases.

The Results

• Precision: Calculation of the total occupied area with a high degree of confidence, enabling the estimation of nominal installed capacity.
• Total Remote Monitoring: Reducing the need for frequent physical site visits to verify work progress (from Early Stage to Completion).
• Independent Audit: Providing frequent reports for monitoring the contractual compliance of installation contractors.

Area covered by solar panels vs time. The area is estimated (grey dots) by processing a time series of SAR images acquired in the period 2016-2026 by using an automatic data pipeline to discriminate solar panels vs the background. The regression curve (in blue) is used to smooth the estimate time series and highlight area trends. The plant reaches its full installed potential in about two years reaching a final extension of about 30 km2 (3000 ha). Source: contains modified Copernicus ESA Sentinel data (2026), processed via Copernicus Data Space Ecosystem (CDSE) by studio MERIDIAN-Data & Geospatial Intelligence.

Strategic Value

Satellite monitoring transforms raw imagery into a Key Performance Indicator (KPI). This allows stakeholders to promptly detect anomalies or schedule delays, optimizing the financial and operational management of energy assets.

Future Developments

  • Data fusion with Sentinel-2 MSI multispectral imagery.
  • Integration of a classification layer into the pipeline to enhance area estimation accuracy.

Stay tuned for more updates!

Disclaimer

The analyses reported herein are for illustrative and methodological purposes only.

Credits

This page contains modified Copernicus ESA Sentinel data (2026), processed via Copernicus Data Space Ecosystem (CDSE) by MERIDIAN – Data & Geospatial Intelligence.

Looking to optimize your asset monitoring?

Discover how to transform geospatial intelligence into a competitive advantage for your business.

Get in touch

Raffaele Grasso, MERIDIAN – Data & Geospatial Intelligence

, , , ,