Deep Learning Recovers 1,353 Historical Astronomical Plates: China's Century-Old Night Sky Photographs Digitally Restored

Available in: 中文
2026-04-07T22:48:17.342Z·1 min read
China has been systematically collecting nighttime astronomical photographs since 1900. Now, a Transformer-based deep learning model has rescued over 1,300 plates that were previously unprocessable...

China has been systematically collecting nighttime astronomical photographs since 1900. Now, a Transformer-based deep learning model has rescued over 1,300 plates that were previously unprocessable due to environmental deterioration.

The Archive

China's historical astronomical plate collection spans over a century of observations, digitized using optical scanners. These plates contain irreplaceable scientific data about:

The Problem

Suboptimal early storage conditions and environmental deterioration caused thousands of digitized plates to fail standard processing:

The AI Solution

A Transformer-based classification model that:

  1. Takes cutouts of detected sources as input
  2. Uses multi-scale feature fusion to identify trustworthy stellar sources
  3. Trains on successful plates to learn what reliable sources look like
  4. Applied to 1,883 failed plates — successfully registered 1,353

Results

MetricValue
Failed plates attempted1,883
Successfully recovered1,353 (71.9%)
MethodTransformer + multi-scale feature fusion
Reference catalogsAstrometry.net + Gaia

Why It Matters

The Chinese astronomical plate archive joins similar digitization efforts worldwide, using AI to unlock scientific data trapped in deteriorating physical media.

↗ Original source · 2026-04-07T00:00:00.000Z
← Previous: US Lunar Spacecraft Toilet Malfunction: $150 Million Waste Management System Breaks Down on NASA MissionNext: Xi Jinping Congratulates Four Jiaotong Universities on 130th Anniversary: China's Engineering Education Legacy →
Comments0