L-AVI Model Demo

Adaptive lunar vision reconstruction for navigation, safety, and exploration.

● TRL-4 RESEARCH DEMO

Model Intelligence Pipeline

L-AVI fuses illumination prediction with reconstruction to recover critical visual features under extreme low-light and lunar dust conditions. This demo highlights the pipeline flow and quality metrics.

Raw Lunar Frames

Illumination Prediction

Visibility Reconstruction

Terrain Insight Output

Performance Signals

Live demo metrics from recent evaluation batches.

PSNR

28.7 dB

Visibility fidelity

SSIM

0.89

Structural accuracy

Latency

42 ms

Realtime readiness

Terrain Gain

+63%

Feature recoverability

Incident Replay Timeline

Example of how L-AVI enhances visibility during critical mission windows.

Sample Outputs

Representative outputs from simulated lunar surface datasets.

Raw Low-Light Frame