cs.CVNov 13, 2026

Lens: Training-Free Semantic Search for Cross-Tile Object and Scene Retrieval in Satellite Imagery

Authors: Ata Akbari Asanjan, Anit Upadhyaya, Jeremy Kravitz, Subash Yeggina

Organizations: Pixxel Space Technologies, El Segundo, United States · Pixxel Space India Pvt Ltd, Bengaluru, India

Abstract

We developed Lens, a training-free system for semantic object and scene search across large satellite mosaics that requires no detector trained for the class of interest. Lens represents each tile with a frozen, self-supervised vision transformer (DINOv3), removes the positional bias in its dense features by projecting them onto the subspace orthogonal to a positional basis estimated once from a constant image, and centers the debiased features on the mean feature computed across the whole area before comparison. Lens forms a query by averaging the centered features of a small neighborhood of patches around a chosen point, or of an entire tile for scene-level search, and matches it against the index by cosine similarity. It stores the centered patch features in a byte-quantized similarity index that supports real-time range search over a mosaic of thousands of patches, and a coarser whole-tile embedding index for scene-level retrieval. A guided upsampling network (AnyUp) runs once per tile to recover pixel-accurate boundaries beyond the native feature grid, and the result is cached so later queries reuse it. We evaluated Lens on the 5,128-tile DFC2020 test set and ran it as an interactive tool on five Sentinel-2 areas spanning agricultural, industrial, and desert terrain; the pipeline is not tied to a specific sensor. On DFC2020 a single click reaches 0.09 to 0.80 pixel F1 across eight land-cover classes at the per-class best threshold, with a mean F1 of 0.39, and it beats a pairwise INSID3 baseline on six of the eight classes at roughly four hundred times lower query cost. Indexing 1.31 million patches takes 169 s on one CPU thread, and a click returns a rendered overlay in 6 to 10 ms.

Figures & tables

Explore similar work

CardsList
  1. Open-SAT: LLM-Guided Query Embedding Refinement for Open-Vocabulary Object Retrieval in Satellite Imagery

    May 6, 2026Md Adnan Arefeen, Biplob Debnath, Ravi K. Rajendran +2Remote Sensing Image UnderstandingImage-Text Retrieval

  2. Zero-Shot Satellite Image Retrieval through Joint Embeddings: Application to Crisis Response

    May 6, 2026James Walsh, William Fawcett, Grace Colverd +1Geospatial Foundation ModelsDisaster Response

  3. GeoDIVE: Geospatial Descent Informed by Value Estimation

    Nov 3, 2026Hamza Iqbal, Dan Cher, Eric Xing +4Image RetrievalMultimodal IR