Pith. sign in

REVIEW 8 cited by

SegEarth-R1: Geospatial Pixel Reasoning via Large Language Model

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2504.09644 v1 pith:OVKPPTWW submitted 2025-04-13 cs.CV

Kaiyu Li , Zepeng Xin , Li Pang , Chao Pang , Yupeng Deng , Jing Yao , Guisong Xia , Deyu Meng
show 2 more authors
Zhi Wang Xiangyong Cao
This is my paper
classification cs.CV
keywords reasoningsegearth-r1implicitsegmentationlanguagemaskremotesensing
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Remote sensing has become critical for understanding environmental dynamics, urban planning, and disaster management. However, traditional remote sensing workflows often rely on explicit segmentation or detection methods, which struggle to handle complex, implicit queries that require reasoning over spatial context, domain knowledge, and implicit user intent. Motivated by this, we introduce a new task, \ie, geospatial pixel reasoning, which allows implicit querying and reasoning and generates the mask of the target region. To advance this task, we construct and release the first large-scale benchmark dataset called EarthReason, which comprises 5,434 manually annotated image masks with over 30,000 implicit question-answer pairs. Moreover, we propose SegEarth-R1, a simple yet effective language-guided segmentation baseline that integrates a hierarchical visual encoder, a large language model (LLM) for instruction parsing, and a tailored mask generator for spatial correlation. The design of SegEarth-R1 incorporates domain-specific adaptations, including aggressive visual token compression to handle ultra-high-resolution remote sensing images, a description projection module to fuse language and multi-scale features, and a streamlined mask prediction pipeline that directly queries description embeddings. Extensive experiments demonstrate that SegEarth-R1 achieves state-of-the-art performance on both reasoning and referring segmentation tasks, significantly outperforming traditional and LLM-based segmentation methods. Our data and code will be released at https://github.com/earth-insights/SegEarth-R1.

Discussion (0). Sign in to comment.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. PixDLM: A Dual-Path Multimodal Language Model for UAV Reasoning Segmentation

    cs.CV 2026-04 unverdicted novelty 7.0 of 10

    The work introduces the UAV Reasoning Segmentation task, the DRSeg benchmark dataset, and PixDLM as a baseline dual-path multimodal language model for reasoning-based segmentation in aerial imagery.

  2. GeoSelect: Spatial-Program Execution for Training-Free Referring Remote Sensing Image Segmentation

    cs.CV 2026-07 conditional novelty 6.5 of 10

    A training-free pipeline synthesises referring expressions into a typed geometric DSL, executes them over scored candidate boxes, and reaches 58.86 mIoU on RRSIS-D—over twice the previous training-free best.

  3. CROSS: Cascaded Distillation and Dual-Constraint Grounding for Remote Sensing Referring Segmentation

    cs.CV 2026-08 conditional novelty 6.0 of 10

    CROSS adds SAM-derived structural distillation and spatial-contrastive negatives to a SigLIP-SAM pipeline, achieving state-of-the-art cIoU on two remote sensing referring segmentation benchmarks.

  4. OSMDA: OpenStreetMap-based Domain Adaptation for Remote Sensing VLMs

    cs.CV 2026-03 conditional novelty 6.0 of 10

    A remote-sensing VLM can be adapted by having it read rendered OpenStreetMap maps paired with satellite images, then fine-tuning it on satellite images alone.

  5. Annotation-Free Open-Vocabulary Segmentation for Remote-Sensing Images

    cs.CV 2025-08 conditional novelty 6.0 of 10

    SegEarth-OV performs annotation-free open-vocabulary segmentation of remote-sensing images by upsampling CLIP features, removing global bias, and distilling optical knowledge into a SAR encoder.

  6. More with Less: a Large Scale Remote Sensing VLM with a Simple Recipe

    cs.CV 2026-07 conditional novelty 5.0 of 10

    An unmodified general-purpose VLM trained with multi-task RL and a SAM3 tool reaches top results on most remote sensing zero-shot benchmarks, with gains the paper attributes to training-data diversity rather than arch...

  7. CangLing-KnowFlow: A Unified Knowledge-and-Flow-fused Agent for Comprehensive Remote Sensing Applications

    cs.AI 2025-12 reject novelty 5.0 of 10

    CangLing-KnowFlow combines a procedural knowledge base, dynamic workflow repair, and memory to beat ReAct/Reflexion on remote-sensing workflow tasks, but the benchmark is drawn from the same tasks used to build its kn...

  8. Reasoning Segmentation for Images and Videos: A Survey

    cs.CV 2025-05 conditional novelty 3.0 of 10

    The paper organizes the field of reasoning segmentation into image and video tracks, cataloging 26 methods, 12 metrics, and 29 datasets.

Pith tools