{"paper":{"title":"Multimodal Object Detection Under Sparse Forest-Canopy Occlusion","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"Multimodal fusion of thermal-visible imagery and airborne optical sectioning improves human detection under sparse forest canopy where LiDAR penetration proves limited.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Mangal Kothari, Nitik Jain","submitted_at":"2026-05-14T18:39:51Z","abstract_excerpt":"Reliable detection of humans beneath forest canopy remains a difficult remote-sensing challenge due to sparse, structured, and viewpoint-dependent occlusion. This paper presents a multimodal proof-of-concept pipeline that integrates three complementary approaches: (i) experimental evaluation of LiDAR returns through vegetation to assess the feasibility of active sensing, (ii) visible--thermal image fusion using a multi-scale transform and sparse-representation framework to enhance human saliency, and (iii) synthetic-aperture image formation via Airborne Optical Sectioning (AOS) to suppress can"},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Results show that the tested terrestrial LiDAR configuration provides limited penetration for object-level detection, while visible-thermal fusion improves target visibility in low-contrast scenes and AOS enhances ground-plane detection in synthetic forest imagery. The fine-tuned YOLOv5 achieves a mean average precision of ~0.83 on the top three FLIR classes.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"The assumption that performance on the Teledyne FLIR thermal dataset and synthetic forest imagery will translate to real-world sparse forest-canopy occlusion scenarios with actual UAV or ground-based captures (abstract, results paragraph).","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"A proof-of-concept multimodal pipeline using LiDAR, visible-thermal fusion, AOS, and fine-tuned YOLOv5 reports mAP of ~0.83 on thermal classes and notes limited LiDAR penetration plus improved visibility from fusion and AOS in forest settings.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Multimodal fusion of thermal-visible imagery and airborne optical sectioning improves human detection under sparse forest canopy where LiDAR penetration proves limited.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"740ad302ee08c7c3e425c26742a324eeea8fcaffbe273dff79ac922d2fb6a36e"},"source":{"id":"2605.15326","kind":"arxiv","version":1},"verdict":{"id":"4a9905e6-0b46-4dda-ab61-f29d66db2f7b","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-19T16:03:47.644895Z","strongest_claim":"Results show that the tested terrestrial LiDAR configuration provides limited penetration for object-level detection, while visible-thermal fusion improves target visibility in low-contrast scenes and AOS enhances ground-plane detection in synthetic forest imagery. The fine-tuned YOLOv5 achieves a mean average precision of ~0.83 on the top three FLIR classes.","one_line_summary":"A proof-of-concept multimodal pipeline using LiDAR, visible-thermal fusion, AOS, and fine-tuned YOLOv5 reports mAP of ~0.83 on thermal classes and notes limited LiDAR penetration plus improved visibility from fusion and AOS in forest settings.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"The assumption that performance on the Teledyne FLIR thermal dataset and synthetic forest imagery will translate to real-world sparse forest-canopy occlusion scenarios with actual UAV or ground-based captures (abstract, results paragraph).","pith_extraction_headline":"Multimodal fusion of thermal-visible imagery and airborne optical sectioning improves human detection under sparse forest canopy where LiDAR penetration proves limited."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2605.15326/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_title_agreement","ran_at":"2026-05-19T16:31:18.294414Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"doi_compliance","ran_at":"2026-05-19T16:16:01.154453Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"claim_evidence","ran_at":"2026-05-19T14:41:54.199758Z","status":"completed","version":"1.0.0","findings_count":0},{"name":"ai_meta_artifact","ran_at":"2026-05-19T13:33:22.765016Z","status":"skipped","version":"1.0.0","findings_count":0}],"snapshot_sha256":"7b88d969798fb16252c6710591991c5ab1c7290d6d0f7550562588f0f675e9a7"},"references":{"count":10,"sample":[{"doi":"","year":2018,"title":"Survey of computer vision algorithms and applications for unmanned aerial vehicles,","work_id":"95f61f48-b142-4b88-8a60-1946153d1f38","ref_index":1,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"You Only Look Once: Unified, Real-Time Object Detection,","work_id":"0c9c853a-a296-4362-aa28-d84242ec77b2","ref_index":2,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2018,"title":"YOLOv3: An Incremental Improvement","work_id":"d737b3cc-9bd1-43d6-8310-c91e64b510f7","ref_index":3,"cited_arxiv_id":"1804.02767","is_internal_anchor":true},{"doi":"","year":2020,"title":"VIFB: A Visible and Infrared Image Fusion Benchmark,","work_id":"38bbcc5d-13ca-4b48-9828-016de6f33ee4","ref_index":4,"cited_arxiv_id":"","is_internal_anchor":false},{"doi":"","year":2016,"title":"Image Fusion with Convolutional Sparse Representation,","work_id":"db9f1ad4-63e7-4a08-b2c0-8460e7d4dc77","ref_index":5,"cited_arxiv_id":"","is_internal_anchor":false}],"resolved_work":10,"snapshot_sha256":"26a75a9785a2ff6ec38b80da59d31e3cc36dd7e76c7fc08048e4103d292ac50c","internal_anchors":1},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}