SatoyamaCT: A Multi-Axis Night-IR Camera-Trap Benchmark for Monitoring Crop-Damaging Wildlife in Japanese Agroforestry
Abstract
Crop and forest damage from sika deer, wild boar, and Japanese macaque is a serious economic problem in Japanese satoyama, where farmland and forest intermingle. Camera traps enable scalable monitoring, yet existing benchmarks evaluate recognition in-distribution, rarely prioritize night infrared imagery, and do not jointly address regional domain shift, novel-species detection, and uncertainty-based triage on a single dataset. We introduce the Satoyama Camera Trap Dataset (SatoyamaCT), 12,642 expert-verified crops from night-IR-dominant camera traps across three satoyama regions. An iterative annotation protocol combining BioCLIP embeddings with confidence-ordered confirmation reduced expert effort while all labels were verified by domain ecologists; species-level inter-annotator agreement reached Cohen's kappa = 0.900. A multi-axis protocol jointly evaluates domain generalization across region, camera placement, and illumination; open-set novel-species detection; and selective prediction, all under capture-event-based leakage control. Difficulty separates into distinct failure modes: a data-inherent regional gap of 16 to 27 percentage points persists in Wakayama across four backbones and domain-generalization methods including CORAL, DANN, and GroupDRO, and open-set detection reaches AUROC 0.93 to 0.96 overall yet degrades jointly with classification in Wakayama. Selective prediction recovers Wakayama accuracy from 0.593 to 0.815 at 50% coverage, supporting an uncertainty-aware triage workflow for practical pest monitoring in agroforestry.
Related
- Benchmark indicators for evaluating sea turtle rehabilitation outcomes at a wildlife rehabilitation center on Tenerife island, Spain (1982–2025)
- Toward closing a heartwater monitoring gap with a new pan-species ELISA suitable for domestic ruminants, wildlife, and camels.
- Opportunities and challenges in training volunteer wildlife detection dog-handler teams for invasive plant monitoring
- Dolycoris baccarum (L.) (Hemiptera: Pentatomidae) Infestation Induced Differential Changes in Plant Physiology: Preference and Damaging Nature to Different Sesame (Sesamum indicum L.) Varieties
- Trimethoprim suppress phage activity and reduces toxin production in a clinical STEC strain, even in the presence of DNA-damaging antibiotics
BibTeX
@misc{inoshita2026satoyamact,
title = {{SatoyamaCT: A Multi-Axis Night-IR Camera-Trap Benchmark for Monitoring Crop-Damaging Wildlife in Japanese Agroforestry}},
author = {Keito Inoshita and Kohei Hisayama and Haruto Sugeno and Kota Nojiri},
howpublished = {arXiv},
year = {2026},
month = sep,
doi = {10.48550/arxiv.2609.30278},
eprint = {2609.30278},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2609.30278},
}
RIS
TY - UNPB
TI - SatoyamaCT: A Multi-Axis Night-IR Camera-Trap Benchmark for Monitoring Crop-Damaging Wildlife in Japanese Agroforestry
AU - Keito Inoshita
AU - Kohei Hisayama
AU - Haruto Sugeno
AU - Kota Nojiri
PY - 2026
DA - 2026/09/01
PB - arXiv
DO - 10.48550/arxiv.2609.30278
UR - https://arxiv.org/abs/2609.30278
AB - Crop and forest damage from sika deer, wild boar, and Japanese macaque is a serious economic problem in Japanese satoyama, where farmland and forest intermingle. Camera traps enable scalable monitoring, yet existing benchmarks evaluate recognition in-distribution, rarely prioritize night infrared imagery, and do not jointly address regional domain shift, novel-species detection, and uncertainty-based triage on a single dataset. We introduce the Satoyama Camera Trap Dataset (SatoyamaCT), 12,642 expert-verified crops from night-IR-dominant camera traps across three satoyama regions. An iterative annotation protocol combining BioCLIP embeddings with confidence-ordered confirmation reduced expert effort while all labels were verified by domain ecologists; species-level inter-annotator agreement reached Cohen's kappa = 0.900. A multi-axis protocol jointly evaluates domain generalization across region, camera placement, and illumination; open-set novel-species detection; and selective prediction, all under capture-event-based leakage control. Difficulty separates into distinct failure modes: a data-inherent regional gap of 16 to 27 percentage points persists in Wakayama across four backbones and domain-generalization methods including CORAL, DANN, and GroupDRO, and open-set detection reaches AUROC 0.93 to 0.96 overall yet degrades jointly with classification in Wakayama. Selective prediction recovers Wakayama accuracy from 0.593 to 0.815 at 50% coverage, supporting an uncertainty-aware triage workflow for practical pest monitoring in agroforestry.
ER -