Pith. sign in

REVIEW 3 major objections 5 minor 26 references

FloodVision: Urban Flood Depth Estimation Using Foundation Vision-Language Models and Domain Knowledge Graph

T0 review · 3 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A zero-shot framework that fuses a vision-language model with a domain knowledge graph estimates urban flood depth to a mean error of 8.17 cm on crowdsourced images.

desk verdict A sensible zero-shot VLM-plus-knowledge-graph flood depth estimator whose headline 8.17 cm MAE is only as trustworthy as the unvalidated citizen-reported depths used as ground truth. read the letter →

arxiv 2509.04772 v1 pith:R5TRUZFB submitted 2025-09-05 cs.CV cs.AI

Zhangding Liu , Neda Mohammadi , John E. Taylor This is my paper · ORCID
classification cs.CVcs.AI
keywords urbanflooddepthestimationvision-languagemodelsknowledgegraphzero-shotlearningemergencyresponseresiliencecrowdsourcedimagerysubmergenceratio
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

FloodVision sets out to show that a foundation vision-language model—a model that reads images and answers in text—can estimate how deep floodwater covers a road, if its numeric guesses are pinned to measured dimensions of everyday objects. The method asks the model to name visible reference objects like tires, curbs, and people's knees, replaces the model's guessed heights with canonical heights from a curated knowledge graph, and multiplies each height by the object's estimated submerged fraction. On 110 crowdsourced flood photos, the averaged estimate lands within 8.17 cm of the reported depth, a 20.5% improvement over the model working alone (10.28 cm) and better than earlier CNN-based approaches. The appeal is practical: depth estimates with no training data and near real-time speed could support road-closure decisions, emergency routing, and flood mapping in cities that lack dense sensor networks.

What carries the argument

The load-bearing mechanism is FloodKG, a curated knowledge graph of canonical real-world dimensions for urban objects—vehicles, people, and infrastructure—stored as mean heights with standard deviations and linked by subClassOf and partOf relations. The pipeline uses the graph as an override: GPT-4o names candidate reference objects, matching graph entries replace the model's provisional height, and only the submergence ratio comes from the model's visual estimate. Per-object depth is height times submergence ratio, followed by statistical outlier filtering and aggregation into minimum, average, and maximum estimates. This retrieval-and-replace step is what turns qualitative scene understanding into physically anchored numbers, and its absence in the GPT-4o-only baseline is what the 20.5% error reduction measures.

What would settle it

Collect flood photos with independent water-depth measurements from the same streets and times—staff gauges, surveyed high-water marks, or pressure-sensor logs—run FloodVision on those photos, and compare its outputs to the independent values. If the error against independent truth is substantially larger than 8.17 cm, the claimed accuracy depends on the proxy labels.

Watch

Extended reading notes

Core claim

The central claim is that physically grounding a vision-language model's quantitative output removes enough hallucination to make zero-shot flood depth estimation viable. FloodVision prompts GPT-4o to identify up to three visually distinct reference objects, canonicalizes their names, retrieves verified height statistics from the FloodKG knowledge graph, estimates each object's submergence ratio (the fraction of its height under water), then filters outliers and aggregates into minimum, average, and maximum depth. The average variant achieves a mean absolute error of 8.17 cm and Pearson r of 0.51 on the 110-image test set, versus 10.28 cm and 0.44 for the GPT-4o-only baseline; the minimum and maximum variants also beat the baseline. The authors present this as preliminary evidence that semantic scene understanding plus external physical knowledge generalizes across diverse urban scenes where fixed object detectors and task-specific training fall short.

Load-bearing premise

The load-bearing premise is that the depth values residents typed into the crowdsourced flood-report platform are accurate enough to serve as ground truth; if those self-reports are biased or noisy, the reported 8.17 cm error does not reflect true estimation accuracy.

Editorial extensions

If this is right

  • Because FloodVision needs no task-specific training, the same pipeline can be pointed at a new city or event immediately, as long as visible reference objects exist in the photos.
  • The 20.5% drop over the VLM-only baseline indicates that external physical knowledge is a practical cure for quantitative hallucination in vision-language models, not just for flood depth.
  • The minimum/average/maximum aggregation turns one image into a depth range, which supports uncertainty-aware road-closure and routing decisions during active flooding.
  • Near-real-time operation on geotagged citizen photos or CCTV frames makes the method a plausible perception module for urban flood digital twins and alert systems.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The same retrieve-and-override pattern could transfer to other VLM measurement tasks—snow depth against parked cars, debris height against curbs, or wrack-line surveys—wherever a stable catalog of reference sizes can be built.
  • Because unmatched objects fall back to the model's own estimate, graph coverage is a likely driver of accuracy; a testable prediction is that MAE shrinks monotonically as FloodKG adds reference categories.
  • A natural ablative test the paper does not run is swapping GPT-4o for an open-weight vision-language model; if accuracy holds, the knowledge graph rather than the specific model is carrying the improvement.
  • The paper leaves open how to handle images with no recognizable reference objects; future work could combine water-surface texture or reflection cues with the graph to cover those cases.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 5 minor

Summary. FloodVision proposes a zero-shot flood depth estimation framework that combines GPT-4o's visual reasoning with a curated knowledge graph (FloodKG) of canonical object dimensions. The system identifies visible reference objects (e.g., car tires, curbs), retrieves physical heights from FloodKG, estimates submergence ratios via prompt-guided GPT-4o, and aggregates per-object depth estimates with outlier filtering. Evaluated on 110 crowdsourced images from MyCoast New York, the average estimate achieves a mean absolute error of 8.17 cm, a 20.5% improvement over a GPT-4o-only baseline, and is claimed to surpass prior CNN-based methods. The paper also discusses integration with digital twins and emergency response.

Significance. If the evaluation is trustworthy, FloodVision contributes a practical zero-shot alternative to task-specific CNN flood-depth estimators, addressing VLM quantitative hallucination through knowledge graph grounding. The approach is novel in combining a structural knowledge graph with a foundation VLM for physical measurement, and the authors provide a reproducible workflow (clear prompting strategy, RDF-based graph construction, public data source). The claimed 20.5% error reduction and near-real-time operation, if confirmed, would be valuable for smart city flood resilience. However, the significance is currently tempered by evaluation weaknesses: unvalidated citizen-reported ground truth, lack of uncertainty quantification, and cross-dataset comparisons that do not support the 'surpassing' claim.

major comments (3)
  1. [Section 3.1] The evaluation rests entirely on resident-reported depth estimates from MyCoast New York being treated as proxy ground truth, with no independent water-level measurements, sensor/gauge comparison, or screening for reporting bias. If these self-reports are rounded (e.g., to whole inches), biased by visual severity, or influenced by the same visual cues that GPT-4o uses, then the 2.11 cm MAE improvement over the baseline falls within plausible label noise. The paper must provide evidence for the reliability of these labels, such as comparison with gauge data, an analysis of report rounding, or a sensitivity study showing the conclusions are robust to label perturbations.
  2. [Table 1] Table 1 reports MAE and Pearson r values without error bars, confidence intervals, or significance tests. With n=110, the differences between FloodVision (average) at 8.17 cm and FloodVision (min) at 8.44 cm, or between FloodVision variants and the 10.28 cm baseline, may not be statistically significant. The authors should provide bootstrap confidence intervals or paired significance tests (e.g., Wilcoxon signed-rank test) to support the claim of improvement.
  3. [Section 3.2] The comparison with prior CNN-based methods (Chaudhary et al., Li et al., Alizadeh Kharazi and Behzadan, Akinboyewa et al.) is performed on different datasets with different label definitions, as the text itself concedes. The claim that FloodVision is 'surpassing prior CNN-based methods' is not supported by these incomparable numbers. A fair comparison would require re-running prior methods on the MyCoast dataset or evaluating FloodVision on the same benchmark used by those works; at minimum, the claim should be softened to 'does not require task-specific training' rather than 'surpasses'.
minor comments (5)
  1. [Section 2.2] The prompt description says the model 'outputs its reasoning' but the actual output is a constrained JSON of object, height, and submergence ratio; clarify whether the JSON includes textual reasoning or only structured fields.
  2. [Section 2.3] The knowledge graph construction cites 'verified' dimensions from sources including Wikipedia, but the aggregation and verification procedure is not detailed; a reader cannot assess how canonical heights for objects like 'curbs' or 'trash cans' were selected from heterogeneous sources.
  3. [Section 3.1] The paper does not describe the image selection criteria within the MyCoast database; if the 110 images were not randomly or systematically sampled, the reported performance may not reflect typical conditions.
  4. [Section 3.2 / Figure 3] The scatter plots and error distributions in Figure 3 are described qualitatively; adding numerical summaries of the residuals (e.g., median error, interquartile range) would strengthen the comparison.
  5. [General] There are minor typographical errors, e.g., 'anomalously depths' should be 'anomalous depths' and 'the 45°line' should be 'the 45-degree line'.

Circularity Check

0 steps flagged · score 1.0 of 10

No load-bearing circular step found; the knowledge graph and submergence-ratio model are independent of the target labels, with only minor non-central self-citations.

full rationale

FloodVision derives flood depth as object canonical height (from an externally sourced knowledge graph) multiplied by a VLM-estimated submergence ratio. Neither input is fitted to the MyCoast evaluation labels, and no equation in the paper defines the output in terms of the ground truth or vice versa. The knowledge graph heights come from AutomobileDimension, CDC NHANES, and design manuals/Wikipedia, i.e., independent physical references, not from the evaluation set. The submergence ratio is estimated by GPT-4o from visual waterline cues, which is a modeling choice rather than a circular reduction. The paper explicitly notes that prior CNN comparisons were obtained on different datasets, so those comparisons are not constructed to force a favorable outcome. The main weakness is that resident-reported depths are used as proxy ground truth without independent water-level validation; that is a data-validity threat, not a circularity in the derivation chain. The paper contains several self-citations (e.g., to the authors' prior digital twin and damage assessment work), but none of them is load-bearing for the central depth-estimation claim. There is also a mild post-hoc selection of the 'average' aggregation variant because it achieved the lowest MAE on the same test set, which raises multiple-comparison concerns but does not make the result true by construction. Overall, the central derivation is self-contained and empirically grounded, so circularity is minimal.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The system does not fit free parameters to the evaluation data; the knowledge graph heights come from external sources. The main assumptions are about the trustworthiness of the crowdsourced ground truth and the accuracy of the vision-language model's ratio estimates. No new physical entities are introduced.

assumptions (4)
  • domain assumption Resident-reported depth estimates from MyCoast New York are a valid proxy for true flood depth.
    Section 3.1 states these are 'treated as proxy ground truth values' without validation against measured water levels. The entire evaluation depends on this assumption.
  • domain assumption GPT-4o's submergence ratio estimates are reasonably accurate and unbiased.
    The method multiplies the model's ratio by known heights; if ratio estimates are biased, the final depth is biased. No independent check of ratio accuracy is provided.
  • domain assumption Canonical object heights in FloodKG match the actual objects in the test images.
    The knowledge graph stores mean heights from external sources, but the system assumes these apply to the specific vehicles, people, and infrastructure in each photo.
  • domain assumption The 110 MyCoast images are representative of urban flood scenarios.
    The dataset is self-selected by citizen reporters and may not cover the full range of flood scenes, which limits the generality of the reported performance.

how reviews work

0 comments
Cite this review

Pith. "Pith review of FloodVision: Urban Flood Depth Estimation Using Foundation Vision-Language Models and Domain Knowledge Graph." pith.science (2026). https://pith.science/paper/R5TRUZFB

@misc{pith2026250904772,
  author       = {Pith},
  title        = {Pith review of: FloodVision: Urban Flood Depth Estimation Using Foundation Vision-Language Models and Domain Knowledge Graph},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/R5TRUZFB}},
  note         = {Machine review of arXiv:2509.04772}
}
read the original abstract

Timely and accurate floodwater depth estimation is critical for road accessibility and emergency response. While recent computer vision methods have enabled flood detection, they suffer from both accuracy limitations and poor generalization due to dependence on fixed object detectors and task-specific training. To enable accurate depth estimation that can generalize across diverse flood scenarios, this paper presents FloodVision, a zero-shot framework that combines the semantic reasoning abilities of the foundation vision-language model GPT-4o with a structured domain knowledge graph. The knowledge graph encodes canonical real-world dimensions for common urban objects including vehicles, people, and infrastructure elements to ground the model's reasoning in physical reality. FloodVision dynamically identifies visible reference objects in RGB images, retrieves verified heights from the knowledge graph to mitigate hallucination, estimates submergence ratios, and applies statistical outlier filtering to compute final depth values. Evaluated on 110 crowdsourced images from MyCoast New York, FloodVision achieves a mean absolute error of 8.17 cm, reducing the GPT-4o baseline 10.28 cm by 20.5% and surpassing prior CNN-based methods. The system generalizes well across varying scenes and operates in near real-time, making it suitable for future integration into digital twin platforms and citizen-reporting apps for smart city flood resilience.

Figures

Figures reproduced from arXiv: 2509.04772 by the authors.

Figure 1
Figure 1. illustrates the FloodVision framework's workflow for estimating urban flood depth from RGB images. The process begins with GPT-4o performing semantic scene analysis to identify reference objects for depth estimation. Identified objects are queried from our FloodKG, which provides verified physical dimensions. GPT-4o then estimates each object's submerged ratio and returns the results in structured JSON. In the post-… view at source ↗

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

26 extracted references · 18 canonical work pages

  1. [1]

    Naser Lessani, and Zhenlong Li

    Temitope Akinboyewa, Huan Ning, M. Naser Lessani, and Zhenlong Li

    work page

  2. [2]

    Behzadan

    Bahareh Alizadeh Kharazi and Amir H. Behzadan. 2021. Flood depth mapping in street photos with image processing and deep neural networks. Computers, Environment and Urban Systems 88, (July 2021), 101628. https://doi.org/10.1016/j.compenvurbsys.2021.101628

    arXiv 2021

  3. [3]

    Athira, Yashwant B

    S. Athira, Yashwant B. Katpatal, and Digambar S. Londhe. 2023. Flood Modelling and Inundation Mapping of Meenachil River Using HEC-RAS and HEC-HMS Software. In Climate Change and Ocean Renewable Energy ,

    work page 2023

  4. [4]

    Chaudhary, S

    P. Chaudhary, S. D’Aronco, M. Moy De Vitry, J. P. Leitão, and J. D. Wegner

    work page

  5. [5]

    Fryar, Margaret D

    Cheryl D. Fryar, Margaret D. Carroll, Qiuping Gu, Joseph Afful, and Cynthia L. Ogden. Anthropometric reference data for children and adults : United States, 2015 -2018. Retrieved July 3, 2025 from https://stacks.cdc.gov/view/cdc/100478

    work page 2015

  6. [6]

    Aditya Kane and Sahil Khose. 2022. An Efficient Modern Baseline for FloodNet VQA. Retrieved March 21, 2024 from http://arxiv.org/abs/2205.15025

    arXiv 2022

  7. [7]

    Jingru Li, Ruying Cai, Yi Tan, Haijun Zhou, Abdul -Manan Sadick, Wenchi Shou, and Xiaoling Wang. 2023. Automatic detection of actual water depth of urban floods from social media images. Measurement 216, (July 2023), 112891. https://doi.org/10.1016/j.measurement.2023.112891

    arXiv 2023

  8. [8]

    Zhenlong Li, Wang ,Cuizhen, Emrich ,Christopher T., and Diansheng and Guo. 2018. A novel approach to leveraging social media for rapid flood mapping: a case study of the 2015 South Carolina floods. Cartography and Geographic Information Science 45, 2 (March 2018), 97 –110. https://doi.org/10.1080/15230406.2016.1271356

    arXiv 2018

Show all 26 references
  1. [9]

    Zhangding Liu, Neda Mohammadi, and John E. Taylor. 2025. Multi -Label Classification Framework for Hurricane Damage Assessment. https://doi.org/10.48550/arXiv.2507.02265

    2025 doi arXiv

  2. [10]

    Yujie Lu, Bo Liu, Wei Wei, Bo Xiao, Zhangding Liu, and Wensheng Li

  3. [11]

    Neda Mohammadi and John E. Taylor. 2021. Thinking fast and slow in disaster decision-making with Smart City Digital Twins. Nat Comput Sci 1, 12 (December 2021), 771–773. https://doi.org/10.1038/s43588-021-00174-0

    2021 doi

  4. [12]

    OpenAI et al. 2024. GPT -4 Technical Report. https://doi.org/10.48550/arXiv.2303.08774

    2024 doi arXiv

  5. [13]

    Xiyu Pan, Neda Mohammadi, and John E. Taylor. 2024. Predicting the Response Time of Emergency Medical Services during Urban Flooding: From Rule -Based to Data -Driven Approach. In Computing in Civil Engineering 2023, January 25, 2024. American Society of Civil Engineers, Corva...

    2024 doi

  6. [14]

    Wilkinson, and Richard J

    Maria Pregnolato, Alistair Ford, Sean M. Wilkinson, and Richard J. Dawson

  7. [15]

    Yuxuan Shao and Liwen Xu. 2025. Multimodal Natural Disaster Scene Recognition with Integrated Large Model and Mamba. Applied Sciences 15, 3 (January 2025), 1149. https://doi.org/10.3390/app15031149

    2025 doi

  8. [16]

    Mingrui Wu, Jiayi Ji, Oucheng Huang, Jiale Li, Yuhang Wu, Xiaoshuai Sun, and Rongrong Ji. 2024. Evaluating and Analyzing Relationship Hallucinations in Large Vision -Language Models. https://doi.org/10.48550/arXiv.2406.16449

    2024 doi arXiv

  9. [17]

    Nishant Yadav, Samrat Chatterjee, and Auroop R. Ganguly. 2020. Resilience of Urban Transport Network -of-Networks under Intense Flood Hazards Exacerbated by Targeted Attacks. Sci Rep 10, 1 (June 2020), 10350. https://doi.org/10.1038/s41598-020-66049-y

    2020 doi

  10. [18]

    Yueming Zhu, Xiyu Pan, Neda Mohammadi, and John Taylor. 2025. A Digital Twin Approach to Advancing River Emergency Response Systems in Smart Cities. 2025. . https://doi.org/10.24251/HICSS.2025.166

    2025 doi

  11. [19]

    AutomobileDimension.com

    Car dimensions of all makes with size comparison tools. AutomobileDimension.com. Retrieved July 3, 2025 from https://www.automobiledimension.com/

    2025

  12. [20]

    Retrieved July 3, 2025 from https://www.wikipedia.org/

    Wikipedia, the free encyclopedia. Retrieved July 3, 2025 from https://www.wikipedia.org/

    2025

  13. [21]

    Retrieved July 3, 2025 from https://mycoast.org/ny/flood-watch

    Flood Watch - MyCoast. Retrieved July 3, 2025 from https://mycoast.org/ny/flood-watch

    2025

  14. [2017]

    Transportation Research Part D: Transport and Environment 55, (August 2017), 67–81

    The impact of flooding on road transport: A depth-disruption function. Transportation Research Part D: Transport and Environment 55, (August 2017), 67–81. https://doi.org/10.1016/j.trd.2017.06.020

    2017 doi

  15. [2019]

    ISPRS Ann

    FLOOD -WATER LEVEL ESTIMATION FROM SOCIAL MEDIA IMAGES. ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci. IV-2/W5, (May 2019), 5–12. https://doi.org/10.5194/isprs-annals-IV-2-W5-5-2019

    2019 doi

  16. [2023]

    https://doi.org/10.1007/978-3-031-26967-7_9

    Springer Nature Switzerland, Cham, 113 –130. https://doi.org/10.1007/978-3-031-26967-7_9

    doi

  17. [2024]

    Comput.Urban Sci

    Automated floodwater depth estimation using large multimodal model for rapid flood mapping. Comput.Urban Sci. 4, 1 (May 2024), 12. https://doi.org/10.1007/s43762-024-00123-3

    2024 doi

  18. [2025]

    Developments in the Built Environment 21, (March 2025), 100610

    Generating synthetic images for construction machinery data augmentation utilizing context-aware object placement. Developments in the Built Environment 21, (March 2025), 100610. https://doi.org/10.1016/j.dibe.2025.100610

    2025

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.