diagnosed_defect string | case_prompt string | probable_causes list | assistant_answer string | evidence_reasoning string | recommended_checks string |
|---|---|---|---|---|---|
Ghosting (doubled geometry) in the reconstructed tree crowns. | A 3D reconstruction of a tree-lined drainage channel shows doubled outlines and overlapping lobes along several tree crowns on the north bank. The corresponding ground surface and nearby fence are aligned without visible duplication. In the overlapping source frames, branches occupy different positions, and the flight ... | [
"Wind-driven movement of branches between overlapping exposures.",
"Inconsistent image matching and depth estimation where the same moving foliage appears in different positions across frames."
] | The likely defect is ghosting in the tree crowns. Branch movement between overlapping exposures is the strongest indicated cause: the source frames show the branches in different positions, and gusty conditions were recorded. Inconsistent matching of that moving foliage can then produce doubled geometry. The aligned fe... | The duplicated outlines are confined to tree crowns, while the nearby fence and ground remain aligned. The source frames show branch positions changing between exposures, and gusty conditions are noted for the pass. Together, these clues support motion-related ghosting in vegetation rather than a displacement affecting... | Review the overlapping source frames and image timestamps for the affected crowns, and compare their alignment with stable features nearby. Check tie points and depth estimates in the foliage; if reprocessing is warranted, consider masking moving vegetation or excluding frames with visibly displaced branches. |
A hole in the reconstructed wall beneath the loading canopy. | A 3D model of a masonry service building contains a broad, irregular hole in the wall directly beneath a projecting loading canopy. The exposed wall on either side is reconstructed, but the surface inside the void has no mesh. Inspection shows that most images looking toward this area are blocked by the canopy; only on... | [
"Occlusion by the projecting canopy limited the number of images that captured the affected wall surface.",
"Insufficient clear, overlapping imagery from the single flight line may have left too few reliable correspondences for reconstruction.",
"Soft image detail in two overlapping frames may have further weak... | The defect is a hole in the reconstructed wall beneath the canopy. The inspection clues point first to occlusion: most images were blocked, and only one flight line clearly captured the surface. That limited coverage, together with soft detail in two overlapping frames, could leave too few reliable matches to build the... | The missing mesh is localized beneath the canopy, where inspection says most viewing images were blocked. The wall on both sides is present, which is consistent with a coverage problem specific to the occluded area rather than a general failure across the facade. The limited clear coverage and soft frames provide plaus... | Inspect the original images and camera positions to confirm which frames actually show the wall beneath the canopy. Review image sharpness, overlap, and tie-point coverage in that area. If additional capture is possible, acquire clear oblique views from angles that see beneath the canopy; otherwise, assess whether the ... |
UAV Photogrammetry Reconstruction Defect Root-Cause Question-Answer Dataset
This question-answer dataset describes quality defects in UAV photogrammetry reconstruction outputs alongside relevant inspection clues. It covers common issues such as local model bulging, ghosting, holes, and blurred edges; answers use the clues to infer possible acquisition or processing causes and explain the supporting evidence. The dataset supports defect diagnosis and root-cause reasoning for engineering surveying deliverables, quality inspection, acceptance, and technical support.
Technical Specifications
| Field | Type | Description |
|---|---|---|
| case_prompt | string | The original text describing the photogrammetry defect and any related quality inspection clues. |
| probable_causes | array | Likely data acquisition or processing causes inferred from the defect and inspection clues. |
| assistant_answer | string | The final user-facing answer combining the diagnosis, likely causes, and supporting evidence. |
| diagnosed_defect | string | The defect identified from the case prompt, such as local bulging, ghosting, holes, or blurred edges. |
| evidence_reasoning | string | An explanation connecting the inspection clues to the diagnosed defect and probable causes. |
| recommended_checks | string | Suggested follow-up checks or actions to validate the diagnosis or address the likely causes. |
Compliance Statement
| Authorization Type | CC-BY-NC-SA 4.0 (Attribution–NonCommercial–ShareAlike) |
| Commercial Use | Requires exclusive subscription or authorization contract (monthly or per-invocation charging) |
| Privacy and Anonymization | No PII, no real company names, simulated scenarios follow industry standards |
| Compliance System | Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs |
Source & Contact
- Downloads last month
- -