Video demonstrations

Click a video to play, or Explore in 3D to rotate, zoom and scrub through the flight.

Figure 2 on more scenes

SAME LAYOUT AS FIGURE 2 IN THE MAIN PAPER
Fushimi: trajectories of FAR, MAGICIAN and EPIC with GT depth (top), and of the FAR planner driven by FAR, PanoVGGT and DAP geometry (bottom)
Full size

(a) Planners with GT depth. (b) The FAR planner driven by each geometry source, rendered with the point clouds from predicted depth. Color encodes flight progress.

Shared reviewer questions

Reconstruction ablations

* qiHD Q1; eJ88 Q1

FAR combines panoramic depth prediction, metric alignment and persistent mapping to select the next capture position.

INPUTPanoramic RGB
+ metric poses
PERCEPTIONSRA depth
prediction
GEOMETRYMetric alignment
+ persistent map
PLANNINGNext capture
position

FAR achieves the lowest AbsRel among the shown reconstruction ablations.

Selected depth ablation results from Table 3 in the main paperw/o SRA0.1914w/o alignment0.2073w/o refinement0.1337FAR0.1196AbsRel ↓
Selected AbsRel results cited from Table 3 in the main paper. Lower is better.

With the same planner, FAR achieves higher coverage and AUC than PanoVGGT using predicted depth.

Selected closed loop results cited in the reply (Table 1 in the main paper).
MethodDepth sourceCoverage at ε = 1.0 m ↑AUC ↑
FARGT depth (reference)85.6%0.663
PanoVGGTPredicted depth69.2%0.421
FARPredicted depth83.3%0.634

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

Same FAR planner. Only the two predicted depth rows are compared for highlighting; GT depth is a reference.

Comparison under matched sensing

* qiHD Q2; zcV1 Q1

OMNIDIRECTIONAL · 50 M · GT DEPTH

With matched sensing, FAR achieves the highest coverage.

Coverage under the same omnidirectional 50 m sensorEPIC67.4%MAGICIAN80.5%FAR85.2%Coverage at ε = 1.0 m (%) ↑

With the sensor fixed, FAR retains higher coverage and shorter paths.

Table R1. Comparison with the same sensor configuration.
Sensor configurationMethodCoverage at ε = 1.0 m ↑Path length (m) ↓AUC ↑
Native (Table 1 in the main paper)EPIC65.6%18070.377
Native (Table 1 in the main paper)MAGICIAN79.4%14000.500
Native (Table 1 in the main paper)FAR85.6%11110.663
Omnidirectional, 50 mEPIC67.4%17560.437
Omnidirectional, 50 mMAGICIAN80.5%17950.604
Omnidirectional, 50 mFAR85.2%15710.623

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

Tables R1/R4 · GT depth, ε = 1.0 m. Native and matched sensor groups are compared separately. Added results are averaged over multiple start positions per scene.

Planning ablations

* zcV1 Q1; eJ88 Q4

Compared with FAR, removing Adaptive Initialization lowers final coverage; removing peak ratio termination lengthens the path with little additional coverage.

Coverage versus relative flight path (ε = 1.0 m).

Adaptive Initialization adds 4.1 pp coverage; removing peak ratio termination adds only 0.4 pp but increases path length by 21%.

Table R5. Planning component ablations.
SettingCoverage at ε = 1.0 m ↑Path length (m) ↓
FAR85.6%1111
w/o Adaptive Initialization (one fixed parameter set for all scenes)81.5%888
w/o termination based on the peak ratio86.0%1345
MAGICIAN79.4%1400
EPIC65.6%1807

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

Tables R5/R13 · Native sensing, GT depth, ε = 1.0 m. Added results are averaged over multiple start positions per scene. Fixed settings use Colosseum’s adaptive values for all scenes. Removing peak ratio termination means η = 0.

Computational cost and decision rate

* qiHD Q3; zcV1 Q4

FAR has lower reported computational and memory costs than MAGICIAN and plans faster than EPIC.

Table R2. Computational costs and memory usage.
MethodPerception (ms) ↓Mapping (ms) ↓Planning (ms) ↓RAM (GB) ↓GPU memory (GB) ↓
EPIC—55592.20
MAGICIAN128217538434.115.1
FAR912149442.58.7

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

R2/R8 · AMD EPYC 9T24, 32 vCPUs; RTX 5880 Ada, 48 GB. Latency: mean of per scene medians. Memory: mean of per run peaks.

VIEWPOINT DECISIONS0.78 Hz0.72–0.81 Hz across 11 scenes

Decision frequency, not total mission speed.

FAR achieves a viewpoint decision rate of 0.78 Hz.

Table R9. FAR latency and closed loop rate.
MetricMedianP95
Panorama rendering and readback (ms)4272
Perception per window (ms)9121002
Mapping (ms)149221
Planning per decision (ms)44133
Closed loop rate (Hz)0.78—

Reviewer specific questions

qiHDCommon rendering pipeline

Rendering quality with a common FastGS pipeline

* qiHD Q4

With FastGS fixed, FAR has higher PSNR and SSIM and lower LPIPS than MAGICIAN.

Table R3. 3DGS rendering quality comparison.
Method3DGS trainingInitial geometryPSNR ↑SSIM ↑LPIPS ↓
MAGICIANFastGSGT depth13.140.3970.447
FAR (original paper)referenceSPaGSPredicted depth15.940.4520.429
FAR (FastGS)FastGSPredicted depth15.290.4340.443

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

FastGS rows are compared; original SPaGS is reference only. Same ten ERP test positions per scene and matched FastGS training configuration. MAGICIAN: GT geometry; FAR: predicted geometry.

zcV1Training and pose inputs

Training data and SRA ablations

* zcV1 Q2

With the same backbone and training data, SRA reduces AbsRel from 0.1914 to 0.1196.

Table R6. Training data and SRA comparisons.
ModelAdditional training dataAbsRel ↓RMSE (m) ↓δ₁.₂₅ ↑
PanoVGGTNone0.166727.180.7684
PanoVGGTOurs (FAR26 + Holo360D)0.127424.360.8473
VGGT-Ω w/o SRAOurs (FAR26 + Holo360D)0.191431.460.6821
VGGT-Ω + SRA (ours)Ours (FAR26 + Holo360D)0.119623.290.8647

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

All rows use the same alignment. PanoVGGT’s additional training uses the same data, steps and mixture as FAR; w/o SRA updates only the depth head. Macarons++ evaluation scenes are absent from training.

Comparison with MapAnything

* zcV1 Q3

Within each alignment group, FAR outperforms MapAnything on all three reported metrics.

Table R7. Comparison with MapAnything.
MethodUse of posesAbsRel ↓δ₁.₂₅ ↑RMSE (m) ↓
MapAnything (cubemap)Within the network0.35550.348155.8
Ours w/o alignmentWithin the network0.20730.587237.72
MapAnything (cubemap) + alignmentWithin the network and for alignment0.21890.638536.54
Ours (SRA + alignment)Within the network and for alignment0.11960.864723.29

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

Without alignment, FAR outperforms MapAnything on all test scenes. Both receive the same metric poses. Compare within each alignment group; MapAnything uses cubemap to ERP adaptation.

eJ88Robustness, detail and disclosure

Component adaptations

* eJ88 Q1

FAR adapts ray encoding, persistent mapping and planning to panoramic active reconstruction.

Component adaptations
Prior componentFAR adaptation
UCPEERP seam handling, pole frame fallback, metric camera offsets.
AMB3RA global voxel map that persists across windows.
FUELRouting retained; Adaptive Initialization and peak ratio termination added.

Reconstruction evidence ↑ · Planning evidence ↑

Parameter sensitivity

* eJ88 Q2

Across all 25 settings, coverage changes range from −4.4 to +1.6 pp; the minimum is 81.2%.

Table R10. Coverage changes relative to the default setting (pp).
α / default αη = 0.03η = 0.05η = 0.07η = 0.10η = 0.15
×0.5−1.0−1.2−1.4−1.5−1.9
×0.7+1.6+1.6+0.9+0.70.0
×1+0.5+0.30 (default)−0.3−1.5
×1.4−1.6−1.8−2.0−2.3−2.9
×2−3.6−3.6−3.7−3.8−4.4

At η = 0.07, finer voxel settings increase path length by 33–54%.

Table R11. Path length relative to the default setting.
α / default αη = 0.03η = 0.05η = 0.07η = 0.10η = 0.15
×0.51.761.631.541.411.25
×0.71.511.441.331.251.14
×11.191.091.00 (default)0.940.87
×1.40.880.860.830.790.76
×20.670.670.660.650.63
Select a cell to compare the same setting in both tables.

GT depth, 11 scenes, multiple start positions per scene. Default: α = 0.0273, η = 0.07, coverage 85.6%. Coverage uses ε = 1.0 m; path values are ratios to the default. Shading shows values; the outline marks the default.

Local detail at matched flight budgets

* eJ88 Q3

Adaptive voxels achieve the highest detail AUC in both scenes under the same path budget.

Table R12. Detail coverage and path length at different voxel resolutions.
SceneVoxel size (m)Coverage at ε = 0.5 mAUC at ε = 0.5 m under the same path budgetPath length (m)
Colosseum2.2564.6%0.5122477 (+53%)
Colosseum3.1563.3%0.5182103 (+30%)
Colosseum4.5 (adaptive)62.0%0.5331615
Dunnottar2.2298.8%0.7421883 (+34%)
Dunnottar3.1198.0%0.7831470 (+5%)
Dunnottar4.45 (adaptive)98.4%0.8131401

Best results within each comparable group are shown in bold black text with an orange background. Ties are highlighted equally.

ε = 0.5 m; compare within each scene. AUC uses the same path budget.

Closed loop comparisons

* eJ88 Q4

Initialization and termination ablations are provided; the persistent versus per window closed loop comparison is not included.

Closed loop evidence coverage
ComparisonAvailable evidence
Fixed vs adaptive parametersR5/R13 and the planning curve above.
With vs without peak ratio terminationR5/R13 and the planning curve above.
Persistent vs per window refinementClosed loop coverage/path comparison not included.

Go to the planning comparison ↑

Parameter settings and task boundary

* eJ88 Q5; eJ88 Q6

All 11 scenes share the same parameters, and the task boundary closely encloses the target structure.

Reported settings and boundary
ItemReported setting
α default0.0273; selected for mapping and planning efficiency.
η default0.07; coverage saturation criterion.
Across scenesOne shared setting across all 11 scenes.
Task boundary BOfficial Macarons++ evaluation box, closely enclosing the target.

Figure 1 provenance and AI disclosure

* eJ88 AI Disclosure Concern

The central “Omnidirectional View” panel in Figure 1 combines three elements: a 3D scene generated with OpenAI’s ChatGPT Image based on one of our existing scenes, a photograph of our actual UAV, and a panoramic sphere in the upper left created from visual materials in PowerPoint. This composition illustrates omnidirectional acquisition, with the generated scene designed to enhance the sense of depth and convey this concept more clearly. All other result images in the paper were produced by our experiments. We apologize for omitting this use of image generation from the AI Use Statement; we will revise the statement in the main manuscript and update the Figure 1 caption to disclose the generated scene and explain the composition.

Planning ablation
Enlarged original planning ablation figure