Interpreting quality report

Hey all,

I’m currently doing a masters thesis, using ODM to generate a 3D point cloud and compare various different vegetation structure parameters to field measurements.

Part of my research involves a quality assessment of the ODM outputs, using both a qualitative assessment and the ODM report.

I have a question about a drone flight I did yesterday. I flew the drone in perfect conditions, with a flight plan and settings to maximize quality. The outputs look great, except for one thing, the ODM quality report states that “Reconstructed Points (Sparse) 647799 over 4818886 points (13.4%)”

Unless I am confused, that seems very low? Even though in all other aspects the quality of the outputs seem quite good.

Can anyone help me make sense of this? See dropbox link to the report below

1 Like

What were your full processing parameters?

Seems low. This is often 80% or above. It’s true, the report otherwise looks fantastic.

1 Like

Here you go:

auto_boundary false
boundary {}
build_overviews true
camera_lens auto
cameras {}
cog true
copy_to null
crop 3
debug false
dem_decimation 1
dem_euclidean_map false
dem_gapfill_steps 3
dem_resolution 2
depthmap_resolution 640
dsm true
dtm true
end_with odm_postprocess
fast_orthophoto false
feature_quality high
feature_type sift
force_gps false
gcp null
geo null
gps_accuracy 10
ignore_gsd false
matcher_distance 0
matcher_neighbors 8
matcher_type flann
max_concurrency 24
merge all
mesh_octree_depth 11
mesh_size 200000
min_num_features 8000
name ef108e8e-ea3b-436c-bc86-05d736a648ba
optimize_disk_space true
orthophoto_compression DEFLATE
orthophoto_cutline false
orthophoto_kmz false
orthophoto_no_tiled false
orthophoto_png false
orthophoto_resolution 2
pc_classify true
pc_csv false
pc_ept true
pc_filter 2.5
pc_geometric false
pc_las true
pc_quality high
pc_rectify false
pc_sample 0.01
pc_tile false
primary_band auto
project_path /var/www/data/ef108e8e-ea3b-436c-bc86-05d736a648ba
radiometric_calibration none
rerun null
rerun_all false
rerun_from null
resize_to 2048
skip_3dmodel false
skip_band_alignment false
skip_report false
sm_cluster null
smrf_scalar 1.25
smrf_slope 0.15
smrf_threshold 0.5
smrf_window 18
split 999999
split_image_groups null
split_overlap 150
texturing_data_term gmi
texturing_keep_unseen_faces false
texturing_outlier_removal_type gauss_clamping
texturing_skip_global_seam_leveling false
texturing_skip_local_seam_leveling false
texturing_tone_mapping none
tiles false
time false
use_3dmesh false
use_exif false
use_fixed_camera_params false
use_hybrid_bundle_adjustment false
verbose true

One thing I just noticed, the ODM report said it used 632 of 632 images. I only took 316 photos from the drone however. Did I somehow copy the image set over twice? That might explain why there are a large amount of points not used in the reconstruction…

1 Like

Ha! Yeah, that might do it.

1 Like

Processed again with 316 images…

Reconstructed Points (Sparse) 525627 over 534591 points (98.3%)

User error strikes again!

1 Like

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.