Skip to content
GitLab
Explore
Sign in
Primary navigation
Search or go to…
Project
wolf
Manage
Activity
Members
Labels
Plan
Issues
Issue boards
Milestones
Code
Merge requests
Repository
Branches
Commits
Tags
Repository graph
Compare revisions
Build
Pipelines
Jobs
Pipeline schedules
Artifacts
Deploy
Releases
Container Registry
Model registry
Operate
Environments
Monitor
Incidents
Service Desk
Analyze
Value stream analytics
Contributor analytics
CI/CD analytics
Repository analytics
Model experiments
Help
Help
Support
GitLab documentation
Compare GitLab plans
Community forum
Contribute to GitLab
Provide feedback
Keyboard shortcuts
?
Snippets
Groups
Projects
Show more breadcrumbs
mobile_robotics
wolf_projects
wolf_lib
wolf
Merge requests
!233
WIP: Apriltag
Code
Review changes
Check out branch
Download
Patches
Plain diff
Merged
WIP: Apriltag
Apriltag
into
devel
Overview
0
Commits
276
Pipelines
63
Changes
2
Merged
Joan Solà Ortega
requested to merge
Apriltag
into
devel
6 years ago
Overview
0
Commits
276
Pipelines
63
Changes
2
Expand
MR open while Gitflow transition
It can stay open until big merge in the future
0
0
Merge request reports
Viewing commit
f596a379
Prev
Next
Show latest version
2 files
+
52
−
14
Inline
Compare changes
Side-by-side
Inline
Show whitespace changes
Show one file at a time
Files
2
Search (e.g. *.vue) (Ctrl+P)
f596a379
Optimize parameters for speed -> barely under 1/30 s for detection of quads on macbook 8 threads
· f596a379
Mederic Fourmy
authored
6 years ago
src/examples/processor_tracker_landmark_apriltag.yaml
+
26
−
9
Options
processor type
:
"
TRACKER
LANDMARK
APRILTAG"
processor name
:
"
tracker
landmark
apriltag
example"
detector parameters
:
quad_decimate
:
0.0
quad_sigma
:
2.0
nthreads
:
4
debug
:
false
refine_edges
:
true
refine_decode
:
true
refine_pose
:
true
ippe_min_ratio
:
3.0
# quite arbitrary, always > 1 (to deactive, set at 0 for example)
quad_decimate
:
1.5
# doing quad detection at lower resolution to speed things up (see end of file)
quad_sigma
:
0.8
# gaussian blur good for noisy images, may be recommended with quad_decimate. Kernel size adapted (see end of this file)
nthreads
:
8
# how many thread during tag detection (does not change much?)
debug
:
false
# write some debugging images
refine_edges
:
true
# better edge detection if quad_decimate > 1 (quite inexpensive, almost no diff)
refine_decode
:
false
# improve detection probability for small tags (quite expensive (~*3)
refine_pose
:
false
# improves detection by maximizing local contrast so that future pose extraction works better (VERY expensive ~*10-20)
ippe_min_ratio
:
3.0
# quite arbitrary, always > 1 (to deactive, set at 0 for example)
ippe_max_rep_error
:
2.0
# to deactivate, set at something big (100)
tag parameters
:
@@ -33,3 +32,21 @@ vote:
voting active
:
true
min_time_vote
:
0.2
# s
min_features_for_keyframe
:
1
# Annexes:
### Quad decimate: the higher, the lower the resolution
# Does nothing if <= 1.0
# Only values taken into account:
# 1.5, 2, 3, 4
# 1.5 -> ~*2 speed up
# Higher values use a "bad" code according to commentaries in the library, smaller do nothing
### Gaussian blur table:
# max sigma kernel size
# 0.499 1 (disabled)
# 0.999 3
# 1.499 5
# 1.999 7
Loading