|
orca-robotics
INTRODUCTION
Overview
Download and Install
Documentation
REPOSITORY
Interfaces
Drivers
Libraries
Utilities
Software Map
DEVELOPER
Dashboard
PEOPLE
Contributors
Users
Project
Download
Mailing lists
|
|
|
An ACFR implementation of a laser feature extractor.
More...
The configuration documentation is split into sections for each feature type
- General
- The uncertainty of point features
PointTargetRangeSd (float) [m]
PointTargetBearingSd (float) [deg]
- Default: 5.0
- Note these values need to take into account:
- Inherent sensor uncertainty
- Sensor offset uncertainty
- Timing uncertainty
- Reflectors
Reflectors.MaxDeltaRangeNearReflector (float) [m]
- Don't extract a feature if there's a jump in the ranges > this threshold near the reflector. This rejects spurious reflectors that are partially-obscured behind a foreground obstacle. It also reduces the angles about which a reflector on a corner can be extracted.
- Default: 0.5
Reflectors.MaxDeltaRangeWithinReflector (float) [m]
- A reflector that includes a jump in range > this threshold will be ignored. Default: 0.3
Reflectors.MinBrightness (int)
- minimum intensity value to be considered a feature
- Default: 1
Reflectors.PFalsePositive (float)
- porbability of a false positive
- Default: 0.1
Reflectors.PTruePositive (float)
- probability of a true positive
- Default: 0.6
- Foreground Points
ExtractForegroundPoints (bool)
FGPoints.MinForegroundWidth (float) [m]
- minimum width of poles/trees
- Default: 0.01
FGPoints.MaxForegroundWidth (float) [m]
- maximum width of poles and trees
- Default: 0.5
FGPoints.MinForegroundBackgroundSeparation (float) [m]
- how far the pole or tree must be from the background or other clutter
- Default: 2.0
FGPoints.PFalsePositive (double)
- Probability of detecting a false-positive foreground point
- Default: 0.4
FGPoints.PTruePositive (double)
- Probability of detecting a true-positive foreground point
- Default: 0.5
- Lines
Lines.MaxDeltaRangeInCluster (float) [m]
- Points are first clustered before fitting lines.
- A change in range greater than this breaks a cluster.
- Default: 0.5
Lines.MaxDeltaInLine (float) [m]
- For a line to be fit, the orthogonal distance of each point from the line must be less than this.
- Default: 0.1
Lines.MinPointsInLine (int)
- Self-explanatory
- Default: 6
Lines.MinLineLengthForLineFeature (float) [m]
- Self-explanatory
- Default: 1.0
Lines.MinLineLengthForVisibleEndpoints (float) [m]
- Self-explanatory
- Default: 1.0
Lines.PFalsePositive (double)
- Probability of detecting a false-positive line
- Default: 0.4
Lines.PTruePositive (double)
- Probability of detecting a line when it's really there
- Default: 0.6
- Corners
ExtractCorners (bool)
PFalsePositive (double)
- Probability of detecting a false-positive corner
- Default: 0.2
PTruePositive (double)
- Probability of detecting a true-positive corner
- Default: 0.6
- The corner extractor looks for intersecting lines. It relies on the line extractor, and is therefore affected by the line extractor's configuration (with the exception of the MinLineLength parameter: the corner detector will look for the intersection of smaller lines than the line detector is configured for).
- Doors
ExtractDoors (bool)
Doors.MinDoorWidth (double)
Doors.MaxDoorWidth (double)
Doors.MinDeltaRangeThruDoor (double)
Doors.RangeSd (double)
Doors.BearingSd (double)
Doors.MinReturnsThruDoor (int)
Doors.MinReturnsBesideDoor (int)
Doors.MaxDeltaRangeWallBesideDoor (double)
- Authors
- Alex Brooks
|
|