FalconSmall

Marker Types + Algorithms

Overview of Markers, their characteristics and scope

MXT

mxt1.jpg Marker type: 5-/6-spot markers (FhG-IOSB)
Definition: Marker diameter / spot diagonal = 1.6
Single spot size = spot diagonal / 4
Spot angle = 90 deg (MXT-5) or 72 deg (MXT-6)
Minimum size approx. 12 pixel
 
Set-up: Automatic detection in search area
 
Tracking: Marker based MarkerXtrackT algorithm (FhG-IOSB)
Model with affined transformation: incl. Rotation and zoom.
Unambiguous detection of close neighboring markers possible applying different relative orientations
 
Feature: „Crash proven“: Extremely robust with regard to variations of illumination and geometrical changes (rotation, shearing).
Additionally to the position you gain the measurement value MXT angle. Measurement accuracy < 0.2 pixel
mxt2.jpg mxt3.jpg
Spots as result of
 preprocessing
Detected MXT-Marker
Data Sheet: FalCon etra MovXact MXT Marker

DOT

dot1.jpg Marker type: 1-spot marker
 
Definition: Ratio marker diameter : inner diameter = 1:1
Minimum size ca. 7 pixel
 
Set-up: Automatic centering in search area
 
Tracking: Marker based DOT algorithm
Model = light circular area with high contrast to the dark background
 
Feature: Fast automatic measuring of quite small markers
Measurement accuracy < 0.2 pixel

QUAD

quad1.jpg Marker type: Quadrant marker (4 segments)
 
Definition: High contrast between the segments
(attention using yellow-black markers!)
Recommended size > 15 pixel
 
Set-up: Automatic centering in search area
 
Tracking: Marker based QUAD algorithm
Model = circular symmetrical pattern
 
Feature: Measurement accuracy > 0.2 pixel.
Tip: avoid practical use for image analysis!

CODE

code1.jpg Marker type: 1-spot marker with code ring (AICON)
 
Definition: Center marker like DOT. Additional ring with width = DOT inner diameter, Code = number of marker
Recommended size > 15 pixel
 
Set-up: Automatic centering and code detection with free search in image
 
Tracking: CODE algorithm similar to DOT, additionally recognition of the correct code
 
Feature: Apply as control points on test fields or in tests with 3D analysis

COR

cor1.jpg Marker type: (structured) image section (= template)
 
Definition: Recommended size > 20 pixel
 
Set-up: Interactive definition of center and size
 
Tracking: COR algorithm (= correlation/matching of image templates)
Standard model using non-adaptive templates (static from set-up image) and translatory displacement (no rotation)
 
Feature: Flexible measuring of not specially marked points.
Recommended only in case of constant orientation

PIX

pix1.jpg Marker type: One single point
 
Definition: Only coordinates of m image element
 
Tracking: Interactive input of the position by cursor (with sub pixel accuracy)
 
Set-up: Automatic motion prediction and interactive input of the positions
 
Feature: Manual measuring of scales/rulers or not signalized points.
Measurement accuracy > 1/zoom magnification pixel

Note