#include <sparse_stereo_types.h>
| stereo::FeatureConfiguration::FeatureConfiguration |
( |
| ) |
|
|
inline |
| bool stereo::FeatureConfiguration::adaptiveDetectorParam |
| bool stereo::FeatureConfiguration::debugImage |
if set to true, the library will generate debug images during the processing of the data
| DESCRIPTOR stereo::FeatureConfiguration::descriptorType |
| DETECTOR stereo::FeatureConfiguration::detectorType |
| int stereo::FeatureConfiguration::distanceFactor |
only used if knn >=2. For two features to be considered corresponding, the next nearest neighbour needs to be (distance * distanceFactor) away from the current neighbour. A value of 1.0 is equal to having knn = 1. A value of 2.0 will make sure matches are quite unique. Usually something like 1.6 is used.
| FILTER stereo::FeatureConfiguration::filterType |
| int stereo::FeatureConfiguration::isometryFilterMaxSteps |
maximum Ransac steps the isometry filter should use
| double stereo::FeatureConfiguration::isometryFilterThreshold |
threshold error value for a point to still be considered an inlier in the isometryFilter
| int stereo::FeatureConfiguration::knn |
number of neares neighbours to check for feature correspondence. a value of 1 will just check the next neighbour. A value of 2 will check the two nearest neighbours and apply the distanceFactor criterion for filtering correspondences.
| int stereo::FeatureConfiguration::maxStereoYDeviation |
| int stereo::FeatureConfiguration::targetNumFeatures |
the target number of features from the detector
The documentation for this struct was generated from the following file:
