NVIDIA DriveOS Linux NSR SDK API Reference

7.0.3.0 Release
dwFeature2DTrackerConfig Struct Reference

Detailed Description

Holds configuration parameters for a feature tracker.

Definition at line 94 of file FeatureTracker.h.

Data Fields

dwFeature2DTrackerAlgorithm algorithm
 Tracking Algorithm defined by dwFeature2DTrackerAlgorithm. More...
 
dwFeature2DDetectorType detectorType
 Type of detector that connects to the tracker, should be the same value as dwFeature2DDetectorConfig::type during detector initialization. More...
 
uint32_t maxFeatureCount
 Upper bound on number of features handled. More...
 
uint32_t historyCapacity
 Upper bound of history in feature history array. More...
 
uint32_t pyramidLevelCount
 Levels of pyramid to track Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST. More...
 
uint32_t imageWidth
 Width of the images that the tracker runs on. More...
 
uint32_t imageHeight
 Height of the images that the tracker runs on. More...
 
uint32_t windowSizeLK
 Window size used in the KLT tracker. More...
 
bool enableAdaptiveWindowSizeLK
 Enable adaptive window size If enabled, the tracker will use windowSizeLK to track only at the first and the last pyramid levels, and use smaller window size at the rest pyramid levels. More...
 
uint32_t numIterTranslationOnly
 Upper bound on number of 2-DOF translation-only KLT iterations per level. More...
 
uint32_t numIterScaling
 Upper bound on number of 3-DOF translation+scaling KLT iterations per level. More...
 
uint32_t numLevelTranslationOnly
 Number of levels in pyramid that will use translation-only KLT tracking, level [maxPyramidLevel - 1, maxPyramidLevel - numTranslationOnlyLevel] will use translation-only track level [maxPyramidLevel - numTranslationOnlyLevel - 1, 0] will use translation-scaling track if numTranslationOnlyLevel = 0, translation-scaling track will be applied to all levels in pyramid. More...
 
bool useHalf
 Set it to true to use half float as intermediate results during tracking It saves register usage and will be faster than full 32-bit float, but will lose a bit of precision. More...
 
uint32_t enableSparseOutput
 Enable sparse output if set to 1. More...
 
float32_t maxScaleChange
 The maximum allowed scale change for the tracked points across consecutive frames. More...
 
float32_t nccKillThreshold
 Features will be killed if the Cross Correlation Score is less than this threshold during tracking. More...
 
float32_t nccUpdateThreshold
 Feature template will be updated if the Cross Correlation Score is less than this threshold during tracking. More...
 
float32_t largeMotionKillRatio
 Features will be killed if the motion is larger than the template size times the large motion killing threshold during tracking. More...
 
float32_t displacementThreshold
 If difference of translation prediction between 2 adjacent KLT iteration is less than this value, it is thought the result is converged and will abort tracking iteration. More...
 
dwProcessorType processorType
 Processor type which determines on which processor the algorithm should be executed on. More...
 

Field Documentation

◆ algorithm

dwFeature2DTrackerAlgorithm dwFeature2DTrackerConfig::algorithm

Tracking Algorithm defined by dwFeature2DTrackerAlgorithm.

Definition at line 97 of file FeatureTracker.h.

◆ detectorType

dwFeature2DDetectorType dwFeature2DTrackerConfig::detectorType

Type of detector that connects to the tracker, should be the same value as dwFeature2DDetectorConfig::type during detector initialization.

Definition at line 101 of file FeatureTracker.h.

◆ displacementThreshold

float32_t dwFeature2DTrackerConfig::displacementThreshold

If difference of translation prediction between 2 adjacent KLT iteration is less than this value, it is thought the result is converged and will abort tracking iteration.

Default value = 0.1F, larger value causes faster convergence, but will lose the precision.

Definition at line 211 of file FeatureTracker.h.

◆ enableAdaptiveWindowSizeLK

bool dwFeature2DTrackerConfig::enableAdaptiveWindowSizeLK

Enable adaptive window size If enabled, the tracker will use windowSizeLK to track only at the first and the last pyramid levels, and use smaller window size at the rest pyramid levels.

Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST

Definition at line 138 of file FeatureTracker.h.

◆ enableSparseOutput

uint32_t dwFeature2DTrackerConfig::enableSparseOutput

Enable sparse output if set to 1.

If enabled, the output feature array will contain invalid features, which can be removed by explicitly calling dwFeature2DTracker_compact

Definition at line 177 of file FeatureTracker.h.

◆ historyCapacity

uint32_t dwFeature2DTrackerConfig::historyCapacity

Upper bound of history in feature history array.

Definition at line 107 of file FeatureTracker.h.

◆ imageHeight

uint32_t dwFeature2DTrackerConfig::imageHeight

Height of the images that the tracker runs on.

Definition at line 119 of file FeatureTracker.h.

◆ imageWidth

uint32_t dwFeature2DTrackerConfig::imageWidth

Width of the images that the tracker runs on.

Definition at line 116 of file FeatureTracker.h.

◆ largeMotionKillRatio

float32_t dwFeature2DTrackerConfig::largeMotionKillRatio

Features will be killed if the motion is larger than the template size times the large motion killing threshold during tracking.

Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST

Definition at line 204 of file FeatureTracker.h.

◆ maxFeatureCount

uint32_t dwFeature2DTrackerConfig::maxFeatureCount

Upper bound on number of features handled.

Definition at line 104 of file FeatureTracker.h.

◆ maxScaleChange

float32_t dwFeature2DTrackerConfig::maxScaleChange

The maximum allowed scale change for the tracked points across consecutive frames.

Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST

Definition at line 183 of file FeatureTracker.h.

◆ nccKillThreshold

float32_t dwFeature2DTrackerConfig::nccKillThreshold

Features will be killed if the Cross Correlation Score is less than this threshold during tracking.

The value should be between -1.0 and 1.0. Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST

Definition at line 190 of file FeatureTracker.h.

◆ nccUpdateThreshold

float32_t dwFeature2DTrackerConfig::nccUpdateThreshold

Feature template will be updated if the Cross Correlation Score is less than this threshold during tracking.

The value should be between -1.0 and 1.0. The value should be no less than nccKillThreshold. Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX

Definition at line 197 of file FeatureTracker.h.

◆ numIterScaling

uint32_t dwFeature2DTrackerConfig::numIterScaling

Upper bound on number of 3-DOF translation+scaling KLT iterations per level.

Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST

Definition at line 150 of file FeatureTracker.h.

◆ numIterTranslationOnly

uint32_t dwFeature2DTrackerConfig::numIterTranslationOnly

Upper bound on number of 2-DOF translation-only KLT iterations per level.

If set to zero the default value will be used. More iterations helps improve tracking results but cost more time

Definition at line 144 of file FeatureTracker.h.

◆ numLevelTranslationOnly

uint32_t dwFeature2DTrackerConfig::numLevelTranslationOnly

Number of levels in pyramid that will use translation-only KLT tracking, level [maxPyramidLevel - 1, maxPyramidLevel - numTranslationOnlyLevel] will use translation-only track level [maxPyramidLevel - numTranslationOnlyLevel - 1, 0] will use translation-scaling track if numTranslationOnlyLevel = 0, translation-scaling track will be applied to all levels in pyramid.

Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST

Definition at line 160 of file FeatureTracker.h.

◆ processorType

dwProcessorType dwFeature2DTrackerConfig::processorType

Processor type which determines on which processor the algorithm should be executed on.

Supported options are: DW_PROCESSOR_TYPE_GPU, DW_PROCESSOR_TYPE_PVA_0, DW_PROCESSOR_TYPE_PVA_1.

Note
DW_PROCESSOR_TYPE_GPU supports DW_FEATURE2D_TRACKER_ALGORITHM_STD, and DW_FEATURE2D_TRACKER_ALGORITHM_EX
DW_PROCESSOR_TYPE_PVA_0/1 supports DW_FEATURE2D_TRACKER_ALGORITHM_STD, DW_FEATURE2D_TRACKER_ALGORITHM_EX

Definition at line 218 of file FeatureTracker.h.

◆ pyramidLevelCount

uint32_t dwFeature2DTrackerConfig::pyramidLevelCount

Levels of pyramid to track Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST.

Definition at line 113 of file FeatureTracker.h.

◆ useHalf

bool dwFeature2DTrackerConfig::useHalf

Set it to true to use half float as intermediate results during tracking It saves register usage and will be faster than full 32-bit float, but will lose a bit of precision.

Valid only when algorithm = DW_FEATURE2D_TRACKER_ALGORITHM_EX or DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST

Definition at line 170 of file FeatureTracker.h.

◆ windowSizeLK

uint32_t dwFeature2DTrackerConfig::windowSizeLK

Window size used in the KLT tracker.

Supported sizes are ** DW_FEATURE2D_TRACKER_ALGORITHM_STD: 6, 8, 10, 12, 14. ** DW_FEATURE2D_TRACKER_ALGORITHM_EX: 10, 12. ** DW_FEATURE2D_TRACKER_ALGORITHM_EX_FAST: 10, 12.

If set to zero the default value will be used. Larger window size provides better tracking results but costs more time.

Definition at line 129 of file FeatureTracker.h.


The documentation for this struct was generated from the following file: