Functions

Function Reference

Common

SensorFeatureTracking.drawfeatureLine! — Function.
drawfeatureLine!(image,feature1, feature2)

Draw a line between 2 features.

source
SensorFeatureTracking.drawfeatureX! — Function.
drawfeatureX!(image,feature [,crosslength = 2])

Draw a + on a feature.

Examples

julia> map(ft->drawfeatureX!(image, ft, length=5),features)
source
SensorFeatureTracking.getApproxBestHarrisInWindow — Function.
getApproxBestHarrisInWindow(image,[n=100, windowSize = 9, k=0.04, stepguess=0.4])

Return the n best Harris features in a window

source
SensorFeatureTracking.getApproxBestShiTomasi — Function.
getApproxBestShiTomasi(image,[n=100, windowSize = 9, k=0.04, stepguess=0.4, threshold = 1e-4])

Return the n aproxamate best Shi Tomasi features in a window

source

Block Matching Feature Tracking

SensorFeatureTracking.compute_sad — Function.
compute_sad(image1, image2, off1X, off1Y, off2X, off2Y, REGION_SIZE)

Compute Sum of Absolute Differences of two regions with size REGION_SIZE.

source
SensorFeatureTracking.compute_ssd — Function.
compute_ssd(image1, image2, off1X, off1Y, off2X, off2Y, REGION_SIZE)

Compute Sum of Squared Differences of two regions with REGION_SIZE.

source
SensorFeatureTracking.compute_ncc — Function.
compute_ncc(image1, image2, off1X, off1Y, off2X, off2Y, REGION_SIZE)

Compute the normalized cross-correlation of two regions with size REGION_SIZE.

source
SensorFeatureTracking.block_tracker! — Function.
block_tracker(tracker::BlockTracker, image1, image2)

Track features between two images.

source

KLT Feature Tracking

Forward and Inverse KLT tracking with pyramids

ImageTrackerSetup
fillNewImageTemplates!
KTL_Tracker!
trackOneFeatureInversePyramid
trackOneFeatureInverse
trackOneFeaturePyramid
trackOneFeature
warping!

Camera Models and Geometry

SensorFeatureTracking.CameraModelandParameters — Type.

Data structure for a Camera model with parameters. Use CameraModel(width,height,fc,cc,skew,kc) for easy construction.

source
SensorFeatureTracking.CameraModel — Function.
CameraModel(width,height,fc,cc,skew,kc)

Constructor helper for creating a camera model.

source

Inertial Sensor Aided Feature Tracking

SensorFeatureTracking.IMU_DATA — Type.

Data structure for holding time, gyroscope, and accelerometer data.

Fields

  • utime – time [μs]

  • acc – accelerometer data

  • gyro – gyroscope data

source
SensorFeatureTracking.estimateRotationFromKeypoints — Function.
estimateRotationFromKeypoints(points_a, points_b, cameraModel, [compensate = false])

Estimate the rotation between 2 sets of Keypoints a and b using HornAbsoluteOrientation.
It is assumed that only possitive Keypoints are valid.
Set compensate to true if rotaion should be around centre off image rather than 0,0

source
SensorFeatureTracking.predictHomographyIMU! — Function.
predictHomographyIMU!(index, current_time, vector_data, CameraK, CameraK_inverse)

Estimate camera rotation from IMU data and compute predicted homography

source
SensorFeatureTracking.integrateGyroBetweenFrames! — Function.
integrateGyroBetweenFrames!(index, current_time, vector_data)

Estimate rotations from IMU data between time stamps.

source
SensorFeatureTracking.HornAbsoluteOrientation — Function.
HornAbsoluteOrientation(a::Matrix{Float64},b::Matrix{Float64})
HornAbsoluteOrientation(a::Vector{Vector{Float64}},b::Vector{Vector{Float64}})

Compute the rotation between rows of (a and b)::Array{Float64,2}.
Rotate b into the frame of a
Returns a quaternion, aQb

source

Index