Class FactoryIntensityPointAlg
java.lang.Object
boofcv.factory.feature.detect.intensity.FactoryIntensityPointAlg
Factory for creating various types of interest point intensity algorithms.
- See Also:
-
Constructor Summary
Constructors -
Method Summary
Modifier and TypeMethodDescriptionstatic <T extends ImageGray<T>>
FastCornerDetector<T>Common interface for creating aFastCornerDetectorfrom different image types.static <D extends ImageGray<D>>
GradientCornerIntensity<D>Common interface for creating aHarrisCornerIntensityfrom different image types.static <D extends ImageGray<D>>
GradientCornerIntensity<D>Common interface for creating aShiTomasiCornerIntensityfrom different image types.
-
Constructor Details
-
FactoryIntensityPointAlg
public FactoryIntensityPointAlg()
-
-
Method Details
-
fast
public static <T extends ImageGray<T>> FastCornerDetector<T> fast(int pixelTol, int minCont, Class<T> imageType) Common interface for creating aFastCornerDetectorfrom different image types.- Parameters:
pixelTol- How different pixels need to be to be considered part of a corner. Image dependent. Try 20 to start.minCont- Minimum number of continue pixels in a circle for it ot be a corner. Can be 9,10,11 or 12.imageType- Type of input image it is computed form.- Returns:
- Fast corner
-
harris
public static <D extends ImageGray<D>> GradientCornerIntensity<D> harris(int windowRadius, float kappa, boolean weighted, Class<D> derivType) Common interface for creating aHarrisCornerIntensityfrom different image types.- Parameters:
windowRadius- Size of the feature it is detects,Try 2.kappa- Tuning parameter, typically a small number around 0.04weighted- Is the gradient weighted using a Gaussian distribution? Weighted is much slower than unweighted.derivType- Image derivative type it is computed from. @return Harris corner
-
shiTomasi
public static <D extends ImageGray<D>> GradientCornerIntensity<D> shiTomasi(int windowRadius, boolean weighted, Class<D> derivType) Common interface for creating aShiTomasiCornerIntensityfrom different image types.- Parameters:
windowRadius- Size of the feature it detects, Try 2.weighted- Should the it be weighted by a Gaussian kernel? Unweighted is much faster.derivType- Image derivative type it is computed from.- Returns:
- KLT corner
-