Class NonMaxBlock

java.lang.Object
boofcv.alg.feature.detect.extract.NonMaxBlock
Direct Known Subclasses:
NonMaxBlock_MT

public class NonMaxBlock extends Object

Non-maximum extractor based on the block algorithm in [1]. The worst case complexity per pixel is 4 - 4/(n+1) where n is the region size. The algorithm works by breaking up the image into a set of evenly spaced blocks with their sides touching. The local maximum is found inside a block and then the region around that maximum is examined to see if it is truly a local max.

Each block check is independent of all the others and no information is exchanged. This algorithm could be paralyzed easily and has no memory overhead.

See NonMaxSuppression for a definition of parameters not described in this document

[1] Neubeck, A. and Van Gool, L. "Efficient non-maximum suppression" ICPR 2006

  • Field Details

    • configuration

      protected final boofcv.alg.feature.detect.extract.NonMaxBlock.Configuration configuration
    • border

      protected int border
  • Constructor Details

  • Method Details

    • process

      public void process(GrayF32 intensityImage, @Nullable @Nullable QueueCorner localMin, @Nullable @Nullable QueueCorner localMax)
      Detects local minimums and/or maximums in the provided intensity image.
      Parameters:
      intensityImage - (Input) Feature intensity image.
      localMin - (Output) storage for found local minimums.
      localMax - (Output) storage for found local maximums.
    • getSearch

      public NonMaxBlock.Search getSearch()
    • setSearchRadius

      public void setSearchRadius(int radius)
    • getBorder

      public int getBorder()
    • setBorder

      public void setBorder(int border)
    • getSearchRadius

      public int getSearchRadius()
    • getThresholdMin

      public float getThresholdMin()
    • setThresholdMin

      public void setThresholdMin(float thresholdMin)
    • getThresholdMax

      public float getThresholdMax()
    • setThresholdMax

      public void setThresholdMax(float thresholdMax)