Inheritance diagram for prapi::GlobalPeakDetector< T >:
The detection method is as follows. First, the input image is globally thresholded to produce a binary image (hence the name). Second, a binary opening operation is performed to reduce noise. Third, connected components are found. Finally, the found components are sorted according to their magnitude (the magnitude of the input image at the center of the mass of the component) and returned. The user can decide how many peaks are returned at most.
|GlobalPeakDetector (T threshold=0, int maxPeaks=10)|
|Create a PeakDetector that binarizes input images using the given threshold. |
|void||setOpenMaskRadius (unsigned int radius)|
|Set the radius of the circular structuring element used to reduce noise. |
|unsigned int||getOpenMaskRadius () const|
|Get the radius of the structuring element used to reduce noise.|
|void||setAutoThreshold (bool autoThreshold)|
|Set automatic thresholding on. |
|void||getAutoThreshold () const|
|See if the automatic thresholding is on.|
|void||setThreshold (T threshold)|
|Set the threshold.|
|T||getThreshold () const|
|Get the binarization threshold.|
|void||setMaxPeaks (int peaks)|
|Set the maximum number of peaks to return.|
|int||getMaxPeaks () const|
|Get the maximum number of peaks.|
|util::List< util::Pair< T,|
binary::Component > >
|detectPeaks (const util::Matrix< T > &mat) const throw (ImageTransformException&)|
|Detect peaks in an image or transform domain. |
Create a PeakDetector that binarizes input images using the given threshold.
The maximum number of peaks returned is given by the maxPeaks parameter, which defaults to ten.
Detect peaks in an image or transform domain.
The result is a list of Pair objects that store the magnitude of the peak, and the properties of the binary connected component that forms the peak.
Set automatic thresholding on.
If automatic thresholding is on, the input image is binarized with the mean of the image plus two times standard deviation.
Set the radius of the circular structuring element used to reduce noise.
In noise-free cases, this value can be set to zero, disabling the morphological opening. The size of the structuring element will be (radius*2+1)-by-(radius*2+1). The default value is one.