#include <DLCannyEdgeDetect.h>
Public Member Functions | |
DLCannyEdgeDetect () | |
DLCannyEdgeDetect (const DLImage &image, DL_BYTE lowT=1, DL_BYTE highT=255) | |
virtual | ~DLCannyEdgeDetect () |
DLImage | dlGetMagnitudeImage () |
DLImage | dlGetOrientationImage () |
Two threshold parameters highT and lowT (i.e. highT > lowT) in the range of [0, 255] control the quality of computed edges as they are used globally at the edge tracking step of the Canny edge detector. Edge tracking starts at a point that has edge magnitude higher than highT and then continues outwards from that starting point until it reaches a point that has an edge magnitude lower than lowT. This hysteresis reduces the chance that noisy edges are broken up into multiple edge fragments.
Users can obtain these information from its member functions dlGetMagnitudeImage() and dlGetOrientationImage(), respectively. Both the edge magnitude and orientation are encoded within the intensity range of grayscale image, i.e. [0, 255].
Reference:
J. Canny, "A Computational Approach to Edge Detection," IEEE Trans. Pattern Analysis and Machine Intelligence, vol. 8, no. 6, pp. 679--698, 1986.
Definition at line 43 of file DLCannyEdgeDetect.h.
DLCannyEdgeDetect::DLCannyEdgeDetect | ( | ) |
Default DLCannyEdgeDetect Constructor
DLCannyEdgeDetect::DLCannyEdgeDetect | ( | const DLImage & | image, | |
DL_BYTE | lowT = 1 , |
|||
DL_BYTE | highT = 255 | |||
) |
DLCannyEdgeDetect Constructor
image | pointer to the source DLImage | |
lowT | lower threshold used in edge tracking, setting lowT too high causes noisy edges to break up | |
highT | higher threshold used in edge tracking, setting highT too low increases the number of spurious and undesirable edge fragments |
virtual DLCannyEdgeDetect::~DLCannyEdgeDetect | ( | ) | [virtual] |
Default DLCannyEdgeDetect Destructor
DLImage DLCannyEdgeDetect::dlGetMagnitudeImage | ( | ) |
Return magnitude of detected edges (edge map) as a grayscale image
DLImage DLCannyEdgeDetect::dlGetOrientationImage | ( | ) |
Return orientation of detected edges (scaled within an unsigned char 0-255) as a grayscale image