Title :
Improved entropic edge-detection
Author :
Gomez Lopera, Juan Francisco ; Ilhami, Naima ; Escamilla, Pedro Luis Luque ; Aroza, José Martínez ; Roldán, Ramón Román
Author_Institution :
Dept. de Fisica Aplicada, Granada Univ., Spain
Abstract :
Using the Jensen-Shannon divergence of grey level histograms obtained by sliding a double window over an image, an edge detector is presented. A new technique for linking unconnected edge points, based on estimated directions, is also described. The method is suitable for textured images and has proved to be fairly robust in the presence of impulsive noise or spurious patterns. An application to images of diesel spray injection is shown, that may be used to improve combustion in diesel motors
Keywords :
edge detection; entropy; image texture; impulse noise; internal combustion engines; mechanical engineering computing; parameter estimation; statistical analysis; Jensen-Shannon divergence; diesel motor combustion; diesel spray injection; double window sliding; entropic edge detection; estimated directions; grey level histograms; impulsive noise; spurious patterns; textured images; unconnected edge point linking; Combustion; Histograms; Image edge detection; Image processing; Image segmentation; Image texture analysis; Joining processes; Noise robustness; Probability distribution; Spraying;
Conference_Titel :
Image Analysis and Processing, 1999. Proceedings. International Conference on
Conference_Location :
Venice
Print_ISBN :
0-7695-0040-4
DOI :
10.1109/ICIAP.1999.797591
