Title :
Pattern extraction of sonar images using an LPC edge detector with entropy thresholding
Author :
Zhang, James Z. ; Punch, Andrew J.
Author_Institution :
Dept. of Eng. & Technol., Western Carolina Univ., Cullowhee, NC, USA
Abstract :
In this paper, we present an object pattern extraction method using a recently proposed edge detection approach based on Linear Prediction (LP), with specific applications to pattern extraction of underwater sonar images. The novel edge detection method is illustrated, an entropy method used to calculate the threshold that separates the object and the background is discussed, and the application of this edge detection method in feature extraction is presented. Experimental results show that this method is effective and in many cases, outperforms some well known edge detection methods such as Canny and Sobel edge detectors.
Keywords :
edge detection; entropy; feature extraction; image segmentation; sonar imaging; LPC; edge detection; entropy thresholding method; feature extraction; linear prediction; object pattern extraction; underwater sonar image; Detectors; Entropy; Feature extraction; Image edge detection; Noise; Shape; Sonar; edge detection; entropy thresholding; feature extraction; linear prediction; object detection; pattern recognition;
Conference_Titel :
OCEANS, 2012 - Yeosu
Conference_Location :
Yeosu
Print_ISBN :
978-1-4577-2089-5
DOI :
10.1109/OCEANS-Yeosu.2012.6263408