Segmentation Of Sonar Images
Sonars are used for underwater sensing and the images obtained tends to be noisy making it difficult
for interpretation. This paper focuses on development of segmentation techniques to extract regions of interest
from sonar images. The noisy images are pre-processed using various filtering techniques and are enhanced
using contrast stretching methods. The gray level co-occurrence features and roundness features are extracted
from the regions and nearest neighbor classifier is used to classify them.