Object Identification Using Soft Computing in SONAR images

Objective of this analysis is to enhance the quality of a sonar image in order to get a better interpretation than using a raw (unrefined) sonar image. The overall analysis is accomplished by various module such as identifying the type of noise present in the sonar image, for finding suitable filter to reduce the image noise, applying standard enhancement techniques, implementing efficient edge detection and segmentation techniques to split image objects from its background, discovering whether the changes occurred in the chosen seabed area over the period of time and recognizing object using its regional properties. The performance of the resultant image is evaluated by quality metrics using neural networks. It is found that the resultant image of all the work modules are all of better quality and preserves all the essential and useful information from the source images.