Feature Matching and Assessment of Similarity rate on geometrically Distorted Side Scan Sonar Images

This paper presents the geometric transformation on different shades of Side Scan Sonar (SSS) images and application of different feature extraction and matching algorithms for submerged (debris or salvage) object contained images in the underwater environment. These SSS images can be classified into semi shades (Sh1), high shades (sh2), low shades (Sh3) and no shades (Sh4) by depending on the shadows in the images. During SSS acquisition, geometric distortion can occur due to the variations in the trajectory, speed, or orientation of the tow-fish. So, We have applied the geometric transformation such as rotation, translation and scaling transformation on each shade of the SSS images and feature extraction is done using Scale Invariant Feature Transform (SIFT), Speeded Up Robust Feature (SURF) and Oriented FAST and rotated BRIEF (ORB) and then matching algorithms Fast Library for Approximate Nearest Neighbour Search (FLANN) and RANdom Sample Consensus (RANSAC) is used. The evaluation of feature matcher is done Qualitatively and Quantitatively by using the matching accuracy between the images. Finally, we had concluded that which algorithm will work better for different transformations and shades of SSS images.