Side-scan sonar image processing using thin plate splines and control point matching (Q1315190)

The authors are primarily interested in small scale changes that occur in a given underwater scene observed at different times. They describe the use of thin plate splines in registering pairs of digital images composed of millions of pixels. The development of reliable software tools to aid human interpretation of complex digital images is of increasing importance. In particular, change detection requires that two or more images be carefully overlaid (registered) using digital techniques that compensate for picture distortions caused by sensor motion, ambient conditions, etc. They also describe on selecting and matching the control points needed for image registration. One example is also provided to demonstrate the algorithm.