Accessibility statement

Advanced Signal Processing Techniques for Underwater Acoustic Imaging

There is growing interest and activity in using underwater vehicles for sea exploration. Accurate detection, identification and localisation of different features in an underwater environment is a key challenge that researchers are focusing on. In particular, exploration of natural aquatic environments and maintenance of artificial structures requires efficient imaging techniques, such as three-dimensional (3D) reconstruction of an underwater scene from the data provided by a sonar imaging system. This kind of representation could be applied in specific functions like navigation, exploration for submarine resources, marine science, observation, maintenance and repair of undersea structures, measurement of surface area and other morphological characteristics of corals and similar irregular objects, detecting, classifying, and tracking underwater targets for modern underwater defence systems, mapping of archaeological sites and others.

Some underwater 3D reconstruction techniques from sonar image frames have recently been developed with encouraging results. With my PhD research, I aim to develop an improved method for conducting this task, applying novel computational techniques such as machine learning to object characteristic recognition from the image frames. Creation of our own data set of images is one of that activities being carried out. For this, a sonar image simulator has been developed. In addition, data sets will be acquired from a real sonar imaging system.

Members

  • Jose Almanza Medina
  • Yuriy Zakharov
  • Ben Henson

Funding

Mexican Government

Dates

Start: October 2018 - Sep 2021

Research