Research

The Underwater Vision and Imaging Laboratory (UVIL) at the ECE Department is engaged in the development of advanced computer vision and imaging technologies that transform 2-D images of 3-D scene into high-level knowledge and scene representation. The activities cover methodologies for the processing of data from both optical and 2-D forward-look imaging sonar systems, as well as the integration of visual cues from both imaging modalities.

We originally introduced the novel opti-acoustic stereo imaging paradigm whereby stereovision principles were generalized to a pair of multi-modal optical and sonar devices with overlapping fields of view. A core contribution of this novel paradigm is the adaptation of the epipolar geometry and constraints— a fundamental principle in stereo vision— that are derived from the mathematical relationships between the two different projection models [Oceans’05, PAMI’08]. This original contribution also shows that the true complementary properties of these two modalities enables each to overcome the shortcomings of the other.

One very distinct advantage is that traditional stereo systems require larger baselines to improve 3-D reconstruction accuracy, but an opti-acoustic stereo system not only works equally well with a zero baseline. In fact, this is preferred in order to overcome the fundamental matching problem at the occluding contours, leading to a novel 3-D object modeling technique that integrates the reconstructed 3-D contours from 360 stereo views into a 3-D volumetric representation [Oceans’13, CVIU’15].

Another key contribution in the integration of visual motion cues in 2-D multi-modal opti-acoustic stereo sequences, leading to another solution that bypasses the opti-acoustic correspondence problem [Oceans’08, CVIU’10], as well as the application of bundle adjustment to opti-acoustic stereo sequences [Oceans’08].

Accordion Group

Open All Tabs
  • Awards and Recognitions

    • 2nd Best Student Poster Award: Haghighat, X. Li, Z. Fang, Y. Zhang, and S. Negahdaripour (2016). Segmentation, classification and modeling of two-dimensional forward-scan sonar imagery for efficient coding and synthesis. Proc. IEEE/MTS Conf. Oceans'16, Monterey, CA, September.
    • 3rd Best Student Poster Award: Babaee, and S. Negahdaripour (2015). Improved range estimation and underwater image enhancement under turbidity by opti-acoustic stereo imaging. Proc. IEEE Oceans'15 Conference, Genova, Italy, May.
    • Best Paper Award: Negahdaripour, H. Sekkati, H. Pirsiavash (2007). Opti-acoustic stereo imaging, system calibration and 3-D reconstruction. Proc. IEEE Int. Workshop on Beyond Multiview Geometry: Robust Estimation and Organization of Shapes from Multiple Cues, held in conjunction with CVPR'07, Minneapolis, MN, June.
    • Siemens Corp. Best Paper Award: Firoozfam, S. Negahdaripour (2003). Multi-camera conical imaging; calibration and robust 3-D motion system for robust 3D motion estimation, positioning and mapping from UAVs. Proc. IEEE Int. Conference Advanced Video and Signal Based Surveillance, Miami, FL, July.
    • Finalist for Best Student Paper: Xu, S. Negahdaripour (2001). Application of extended covariance intersection principle for mosaic-based optical positioning and navigation of vehicles. IEEE Conference on Robotics and Automation, Seoul, South Korea, May.

Top