Computer vision has become a popular method to explore the underwater environment. Underwater cameras are used for scientific exploration seafloor structure, monitoring undersea habitats, and inspecting industrial equipment. Camera calibration is required for recovering 3D geometry of a scene using 2D images. Despite the remarkable success for land-based camera systems, underwater camera calibration problems such as image distortion and error have not been addressed until recently. Calibration of underwater cameras systems remains a challenging problem in the computer vision area.
Researchers at the University of Alberta have developed a new method to calibrate and perform underwater 3D image reconstruction using a single camera by exploiting triple wavelength dispersion. This technology provides a closed-form solution to the interface distance and can estimate camera parameters more accurately than any other existing method reported. This is the first technology of its kind to use triple wavelength dispersion and does not require a calibration object. The use of a single undersea camera instead of a stereo rig with two cameras is cheaper to deploy and encounters fewer calibration hurdles. This
method is easy to use and can accurately be applied to reconstruct the 3D geometry of an underwater scene.
Figure 1. Reconstructed 3D model of Folger Passage underwater seafloor observations.
- Accurate, easy to use, cost-competitive
- Single-camera, no calibration object
- Unique, triple wavelength dispersion
The invention represents a valuable opportunity for use in the computer vision industry.
Technology Management Group
TEC Edmonton – University of Alberta