3D shape measurement of underwater objects using motion stereo

It is very difficult to measure 3D shape of underwater object from outside of the water by the use of conventional range image sensors, because the measured image is distorted by the effect of the light refraction on the water surface. In this study, we propose the new method for measuring 3D shape of underwater object by the use of motion stereo images obtained with a moving camera over the water surface. In this method, we employ an extended motion stereo method. We derive an equation of a locus of an object point in temporal-spatial space obtained by motion stereo images. The locus is parameterized by the position of the object point and the distance between the camera and the water surface. To obtain the parameters, all edge points in the images are voted into parameter space, like Hough transform. First, the parameter of the water surface distance is determined by the voting. Next, 3D position of the object point can be obtained by finding peak position in parameter space. We performed an experiment for demonstrating the efficacy of the proposed method. In the experiment, shape of an underwater polyhedron is measured by the proposed method. The experimental results show that the proposed method can provide not only 3D shape of underwater object but also the distance to the water surface.