Recognition and Tracking of Texture-free Objects for Submerged Robots in Structured Underwater Environment

Abstract

While visual tracking problem has been actively studied in computer vision discipline, recognition and tracking objects beneath the water surface still remains a challenging problem because the problem often deals with much more difficult condition compared to other land robots. In this context, we propose an efficient and accurate vision based recognition and tracking framework for texture free objects. The challenge is to segment out and to track interesting target objects in the presence of camera motion and scale changes of the objects in the underwater environment. We approached this problem with a two phased algorithm: detection phase and tracking phase. In the detection phase, we extract shape context descriptors that used for classifying objects into predetermined interesting targets. Later on, the matching result is further refined by a minimization technique. In the tracking phase, we resorted to meanshift tracking algorithm based on Bhattacharyya coefficient measurement. The proposed framework is validated with both synthetic image data set and real data set. Besides, the method is compared with a few other well-known methods, and the corresponding test results are quantitatively reported in the paper.