Automatic Exploration Path Planning for 3D Object Observation

This paper proposes a novel method for planning exploration path in 3D object observation. First, we give a viewpoint quality metric, which combines the notion of mesh saliency and fuzzy nearness. Each viewpoint corresponds to a fuzzy set, and the quality of a viewpoint can be measured by fuzzy nearness between a normal fuzzy set and the ideal fuzzy set. Then, we compute an optimal viewpoint set. Viewpoints in the optimal viewpoint set are key points on the exploration path. We use a virtual viewpoint to replace a viewpoint set, and the fuzzy set which corresponds to the virtual viewpoint can be calculated by fuzzy comprehensive evaluation. The quality of the virtual viewpoint can be taken as an objective function and the optimal viewpoint set can be obtained by ant colony optimization algorithm. Finally, we sort the viewpoints in the optimal viewpoint set and connect them using cardinal spline curve. Consequently, the automatic exploration path is constructed. Experimental results show that our method can obtain a good exploration path which can help observer understand an object roundly and quickly.