(英) |
In this paper, a method of extracting the object region using the distance information is proposed. Pixels that difference in the distance is large are extracted as edges of the object, after obtaining the distance information. However, in part the difference in distance is smaller (for examples, left and right of the ridge line / valley line of the object), it is impossible to extract edges. Therefore, edges of the object are extracted by using the gradient of distance. A combined two edges are determined the outline of the object. In experimental results, it was confirmed the usefulness of the outline extraction method. |