3-D object detection based on line sets matching

3-D Object detection plays an important role in automation industry. It is a big challenge to detect 3-D object in images without knowing its pose because under this condition, the appropriate 2-D template cannot be obtained. In this paper, we propose a 3-D object detection method based on line sets matching. The algorithm uses line segments to represent the 3-D model, and applies projection transform to get several 2-D model sets. Then, it extracts line segments from the image by LSD. Hence, the 3-D object detection task is converted to line sets matching problem. Finally, the algorithm uses ISPD as similarity measurement and finds the best matching between data set and model sets by a restricted steepest-descent local matching method. Experimental results show that the proposed method can detect 3-D object in images and obtain its pose simultaneously.