|
|
 |
| |
|
|
|
Rolling Ball Detection for Signs in Aircraft Measurement |
|
|
|
PP: 925-933 |
|
|
|
Author(s) |
|
|
|
Wusheng Luo,
Jingjing Xiao,
Qin Lu,
Pei Li,
|
|
|
|
Abstract |
|
|
| : The key problem of stereo vision in measurement is how to find the anchor signs of objects for matching. In this paper, we firstly propose the method of signs designing to describe the basic structures of aircraft. A novel algorithm called rolling ball is proposed to detect them. The method opens some detecting windows. Through computing the distance between weight center and the geometrical center of each window, the algorithm could capture the start pixels of the signs and find the raw moving vector. Adding the condition of original moving vector, the synthetical vector ensures that the signs track will not be affected by intersection. We could obtain more accurate detection results in terms of the restrictions of width, length and moving degrees. Both of the simulation and aircraft measurement experiments demonstrate that the method could find out the anchor signs automatically and correctly. In addition, we get some useful data which could be regarded as basic parameters when adopting the rolling ball method in stereo matching to detect the signs.
|
|
|
|
|
|
|
|
