As we all know, the use of binocular camera can easily build indoor map, and the mobile robot itself positioning, but because the binocular camera before doing these work, the first to do the image matching, and image matching is a very time-consuming work, The current computer system has not been able to achieve rapid calculation and real-time feedback. In order to solve the problem of real-time self-localization of indoor mobile robot, this paper presents a method of indoor positioning of robot using monocular camera. In this paper, the coordinate transformation relationship between the world coordinate system, camera coordinate system, image coordinate system and computer pixel coordinate system is analyzed firstly, which is the theoretical basis of the camera calibration, then the camera is calibrated, and the internal parameter matrix is separated, and then the necessary field of view angle for the next position is calculated. By using the geometric model of the camera's keyhole imaging and combining the projection relation between the object and the camera's photosensitive plane, the coordinate transformation relation between the ground coordinate system and the camera image coordinate system is deduced in detail, and the conversion relationship between the camera image coordinate system and the computer pixel coordinate system is analyzed. and the similarities and differences between the coordinate transformation and the camera calibration; Finally, a moving object tracking method based on image processing is proposed, which fits a motion curve according to the historical locus point of moving object, and predicts the possible range of moving target in the next moment, and searches in this range, which greatly shortens the search time. , the speed of image processing is improved, and real-time feedback can be realized preliminarily.
The collection
| Author: |
Sheng Hui |
| Subject Major: |
Mechanical engineering |
| Degree awarded: |
Master |
| Degree awarding unit: |
Tianjin Polytechnic University |
| Tutor's name: |
Li Yangmin |
| Degree year: |
2016 |
| Multilingual |
Chi |
| Classification Number: |
TP242 |
| Keywords: |
Mobile robot indoor positioning moving target tracking monocular vision |
| Online Publication date: |
August 31, 2016 |
Http://d.g.wanfangdata.com.cn/Thesis_D819766.aspx
