Research Of Aerial Triangulation Based On Dual Quaternion

The existing aerial triangulation methods are Euler angle method, unit quaternion method, dual quaternion method based on feature points. But Euler angle method and unit quaternion method solve the rotation transformation and translation transformation separately, thus the coupling error is introduced, and the exterior orientation elements can’t be correctly solved. And due to the above two algorithms do not take full consideration of spatial geometric topology and geometric constraint, they are unsuitable for aerial triangulation in the use of remote sensing images with larger system error, such as small UAV and close range images. The exterior orientation elements with lower precision are unable to meet the requirements of the following work of photogrammetry; this problem has become the bottleneck of the development of photogrammetric theory and technology. As a result, how to establish the model of aerial triangulation, which can describe rotation and translation uniformly, make full use of the geometric topology of imaging ray, and resolve the position and attitude of image efficiently, has become a hot topic in the field of photogrammetry, and it is also the main research direction of this paper.This paper puts forward a brand new theory to use dual quaternion as a mathematical tool for photogrammetry, and establishes a line photogrammetry imaging geometric model. First, use Plücker coordinate to represent imaging ray, then through the spiral movement of imaging ray in the space based on dual quaternion, coordinate transformation between image point and corresponding object point is finally achieved. Based upon this, make further research into unified position and attitude resolution of the image sequence, and construct the imaging geometric model and aerial triangulation model based on dual quaternion to improve the geometric positioning accuracy. The experimental results show that, the method is robust in the cases of image with small inclination, large inclination and when random error of image points is introduced. The adjustment accuracy of this new method is better than that of the conventional method and the requirements of the number and distribution of ground control points is almost the same as the conventional method. In addition, this new method can solve the exterior orientation elements of image with large inclination accurately, while the conventional method cannot. The research will improve the positioning accuracy of aerial triangulation, provide research ideas for image of unstable platform such as UAV, and expand the application fields of photogrammetry.