Automatic registration between remote sensing image and vector data based on line features

Abstract

In recent years, registration between the remote sensing image and vector data has seen its increasing usage in photogrammetry-related applications. Many researchers select higher-level primitives instead of traditional control points since the disadvantages of point features. This paper has proposed an automatic registration approach based on line features, which is insensitive to rotational and scale transformation. In this approach, the similarity measure is established according to the distance between conjugate entities in the same reference coordinate system. The collinearity equation is chosen as the transformation function. A modified Hough Transform is adopted to simultaneously estimate the exterior orientation parameters and obtain matching results. The automatic registration between the remote sensing image and vector data is achieved as well as the exterior orientation of the remote sensing image. Experimental results proved the reliability and feasibility of this approach.