Semi-automated Building extraction from CAD-Based Photogrammetry

The method of using pre-defined models to extract spatial information of man-made object digital image is called model-base object extraction, which has been a convincible approach to improve the existing Photogrammetry techniques. This paper used a semi-automated approach to extracting buildings from single aerial image by the Least-squares model-image fitting (LSMIF) algorithm. Each CAD model is associated with a set of pose parameters as well as a set of shape parameters and then projected to the image, by fitting the model to the image, these parameters will reveal both the location parameters and shape parameters. The work flow of this paper is as fellows: Firstly, the CAD model was selected and the approximate model parameters were given manually, then the parameterized model was projected to the image and the true values of parameters were calculated automatically. Experiment was done, and the result is good.