Automatic Reconstruction for Small Archeology Based on Close- Range Photogrammetry

The paper present one approach to reconstruct the small object based on close-range photogrammetry. The goal is to generate the DSM (Digital Surface Model) of the object fast with high accuracy. Because close-range object have more distortion, occlusion, complex shape and surface. So the procedure is difficult to be standardized as aerial photogrammetry. In the sequences of steps (calibration, orientation, measurement, modelling), new algorithms and approach are applied to improve automation. Rotate platform can improve the controllability of photograph; planar scene calibration using 2D-DLT and bundle adjustment can improve the automation of calibration and orientation; A region segmentation algorithm based on color quantify is used to recognize the object region; multi-view matching in the constraint of 3D-TIN can generate the dense and reliable surface points automatically. The result demonstrates the improved approach is applicable and effective for 3D reconstruction of small object with sufficient texture.