تصحيح ساختار هندسي تصاوير رقومي غيرمتريك در فتوگرامتري و بينايي كامپيوتر به منظور استخراج اطلاعات سه بعدي اجسام

Geometric Information Extraction And Digital Surface Modeling Of "Cheshmeh Ali" Historical Monument Using Digital Close Range photogrammetric Technique

This research work is concerned with the digital surface reconstruction of Cheshmeh All historical site using the method of digital close range photogrammetry. Regarding the size of the. site and other photogrammetry factors such as film resolution, camera frame size, etc. four imaging stations were considered. Hasselblad non metric camera was installed on each predetermined station successively and four %80 overlapping images were taken to cover the entire area of the site. These images were then scanned with Intergraph scanner with a resolution of 14 micron pixel size. Non-linear image distortions, which were mainly due to the camera low geometric stability, were then modeled using the Multi Quadric (MQ) interpolation function. The MQ coefficients were determined using a dense pattern of control points. The object space coordinates of the control points were measured by total station. The statistical tests indicated an accuracy level of ±36 micrometer in image space for the checkpoints. The MQ coefficients were then used in the later stage to resample the original images and hence effectively the generated resampled output data were nearly free from distortions. These images were then fed into the ParadEyes DSP system to generate a three dimensional photogrammetric model. The X, Y and Z coordinates of the site were then extracted manually from the 3D models using the operatorʹs interactive capability offered by the DSP. The connectivity of the common features in the adjacent models was ensured using the stereo superimposition potentials of ParadEyes system. The geometric accuracy of the reconstructed 3D surface is estimated about ± 5 millimeter.