OPTIMUM SELECTION OF IMAGES IN INDUSTRIAL PHOTOGRAMMETRY USING A FUZZY COMPUTATION SYSTEM
Optimized achieving the object coordinate points is the most important issue in close range photogrammetry. In this context, with applying network design, suitable angle and distance of imaging stations can be determined in the process of industrial photogrammetry. But in practice, according to the conditions, network design is not fully observed and imaging is performed experimentally. In this paper, employing a fuzzy computation system in order to specify appropriate camera stations based on network design constraints is described. In this fuzzy system, all range network design constraints are fuzzy modeled and then the system is decided whether each captured image is proper to use in bundle adjustment procedure or not. Practical results derived from modeling a piece of the aircraft showed that with the application of the fuzzy system and omission of inappropriate images, the mean accuracy of point coordinates on the object increased about 20 percent rate in comparison with using all images. Consequently, applying fuzzy computation can be used to determine appropriate images before utilization of all images in bundle adjustment procedure.
Amir Sharokh Amini - Islamic Azad University South Tehran Branch, Iran, Islamic Republic of