Theoretical and Empirical Evaluation of Geometric Performance of Multi-head Large Format Photogrammetric Sensors

The geometric performance of large format multi-head photogrammetric sensors, Intergraph DMC and Vexcel UltraCamD, was evaluated theoretically and empirically. Theoretical point determination accuracy coefficients were derived by simulation; point determination accuracy estimates are obtained by multiplying these coefficients by measurement accuracy and scale number. Multihead images have special multi-head distortions. The simulation study showed that the systematic image distortions caused systematic block deformations. In the empirical study three DMC blocks were evaluated. Point determination accuracy was evaluated using blocks with large overlap percentages. The standard error of unit weight was 2.5-3 μm in adjustments without self-calibration and approximately 2 μm in self-calibrating adjustments; the multi-head parameters efficiently reduced the systematic errors of image residuals. Without additional parameters, the point determination accuracy was 2-4 μm in horizontal coordinates in image and 0.030.07‰ of object distance in height; with the best available self-calibration model the corresponding values were 1.5-2 μm and 0.020.05‰. Large-scale blocks collected by DMC, UltraCamD and RC20 all gave fairly similar point determination accuracy values. The results indicated that the new digital sensors have great geometric accuracy potential; in order to take full advantage of this potential, appropriate additional parameter models must be developed for each sensor type.