Block Triangulation by Bundles and Stereo-Units

Tests were run on simulated and real block data to evaluate the comparative accuracies of bundle and independent unit block triangulation. Cases run using the ISP simulated test block were: bundles, pairs, triplets overlapping by one photograph, and triplets overlapping by two photographs with and without the perspective centers being pass points. The bundle method was superior using the data with random errors only, while the units were slightly better for the data with random plus systematic errors. Cases run using the U.S. Geological Survey, Middleburg, Va. real data test block included those run with simulated data and also quadruplets overlapping by one photograph. The horizontal accuracy results of the bundle method were essentially the same as those from units. For vertical accuracy, the bundle was slightly better in all cases except for the quadruplets, where it was slightly worse.