This document summarizes research analyzing the accuracy of 3D models reconstructed from spherical video images. The study acquired 134 spherical images of an indoor environment using a Garmin VIRB 360 camera. Images were extracted from video and reference points were collected. Aerial triangulation, dense image matching, and registration with TLS point clouds were performed. Results showed reconstruction accuracy was higher when reference points were distributed across images rather than clustered in the middle. However, the model had significant noise due to glass surfaces and stitching challenges. While geometric detail and accuracy were low, the 3D model could still enable some applications. Factors like calibration, stitching, resolution, and illumination variability affected the results.