LASER SCANNER versus DRONE PHOTOGRAMMETRY data
The following data contain most of the informations from photos, 3D photogrammetry and laser scanned files. All photographies (10 m, 20m and 40m flights) from the Mavic Pro drone that generate the 3D Zephyr models through photogrammetry for crime scene registration (.zep). All Trimble X7 laser scanner data (.e57) from the six captures also can be found at LS X7 folder. This models were used for comparation with the 3D point of cloud generated from a Laser Scanner Trimble X7 acquired at the same place. Using ANOVA for statistical comparison of the results from know distances (compared with a total station), no significant differences were found. Thus, drone photogrammetry can be indicated as a cheaper, low time consuming and powerful tool for crime scene registration.
Steps to reproduce
All photographs acquired from the drone method were processed in 3D Zephyr software, which has a tool to automatically recognize the GCPs generated from its template. However, excluding the 10 m flight, the autodetection tool was not useful at higher altitudes. This disadvantage was attributed to the size of the A4 paper on which the GCP was printed and the notion that April Tag 16H5 might not be suitable for high-altitude flights. Nevertheless, the software allows refitting to larger sizes, such as A3, or the use of another kind of April tag for GCPs, such as 41H12, which was not available on-site at the time of the experiment. Furthermore, the manual pick tool for GCPs was very useful and was employed in this scenario after the photogrammetry process was finished. For each altitude, the captured photographs were inserted in the 3DZephyr and evaluated on the software workflow to check its basic characteristics for photogrammetry, such as consistency, similarity, and camera calibration. The objective was to build a point cloud regarding the overlay degree by selecting the aerial/nadiral photograph category set up with deep presets, taking approximately 2 minutes to reconstruct the 3D model from the photographs. The next step in the workflow was to generate a denser point cloud, with the same parameters of the aerial/nadiral category and high-detail presets (taking approximately 11 minutes. Next, a mesh extraction procedure was then performed from the dense point cloud again with the same parameters of aerial/nadiral category and high-detail presets (taking approximately 4 minutes, and a textured mesh was obtained (taking approximately 3 minutes). An Avell notebook with an Intel(R) Core(TM) i7–10750H CPU @ 2.60 GHz, 32 GB RAM, and a GPU-dedicated card Nvidia Geforce RTX 2060 with 6 GB RAM was employed to perform the photogrammetry process with the 3DZephyr.