Artikel

Beranda Artikel Drone Photo Processing for 3D Modeling Visualization
Beranda Artikel Drone Photo Processing for 3D Modeling Visualization

Drone Photo Processing for 3D Modeling Visualization

  •  02 Agus 2019
  • Article/Article
  • 4422 viewed
Drone Photo Processing for 3D Modeling Visualization
Foto: Drone Photo Processing for 3D Modeling Visualization

By Bhima Dhanardono

Currently the use of Drones or Unmanned Aircraft has become common, especially related to infrastructure documentation. In general, drones are used to take photo or video data to be presented with editing as needed.
One thing that has not been widely applied in general is processing photos resulting from shooting using drones to create 3-dimensional (3D) visualization models by utilizing the photogrammetric process, namely mapping techniques through aerial photographs, in this case using photos taken from drones. In this paper we will only present an outline of the processing of the photo drone for visualization of 3D modeling. The main requirements needed for data collection and data processing include:

 

  1. Drone type Fixed wings (fixed wings) or rotary wings
  2. The drone has a satellite-based navigation system (GPS, Glonass and so on) to record photo metadata in the form of X, Y, Z (geotaging) location coordinates
  3. High resolution camera with directional and stabilizing system (Gimbal)
  4. Photo processing software for digital photogrammetric processing to produce 3D spatial data

 

The following are the things that need to be considered when establishing a plan for taking photos using drones as photogrammetric processing material:
1. Setting track and survey area

Pilihan pengaturan mode jalur terbang drone dan bentuk area survey
Pilihan pengaturan mode jalur terbang drone dan bentuk area survey

2. The area of the survey area, will determine the duration of flight time, as a basis for calculating fuel capacity requirements or the number of drone batteries to be used
3. The height of the flight, the lower the height of the drone the more accurate the photos will be, but the number of photos needed will be more (the duration of the survey and photogrammetric processing will be longer)
4. Results of shooting for 2D or 3D photogrammetric processing needs?

 

  1. For the 2D survey process it will be shorter because the direction of the camera is fixed, does not change, parallel to the direction of gravity
  2. For 3D it will take longer, because the camera will change the angle of view of each point of the shooting location, but the results will be more accurate and realistic

 

5. Sidelap and frontlap, i.e. the size of the same area or overlap of a photo with subsequent photos in sequence and next to each other to cause parallax, ie the apparent shift of an object against its background caused by changes in camera position. The greater the percentage of sidelap and frontlap, the results will be more accurate but the impact will increase the duration of the survey and photogrammetric processing later.
The author has tried to do some drone photo processing for visualization of 3D modeling with photographs taken from the Sigandul Bridge photograph on March 15, 2018 from 13.27 to 13.55 on the BTS road section. Kab. Wonosobo - Parakan Km. MGL. 40 + 310.
Taking photos is done using DJI Mavic Pro drones that produce 696 photos with a resolution of 72 dpi, the size of each photo is 4000x3000 pixels. Photos used for photogrammetric processing are not through the editing process at all.

Kumpulan foto-foto hasil pemotretan dengan Drone

Kumpulan foto-foto hasil pemotretan dengan Drone

 

Photogrammetry processing using commonly used software, there are many choices that can be taken such as Pix4Dmapper, VisualSfM, 3DF Zephyr Pro, Agisoft Photoscan and so on, mostly paid software.
The first step is setting the reconstruction of the camera position through the photos used. The system in the software will automatically adjust the position and direction of the camera angle based on the metadata of each photo file.



Posisi kamera sebelum diatur
Camera position before set

Tahap pengaturan rekonstruksi posisi dan arah kamera

Stage of reconstruction of camera position and direction

After setting the camera, the next step is to build Dense Cloud, which is a data set of points that has X, Y, Z, Dense Cloud coordinates, which will form 3D objects. The more photos used, the more accurate the results will be, but the effect will make the processing duration longer. Likewise, setting the level of processing accuracy, the more accurate, the longer the process.

Following are the results of photogrammetric processing that has formed 3D object visualization.

Hasil pemrosesan fotogrametri

Photogrammetric processing results

 

Foto asli

Original photo

 

Hasil pemrosesan fotogrametri

Photogrammetric processing results

Foto asli

Original Photo

 

Hasil pemrosesan fotogrametri
Photogrammetric processing results
Foto asli
Original Photo

 

Hasil pemrosesan fotogrametri
Photogrammetric processing results

 

Foto asli
Original Photo
Hasil pemrosesan fotogrametri
Photogrammetric processing results
 

 

Foto asli

Original Photo
Tampak Atas
Top view

Making 3D model visualization can be used for visual communication with various parties both lay and experts about the shape of the face of the earth or the shape of existing infrastructure objects for example for the basis of preliminary design, modeling of landslides, modeling scouring rivers, bridges, flyovers, roads and so on.

Author: Bhima Dhanardono, ST, M.Eng
Road and Bridge Engineering PPK Young Experts Planning,
Planning and Supervision of the National Roads of Central Java Province,
Center for Implementation of National Road VII Semarang