ZEMCH 2019 International Conference Proceedings April.2020 | Page 370
(a) (b)
Figure 2. The captured images of used python code for S‐CNN development (WBS2) : (a) Image size
revision; (b) Training and testing for WBS level 2 categories
2.2. BIM Model Update Using Digital Photo Information
2.2.1. Extract Location Information from Digital Photographs
Digital pictures contain not only the shape of the subject at the moment of taking, but also other
related information. In specific, when confirming the metadata of the digital image created with a
device such as a smartphone, which is having a function of the positioning tracking system, the
recorded date and time as well as the photographed location information are stored in detail. In case of
outside space, the application of Global Positioning System (GPS) ensures the higher accuracy, and in
indoor space’s case, it is more accurate to determine the location information by using WiFi Positioning
System (WPS) [15].
In accordance with the subject of the usage, the code type of GPS is defined as Coarse (C/A) Code
for civil use, and Precise (P) Code for military use. When it comes to the accuracy rate for GPS and WPS,
they are 0.3m and 2.3m respectively, and both of them have limitation in elevation definition [16].
Therefore, in the indoor space, the positional information on the plane is obtained from the input
picture, but the level information is directly selected by the operator.
2.2.2. Objects Matching and 4D BIM Model Update
The method of matching information acquired from the photographs with the object of the
BIM model is as follows:
1.
2.
3.
4.
Designate the origin of project BIM model with accurate latitude and longitude, and then set project
north and real north direction as coincided
Select an element from the model that has the same position and WBS information as that of the
object in the picture
Update the properties of the selected building element, select the status from three steps‐
“beginning”, “in progress”, and “completed”‐, which do not require subjective judgement
Identify the process status of the object at the time of taking a picture by assigning different color
for each input property (“beginning”: red, “in progress”: yellow, “completed”: blue)
359
ZEMCH 2019 International Conference l Seoul, Korea