DESIGN AND IMPLEMENTATION OF A PEOPLE COUNTING SYSTEM IN CROWDS DURING VERTICAL CIRCULATION IN THE ENTERTAINMENT SECTOR BASED ON VIDEO ANALYTICS
Entertainment is one of the most important aspects of living in urban areas. However, the high- level number of people in a crowd has the potential to result in crowd disaster. To minimize the potential of a crowd disaster, it is necessary to identify the conditions of a crowd. Currently, ident...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/85119 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Entertainment is one of the most important aspects of living in urban areas. However, the high-
level number of people in a crowd has the potential to result in crowd disaster. To minimize the
potential of a crowd disaster, it is necessary to identify the conditions of a crowd. Currently,
identification is conducted by monitoring system of a guard with the help of CCTV. However,
the system cannot give feedback and cannot determine whether the crowd condition is safe.
Therefore, a solution is needed to create a crowd detection system that can determine whether a
crowd condition is safe. This research aims to design, implement, test, and evaluate a detection
system in a vertical circulation area. The proposed detection system will be using YOLO
Algorithm to determine the level of service of an area. The methodology that is used in this
research is CRISP-DM. The result of this research is a crowd counting system that can determine
the level of service of an area and gives notification if the crowd condition is not safe. The
system’s final performance is a precision of 0.9661, recall of 0.8507, and an F1-Score of 0.9048. |
|---|