TY  - THES
ID  - digilib7218
UR  - https://digilib.uin-suka.ac.id/id/eprint/7218/
A1  - ANDI FEBRIYANTO , NIM. 08650020
Y1  - 2013/01/22/
N2  - The monitoring system is used to perform oversight functions for determining the
level of pedestrian crowds. To monitor pedestrians need to be created a system
that can detect and calculate the pedestrian crowds automatically. This research
will develop and compare two methods. Detection methods used are background
subtraction and haar like feature method. This research aim to determine the
performance of each method and compare the performance of both methods in
detecting pedestrians.
Background subtraction method can detect the subtraction in the background by
converting images into binary image and determines the sensitivity of
background's pixel changes. Haar like feature method uses 418 positive samples
and 644 negative samples as the data domain. The process of training carried out
on the samples which later became the basis for object detection pedestrians. Then
for both methods performed object detection and calculations processes.
From the test results of both methods with various test conditions, the detection
results using background subtraction method for pedestrian monitoring was
87.9%, while the result of detection using Haar-like feature method for pedestrian
monitoring was 65%. So the background subtraction method is better than Haarlike
Feature method for pedestrian monitoring.
Keywords : background subtraction, haar like feature, pedestrian detection


PB  - UIN SUNAN KALIJAGA
M1  - skripsi
TI  - ANALISIS KINERJA METODE BACKGROUND SUBTRACTION DAN
HAAR-LIKE FEATURE UNTUK MONITORING PEJALAN KAKI
MENGGUNAKAN KAMERA WEBCAM

AV  - restricted
ER  -