This document summarizes a thesis that uses stereo vision with two webcams to measure object distances. The cameras capture 320x240 pixel images divided into 12 pairs of 20x320 pixel subimages. Edge detection is performed on the subimages and objects are matched between left and right images to determine disparity. Disparity values are then mapped to object distances using a linear spline function curve created from sample data, allowing object distances to be calculated from their disparities without other camera technical specifications.

1. http://digilib.its.ac.id
SUMMARY
Stereo Vision Untuk Pengukuran Jarak Objek
Dengan Mendeteksi Tepi Subimage
Stereo Vision for Objects Distance Measurement by Detecting Edge Subimages
Created by Swasono, Dwiretno Istiyadi
Subject : Permainan komputer
Subject Alt : Image processing
Keyword : : Stereo vision ; tepi subimage ; disparitas ; fungsi spline linear ; jarak objek.
Description :
Pada penelitian ini menggunakan stereo vision untuk mengukur jarak
objek yang ditangkap oleh dua webcam yang identik. Hasil pengukuran jarak ini
ditujukan untuk diterapkan pada robot sederhana, misalnya robot pada Kontes
Robot Cerdas Indonesi (KRCI). Dua webcam diletakkan pada posisi sejajar
dengan jarak tertentu dan dengan posisi yang tetap. Ukuran citra yang ditangkap
oleh setiap webcam adalah 320x240 piksel. Jika diamati, terjadi pergeseran posisi
objek yang ditangkap kedua kamera disebut disparitas. Citra stereo dari kedua
kamera dibagi menjadi beberapa subimage masing-masing berukuran 20x320
piksel sehingga didapatkan 12 pasang subimage berukuran 20x320 piksel.
Kemudian dilakukan deteksi tepi pada setiap pasang subimage menggunakan
turunan pertama deret graylevel pada satu baris piksel. Selanjutnya dilakukan
proses cropping dengan ukuran sebesar 20x20 piksel mulai posisi terdeteksinya
tepi. Dari sederetan subimage ukuran 20x20 piksel hasil cropping subimage
20x320 piksel di kiri dicari posisi pasanganya pada subimage 20x320 piksel di
kanan menggunakan metode pencocokan pola (template matching). Disparitas
atau besar pergeseran dua posisi subimage 20x20 piksel pada kedua pasang
subimage digunakan untuk menghitung jarak objek sebenarnya. Jika data-data
teknis kamera diketahui dan sudut yang dibentuk oleh arah pandang kedua kamera
(diharapkan sejajar) diketahui, jarak objek dapat dihitung dengan perhitungan
geometris. Namun data-data teknis tersebut tidak dicantumkan pada webcam dari
produsenya.Untuk itu dilakukan pendekatan lain untuk menghitung jarak objek
dari nilai disparitas. Untuk mengatasi masalah ini dengan menggunakan
pendekatan fungsi spline linear. Kurva spline dibuat dari sampel data-data
disparitas dan jarak objek hasil pengukuran. Kurva ini digunakan untuk
memetakan semua nilai disparitas terhadap jarak objek, sehingga jarak objek
terhadap kamera dapat diketahui.
Description Alt:
This research used stereo vision to measure the object distance captured by
two identical web cameras. This method of object distance measurement was
aimed to be implemented on simple robot, e.g robot for Kontes Robot Cerdas
Indonesia (KRCI). Two web cameras were placed in line with a certain distance
and fixed position. The image size which was captured on each camera was
320x240 pixels. From images that captured by both cameras, we could observe a
gap between object positions that is called disparity. Stereo images from both
Page 1

2. cameras then were divided into some subimages with 20x320 pixels size for each
so we get 12 pairs of 20x320 pixels subimage. The process was continued with
edges detection for each pair of subimages using the first derivative of graylevel
array on one line of pixels. Cropping process to 20x20 pixels size then applied
from the points where the edges were first detected. From 20x20 pixels of
cropped subimages in left array of 20x320 pixels subimages, we search the
position of its pair in right array of 20x320 pixels subimages using template
matching method. Disparity value of two 20x20 pixels subimages from one pair
of 20x320 pixels subimages was used to calculate the real object distance. If
technical datas of each camera and the angle of the two cameras view were
known, object distance could be calculated by geometrical calculation. But those
technical datas were not available, so we needed another approach to calculate
objects distance using their disparity values. To solve this problem, we used
linear spline function approach. Spline curve was created from samples of
disparity datas and the result of objects distance measurement. This curve was
used to map every disparity value to the object distance, so the distance between
the object and the cameras could be known.
Contributor : Prof.Ir. Handayani Tjandrasa, M.Sc Ph.D
Date Create : 11/09/2009
Type : Text
Format : pdf.
Language : Indonesian
Identifier : ITS-master-3100009035401
Collection : 3100009035401
Call Number : RTIf 006.42 Swa s
Source : Master Theses of Informatics Engineering, RTIf 006.42 Swa s, 2009
COverage : ITS Community Only
Right : Copyright @2009 by ITS Library. This publication is protected by copyright and permission
should be obtained from the ITS Library prior to any prohibited reproduction, storage in a
retrievel system, or transmission in any form or by any means, electronic, mechanical,
photocopying, recording, or likewise. For information regarding permission(s), write to ITS
Library
Full file - Member Only
If You want to view FullText...Please Register as MEMBER
Contact Person :
Mr. Edy Suprayitno (edy_supra@its.ac.id)
Mrs. Ansi M. Putri(ansi@its.ac.id)
Mr. Taufik Rachmanu (taujack@its.ac.id)
Mrs. Dewi Eka Agustina (dee@its.ac.id)
Thank You,
Page 2

3. Nur Hasan
ITS Digilib Supervisor
Page 3