• Tidak ada hasil yang ditemukan

20180930Komik 2018 Agus Harjoko IR4.0 Surveilance

N/A
N/A
Protected

Academic year: 2019

Membagikan "20180930Komik 2018 Agus Harjoko IR4.0 Surveilance"

Copied!
38
0
0

Teks penuh

(1)

www.ugm.ac.id

Locally Rooted, Globally Respected

A SURVEILANCE MODEL

IN INDUSTRY 4.0

Agus Harjoko

Department of Computer Science

and Electronics

Universitas Gadjah Mada,

Indonesia

(2)

www.ugm.ac.id

Locally Rooted, Globally Respected

Outline

UGM etc

Overview

SVS

LPU

DCCU

(3)

www.ugm.ac.id

Locally Rooted, Globally Respected

Universitas Gadjah Mada

Located in Yogyakarta, Indonesia

(4)

www.ugm.ac.id

Locally Rooted, Globally Respected

Dept. Computer Science and

Electronics

Study programs

Undergraduate in Computer Science

Undergraduate in Electronics and

Instrumentation

Master in Computer Science

Doctor in Computer Science

Research labs: 5

AI and Data Science

(5)

www.ugm.ac.id

Locally Rooted, Globally Respected

Industry 1.0

4.0

1784:

mechanical

loom

1850: kapal

uap dll

1970:

electronics and

IT

2016:

cyber-physical

(6)

www.ugm.ac.id

Locally Rooted, Globally Respected

Volume Data Growth

International Data Corp, 2017 (1 ZB = 1 trillion giga bytes)

00.05

2

12

47

163

2005

2010

2015

2020

2025

(7)

www.ugm.ac.id

Locally Rooted, Globally Respected

Key factors for data growth

Evolution of data: business to life-critical

Internet activities, socmed, transportation etc

Embedded system and IoT

Smart city, smart agriculture, etc

Mobile and real-time data

Real-time-data-based bussiness

AI systems change the landscape

AI, NLP, machine learning

Secutiry as critical foundation

(8)

www.ugm.ac.id

Locally Rooted, Globally Respected

Country readyness IR4.0

(9)

www.ugm.ac.id

Locally Rooted, Globally Respected

Making Indonesia 4.0

Roadmap Industry 4.0, April 2018

5 sectors

Food and beverage

Textile and apparel

Automotive

Electronics

Chemical

Why?

60% of manufacturing GDP

65% of manufacturing exports

(10)

www.ugm.ac.id

Locally Rooted, Globally Respected

Creating values from converging

technologies

(11)

www.ugm.ac.id

Locally Rooted, Globally Respected

(12)

www.ugm.ac.id

Locally Rooted, Globally Respected

Surveilance in Industri 4.0

In-door and out-door

Traffic Monitoring

Computer Vision Research group

(13)

www.ugm.ac.id

Locally Rooted, Globally Respected

Existing traffic monitoring in

Yogyakarta, Indonesia

(14)

www.ugm.ac.id

Locally Rooted, Globally Respected

Problems & Motivations

14

Missing frames

High computing power

(15)

www.ugm.ac.id

Locally Rooted, Globally Respected

Proposed Solution

15

Distributing computing tasks

Smart Video Sensor (SVS)

Local Processing Unit (LPU)

Data Center and Control Unit (DCCU)

Need communication data format (XML)

(16)

www.ugm.ac.id

Locally Rooted, Globally Respected

Illustration for proposed solution

existing

Proposed solution

(17)

www.ugm.ac.id

Locally Rooted, Globally Respected

SVS Architecture

Video

Sensor

Module

Video

Compression

Module

Feature

Extraction

Module

Feature

Representation

Module

(18)

www.ugm.ac.id

Locally Rooted, Globally Respected

Video Sensor Module

Video

aquisition

function

Resolution,

positioning,

lighting

(19)

www.ugm.ac.id

Locally Rooted, Globally Respected

Sample frame : Galeria

(20)

www.ugm.ac.id

Locally Rooted, Globally Respected

Video Compression Module

Compressing

video for

efficient

transmission

fungsi

Size, quality,

need for video

(21)

www.ugm.ac.id

Locally Rooted, Globally Respected

Feature Extraction Module

Function

Pre-processing,

segmenting,

extracting

features

Parameters

(22)

www.ugm.ac.id

Locally Rooted, Globally Respected

Feature Extraction Module

Frames

Segments

Video stream

Preprocessing

Segmentation

Feature

extraction

Ciri-ciri obyek

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target=“ROI”>1794</featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target=“ROI">2380</featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target=“ROI”>1794</featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target=“ROI">2380</featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target=“ROI”>1794</featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target=“ROI">2380</featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target=“ROI”>1794</featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target=“ROI">2380</featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

(23)

www.ugm.ac.id

Locally Rooted, Globally Respected

Pre-processing example

23

frame

Histogram

Equalization

(24)

www.ugm.ac.id

Locally Rooted, Globally Respected

Frame before and after segmentation

24

frame

Gaussian

Mixture

Model

(25)

www.ugm.ac.id

Locally Rooted, Globally Respected

Morphological Image Processing

25

morphological

opening and

closing

(26)

www.ugm.ac.id

Locally Rooted, Globally Respected

Feature Representation Module

Function

Representing

features

Parameters

(27)

www.ugm.ac.id

Locally Rooted, Globally Respected

Representation scheme

27

Object : 1

F1: value

F2: value

Fn: value

LocationNm: value VideoNm: value FrameNo: value Time: YYYY-MM-DDThh:mm:ss.sTZD

Object : 2

F1: value

F2: value

Fn: value

Object : n

F1: value

F2: value

(28)

www.ugm.ac.id

Locally Rooted, Globally Respected

XML Structure

<?xml version="1.0" encoding="utf-8"?>

<Z feature_XML="list">

<HFrame Value ="LocationNm"> galeria </HFrame>

<HFrame Value ="VideoNm"> galeria_4 </HFrame>

<HFrame Value ="FrameNo">194

<HFrame Value ="FTime">2014-01-16T06:10:15.21+07:00

</HFrame>

<ObjectNo Value ="Object"> 1 </ObjectNo>

<Feature Value ="WBB">156</Feature>

<Feature Value ="HBB">149</Feature>

<Feature Value ="ROI">23244</Feature>

<Feature Value ="centroid">[368 383] </Feature>

<Feature Value ="aspectratio">1</Feature>

<Feature Value ="PBB">610</Feature>

:

<Feature Value ="Dispersedness">15.154932</Feature>

:

<ObjectNo Value ="Object"> 3 </ObjectNo>

<Feature Value ="WBB">22</Feature>

<Feature Value ="HBB">37</Feature>

<Feature Value ="ROI">814</Feature>

<Feature Value ="centroid">[166 156] </Feature>

<Feature Value ="aspectratio">1</Feature>

<Feature Value ="PBB">118</Feature>

:

<Feature Value ="Dispersedness">15.233610</Feature>

</HFrame>

(29)

www.ugm.ac.id

Locally Rooted, Globally Respected

Transmission Module

Function

Transmitting

compressed

video and

features in XML

format

Parameters

(30)

www.ugm.ac.id

Locally Rooted, Globally Respected

Local Processing Unit

Working for local needs

Receives features from the SVS

converts features from XML to numbers,

selects the appropriate features based on

LPU’s

special needs

processes the selected features

results may be communicated with the DCCU

Example:

timing local traffic light controller,

(31)

www.ugm.ac.id

Locally Rooted, Globally Respected

Example: XML reading

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target= ROI > </featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target= ROI"> </featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target= ROI > </featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target= ROI"> </featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target= ROI > </featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target= ROI"> </featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target= ROI > </featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target= ROI"> </featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target= ROI > </featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target= ROI"> </featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

<?xml version="1.0" encoding="utf-8"?> <feature feature_extraction="program"> <headerframe target="frameNumber">90 <headerframe target="currenttime">2014-01-16T06:15:43.23+07:00

</headerframe>

<objectnumber target="Object">1</objectnumber> <featureitem target="WBB">39</featureitem> <featureitem target="HBB">46</featureitem>

<featureitem target= ROI > </featureitem>

<featureitem target="centroid">[378 42]</featureitem> <featureitem target="aspectratio">1</featureitem> <featureitem target="PBB">170</featureitem>

:

<featureitem target="Dispersedness">14.7226427583197</featureitem>

:

<objectnumber target="Object">3</objectnumber> <featureitem target="WBB">34</featureitem> <featureitem target="HBB">70</featureitem>

<featureitem target= ROI"> </featureitem>

<featureitem target="centroid">[553 128]</featureitem> <featureitem target="aspectratio">0</featureitem> <featureitem target="PBB">208</featureitem>

:

<featureitem target="Dispersedness">16.7321836705882</featureitem> </headerframe>

</feature>

XML representation

Convert

XML

numbers

(32)

www.ugm.ac.id

Locally Rooted, Globally Respected

Data Center and Control Unit

Coordinating LPUs in a certain region

Stores compressed video from SVS

May sends signals to LPUs

Receives object features extracted from SVS

Process features for regional needs

Example:

synchronizing traffic light controllers,

(33)

www.ugm.ac.id

Locally Rooted, Globally Respected

Applications

Traffic object classification

Tracking of classified object

Licence plate recognition

Licence plate and brand name recognition

Traffic light controller timing

Illegally parked vehicle identification

Warehouse monitoring

(34)

www.ugm.ac.id

Locally Rooted, Globally Respected

Classification of traffic objects

(35)

www.ugm.ac.id

Locally Rooted, Globally Respected

(36)

www.ugm.ac.id

Locally Rooted, Globally Respected

(37)

www.ugm.ac.id

Locally Rooted, Globally Respected

Thank you

Terima kasih

Matur nuwun

(38)

www.ugm.ac.id

Locally Rooted, Globally Respected

Features

Type of

Feature

Selected Features

Number

Shape

Bounding Box, WBB, HBB, MajorAxis, MinorAxis, ROI,

ObjectArea, Centroid, AspectRatio, Perimeter,

Dispersedness, InversDispersedness, EdgeDensity,

GrayHOG, PHOG, Equivdiameter, ConvexArea, Solidity,

ConvexHull, Orientation, Extent, EulerNumber,

Eccentricity, FilledArea, AreaRatio, AxisRatio,

Circularity, Elongation, Haar like feature, Gabor, SIFT

31

Texsture

LBP, Statistical texture, invariant moment

3

Color

colorHOG, colorMoment, colorHistogram,

colorCorrelogram, colorEnergy

5

Total number of features

39

Referensi

Dokumen terkait