APLIKASI TEKNOLOGI INDERAJA UNTUK MENDUKUNG PROGRAM KETAHANAN PANGAN NASIONAL HYPERSRI-SOFT DISAJIKAN OLEH: SIDIK MULYONO PADA KULIAH UMUM UNIVERSITAS LAMPUNG 05 NOPEMBER 2014
AGENDA KULIAH UMUM • PERKENALAN • PENDAHULUAN • PENGENALAN INDERAJA
• APLIKASI INDERAJA • KESIMPULAN • DISKUSI
•PERKENALAN • PENDAHULUAN • PENGENALAN INDERAJA • APLIKASI INDERAJA
• KESIMPULAN • DISKUSI
CURRICULUM VITAE SIDIK MULYONO (BERKELUARGA + 4 ANAK)
PEKERJAAN: Perekayasa di Badan Pengkajian dan Penerapan Teknologi (Ketua TIM HyperSRI-Soft) KEAHLIAN: Pemodelan Remote Sensing PENDIDIKAN: S1 Faculty of Science and Technology Nihon University, lulus tahun 1991 S2 Faculty of Science and Technology Nihon University, lulus tahun 1993 S3 Faculty of Computer Science, Univeritas Indonesia, lulus tahun …….
CURRICULUM VITAE DAFTAR PAPER YANG TERKAIT DENGAN PEMODELAN RS: A Paddy Growth Stages Classification For PiSAR-L2 Data Using Fuzzy Model, Asean Conference for Remote Sensing, 2013
An Ensemble Incremental Approach of Extreme Learning Machine (ELM) For Paddy Growth Stages Classification Using MODIS Remote Sensing Images, International Conference for Advanced Computer And Information Systems, 2013 Paddy Growth Stages Classification Using MODIS Remote Sensing Images With Support Vector Machines, International Conference for Advanced Computer And Information Systems, 2012
CURRICULUM VITAE DAFTAR PAPER (Lanjutan): Fuzzy Classifier of Paddy Growth Stages Based on Synthetic MODIS Data, International Conference for Advanced Computer And Information Systems, 2012 Genetic Algorithm Based New Sequence of Principal Component Regression (GA-NSPCR) for Feature Selection and Yield Prediction Using Hyperspectral Remote Sensing Data, IGARSS 2012 HyMap Airborne Hyperspectral Technology for Rice Yield Prediction in Case of Karawang Distric, PIT-MAPIN 2012
CURRICULUM VITAE DAFTAR PAPER (Lanjutan): Non linear dynamic and remote sensing based economic-agricultural growth model, UKP4 2012
HyMAP Airborne Hyperspectral Remote Sensing Application for Predicting Rice Production at Karawang District, PIT-MAPIN 2012 Feature Selection of Hyperspectral Remote Sensing And Prediction Model With Genetic Algorithm And Principal Component Regression, PIT-MAPIN 2011 Paddy Yield Prediction With Principal Component Regression Method Using Hyperspectral Data, PIT-MAPIN 2010
CURRICULUM VITAE DAFTAR PAPER (Lanjutan): Principal Component Regression Analysis on Automatic Sleep Apnea Detection from ECG Data, International Conference for Advanced Computer And Information Systems, 2010 Predicting rice harvest with mathematical model, Senarai Teknologi Untuk Bangsa, Pusat Teknologi Inventarisasi Sumberdaya Alam, Badan Pengkajian dan Penerapan Teknologi, 2010.
Weekly Stats Update
SEPUTAR BPPT VISI Pusat unggulan teknologi yang mengutamakan kemitraan melalui pemanfaatan hasil rekayasa teknologi secara Optimal MISI 1. Memacu perekayasaan teknologi untuk meningkatkan daya saing produk industri 2. Memacu perekayasaan teknologi untuk meningkatkan pelayanan publik instansi pemerintah. 3. Memacu perekayasaan teknologi untuk kemandirian bangsa
Research
Operational
Development
Model Prediksi Skala Provinsi
Akuisisi data Hyperspectral Airborne Campaign
(Integrasi model Hyperspectral & Multispectral) Resolusi 1Km
Resolusi 500m
Peta distribusi produktifitas padi Kabupaten Karawang Total produksi: 186,898 Ton Luas panen: 30,966 Ha Produktifitas: 6.00 Ton/Ha
Citra hyperspectral Kab. Karawang
2011
Model Prediksi Skala Kabupaten 2012
Operasionalisasi sistem prediksi padi nasional
Peta distribusi fase tumbuh padi
Peta distribusi produktifitas padi
2013
2014
TUJUAN Mengembankan potensi teknologi penginderaan jauh untuk bidang pertanian dalam rangka mendukung program ketahanan pangan nasional.
Membangun model prediksi fase tumbuh dan produktivitas (panen) padi pulau Jawa berbasis pembelajaran mesin (Machine learning). Melakukan proses alih teknologi (technology transfer), dalam rangka menyiapkan sumberdaya manusia yang handal untuk mengimplementasikan monitoring kondisi terkini tanaman padi dan prediksi produktivitas padi secara operasional pada skala nasional.
Satelit Multispectral MODIS (Spaceborne) > 600 km
System integratio n
Yes
Spaceborne Image
Validation
No Distribution map
Prediction Model & Growthstage classifier
Machine learning
Upscalling & spectral subgroup
HyMap Hyperspectral (Airborne) 2,000 m
Airborne Image
Prediction Model & Growthstage classifier
Supersite
Supersite
Supersite
Machine learning
Spectral Domain
Supersite Collecting spectral data
15
Spectral library
Spectral correction
• PERKENALAN
•PENDAHULUAN • PENGENALAN INDERAJA • APLIKASI INDERAJA
• KESIMPULAN • DISKUSI
BERAS BAGI MANUSIA Sumber bahan pangan kebutuhan 1,2 triliun masyarakat dunia Tingkat kebutuhan akan beras yang semakin meningkat
Memiliki peranan penting dalam ketahanan pangan Berdampak bila kekurangan atau kelebihan
Penanaman padi yang tidak seragam di setiap wilayah
Luas lahan sawah Indonesia: ± 8.1 juta ha
Kebutuhan pangan nasional: 33~38 juta ton per tahun
Spekulasi impor beras Metode konvensional pengamatan padi
Perbaikan sistem prediksi Yang lebih efektif • • • •
Tidak tersedia data stok beras secara spasial
Lebih akurat Secara spasial Lebih cepat Tanpa merusak
Menyempurnakan sistem yang sudah ada dalam rangka mendukung Program Ketahanan Pangan Nasional
APA SOLUSINYA? Pemantauan tanaman padi dengan memanfaatkan data citra satelit
How?
Pemantauan Tingkat Kehijauan & Fase Tanaman Padi Sumber LAPAN: http://www.lapanrs.com/simba/detail/4?i=26977
MODIS NDVI
Kalender Tanam http://katam.litbang.deptan.go.id/main.aspx?mode=fullscreen
• PERKENALAN
• PENDAHULUAN
•PENGENALAN INDERAJA • APLIKASI INDERAJA
• KESIMPULAN • DISKUSI
Apa itu remote sensing? Kata Kunci:
Remote = Jauh, terpencing Sense
= merasa, mengindera
Remote sensing = Peninderaan jauh Inderaja
Apa itu remote sensing? Definisi remote sensing (RS)? Suatu proses pengumpulan informasi mengenai objek yang dilakukan tanpa menyentuh objek tersebut. Alat apa yang terpenting dalam RS? Sensor
Alat remote sensing apa yang ada pada tubuh kita? Mata Alat RS apa yang anda ketahui yang ada disekitar kita? Stethoscopes Apa yang diukur oleh RS?
Gelombang elektromagnetik Apa itu gelombang elektromagmentik?
Gelombang energi yang terdiri dari medan elektrik dan magnetik, yang bergerak saling tegak lurus. Changing magnetic field can induce an electric field, and electric field changes can also cause magnetic fields (James Clerk Maxwell, 1831)
Gelombang elektromagnetik
26
Gelombang elektromagnetik
27
Gelombang elektromagnetik Intensitas gelombang (Cahaya/RADAR/etc) Sensor & converter
Digital number Reflectance/ Brightness/ Backscatter
Satelit Inderaja Satelit dengan sensor pasif
Satelit dengan sensor aktif
RADAR: RAdio Detection And Ranging Microwave
Pengenalan spektrum warna (reflektansi cahaya)
Banyak warna yang dapat terbentuk dari kombinasi warna dasar (Red, Green, Blue) dengan proporsi yang bervariasi
Pengenalan spektrum warna (reflektansi cahaya)
Respon tanaman terhadap cahaya Klorofil yang terutama menyerap cahaya violet, biru, dan merah, Sedangkan cahaya hijau tidak diserap, melainkan dipantulkan, sehingga mengakibatkan daun terlihat hijau. Struktur sel dalam mesofil mengakibatkan reflektansi tinggi cahaya near infrared, sedangkan klorofil bersifat transparan terhadap near infrared (Campbell, 1996).
Fase tumbuh padi
Fase Vegetatif 0 - 40 hari Perkecambahan
Fase Reproduktif 41 - 75 hari
Pembentukan Malai
Pembungaan
Fase Pematangan 76 - 110 hari Gabah Matang
Fase tumbuh padi Vegetative stage
Reproductive stage
Ripening stage
Fase tumbuh padi Di daerah tropis, fase reproduktif 35 hari dan fase pematangan sekitar 30 hari. Perbedaan masa pertumbuhan ditentukan oleh perubahan panjang waktu fase vegetatif. Sebagai contoh, IR64 yang matang dalam 110 hari mempunyai fase vegetatif 45 hari, sedangkan IR8 yang matang dalam 130 hari fase vegetatifnya 65 hari.
Fase tumbuh padi Bulan 1 Bulan 2
Bulan 3
Bulan 3.5
Reflektansi gelombang NIR dan penyerapan gelombang visible maksimum pada vegetasi terjadi pada fase vegetatif akhir/ Reproduktif awal
Fase tumbuh padi Vegetative
Ripening
Reproductive
37
• PERKENALAN
• PENDAHULUAN • PENGENALAN INDERAJA
•APLIKASI INDERAJA • KESIMPULAN • DISKUSI
Satelit dengan sensor pasif
39
SEPUTAR HYPERSPECTRAL
Generasi terbaru teknologi remote sensing yang memiliki ratusan kanal yang mampu mengidentifikasi objek di muka bumi secara rinci dan akurat (resolusi spektral dan spasial tinggi) Kendala-kendala: 1. Pra-pengolahan dan analisis data cukup kompleks 2. Membutuhkan kapasitas RAM komputer yang sangat besar. 3. Biaya pengoperasian mahal 4. Data lapangan yang tersedia sangat terbatas 5. Belum tersedia satelit yang beroperasi.
Hyperspectral vs Multispectral (HyMap vs MODIS for Karawang District)
HyMap Hyperspectral
MODIS Multispectral
Integrasi Model berbasis Hyperspectral & MODIS
Pemodelan Inderaja Pemodelan komputer menggunakan data inderaja untuk mendefinisikan objek, area, atau fenomena melalui suatu algoritma matematika
RS Conventional Method
RS Model
Statistically
Dynamically
Unsupervised
Supervised (data learning)
Use end member
Use data library
For a moment
For general
HYPERSPECTRAL BASED (Spatial resolution 4.5m)
FIELD CAMPAIGN
Indramayu District: 06-28-05.65S, 108-03-03.14E Karawang District: 06-15-46.24S, 107-25-05.10E
500m
10m 10m
500m
44 Quadrate
AIRBORNE CAMPAIGN
The flight lines set for airborne campaign
The composited Image from 10 lines 45
Pre-processing HyMap Airborne HS Airborne HyMap HS
Field DGPS Data
Table of ubinan from Dinas
Spectral data extraction Water Absorption effect elimination
Yield data convertion
Spectral Matrix X
Yield vector Y Prediction model
(Containing water absorption effects)
Gabungan antara Genetic Algorithm (GA) dengan teknik baru NSPCR, yaitu teknik regresi dengan menyusun ulang urutan PC berdasarkan order of importance
Coefficient correlation R2
0.14 0.12 0.1
80% cut-off
0.08 0.06 0.04 0.02 Improved accuracy
0 -0.02 0
10 20 Eigen Number n
30 47
Band chromosome generated by GA 8-round bootstrap resampling method
SVD Factorization Calculating R2 for Score Matrix Yes
Xtrain
Ytrain
Re-order PC
Xtest
Ytest
Genetic Algorithms Yes
Model testing
Gen less then 200
Model calibration
Resampling round less then 8
Error calculation using Lp-norm
No
No
New Sequence Principal Component Regression (NSPCR) *) Sidik Mulyono, et. al, submitted to IGARSS Conference 2012
Model validation
Fitness function If Error smaller then prev. gen Yes 48
No
The Best Model: smallest number of band & smallest error
46 bands
Band reduction using GA
Error distribution
49
Error Method
Total number of bands
RMSEval
RMSEval
Original
112
0.228
0.619
L1-norm
64
0.034
0.422
L2-norm
46
0.027
0.424
L∞-norm
61
0.028
0.420
GA-SPCA(*)
(*) Haibo Yao, Lei Tian, A Genetic-Algorithm-Based Selective Principal Component Analysis (GA-SPCA) Method for High50 Dimensional Data Feature Extraction, IEEE Transactions on Geoscience and Remote Sensing, vol. 41, no. 6, June 2003
PADDY YIELD MAP
The composited Images from 10 lines of HyMap Hyperspectral Airborne Campaign (Karawang 2011)
Paddy yield distribution map with GA-NSPCR
*) This paper was submitted to International Geo-science And Remote Sensing Symposium (IGARSS) 2012, Munich German
VALIDATION RESULT Paddy Yield Prediction Table For Karawang District No
SubDistrict
Total Production/ Harvested Area [Ton/Ha]
Yield [Ton/Ha]
Data from District Office [Ton/Ha]
Accuracy
1 Lemahabang
27,553 / 4,142
6.65
6.14
91.73%
2 Talagasari
22,881 / 3,409
6.71
6.64
98.95%
3 Banyusari
6,068 / 986
6.15
6.47
95.00%
4 Cikampek
9,183 / 1,893
4.85
6.31
76.89%
5 Cilamaya Kulon
14,909 / 2,089
7.14
5.64
73.49%
6 Klari
27,094 / 4,875
5.56
6.06
91.76%
7 Majalaya
6,679 / 1,218
5.48
6.64
82.53%
8 Rawamerta
17,464 / 2,827
6.18
6.47
95.46%
9 Tempuran
21,962 / 2,946
7.46
5.56
65.85%
10 Tirtamulya
16,918 / 3,532
4.79
6.14
77.99%
HYPERSPECTRAL & MULTISPECTRAL INTEGRATED MODEL (Spatial resolution 1 Km)
PADDY YIELD PREDICTION MODEL 1. Pengumpulan data sampel spektral & ubinan
Original HyMap (124 bands) data samples correspond to 55 Yield data samples
MODIS synthetic data samples (band: 1, 2, 4, 5, 6, 7)
PADDY YIELD PREDICTION MODEL 2. Machine Learning Algorithm (NS-PCR) 8-round bootstrap resampling method
Xtrain
SVD Factorization
Re-order PC
Calculating R2 for Score Matrix
Yes
Ytrain
Ytest
Model testing
Model calibration
Resampling round < 8
Xtest
No
Error calculation using Lp-norm
The Best Model
PADDY YIELD PREDICTION MODEL 3. Kalibrasi model machine learning (training & testing)
The Best Model
PADDY GROWTH STAGE PREDICTION MODEL 1. Hyperspectral knowledge approach based growth stages Vegetative Stage
Reproductive Stage
Ripening Stage
PADDY GROWTH STAGE PREDICTION MODEL 2. Pengumpulan data sampel spektral dari citra MODIS No
Kelas 1 2 3 4 5 6 7 8 9 10 11
veg1 veg2 veg3 rep1 rep2 rep3 rip1 rip2 rip3 others cloud Total sampel Jumlah Band MODIS
Jumlah sampel 103 110 120 120 120 117 120 120 120 120 85 1,255 6
PADDY GROWTH STAGE PREDICTION MODEL 3. Classification method
Support Vector Machines with: • Radial basis function (RBF) kernel trick
Software: • LIBSVM introduced by Chih-Wei Hsu et al • Interactive Data Language (IDL) program
PADDY GROWTH STAGE PREDICTION MODEL 4. Structure of Balanced Branches (BB-SVM)
SVM01
(+)
(-)
SVM02
SVM07
(+)
(-)
(+)
(-)
SVM03
SVM05
SVM08
SVM10
(+)
(-)
(+)
(-)
(+)
(-)
(+)
(-)
SVM04
Rep2
SVM06
Rip1
SVM09
Rep1
Rip2
Soil
6
10
12
(+)
(-)
Veg1
Rip3
3
11
7
(+)
(-)
Veg3
Cloud
5
2
9
(+)
(-)
Veg2
Rep3
4
8
PADDY GROWTH STAGE PREDICTION MODEL 5. Configurasi of class label No Class
SVM01 SVM02 SVM03 SVM04 SVM05 SVM06 SVM07 SVM08 SVM09 SVM10
1
Veg1
1
2
Veg2
-1
3
Veg3
1
4
Rep1
-1
5
Rep2
1
6
Rep3
-1
7
Rip1
1
8
Rip2
-1
9
Rip3
1
10 Cloud
-1
11 Others
1
1
1
1
-1 1
1
-1
1
1
-1
1
1
1
1
-1
-1 1
1
-1
-1 1
-1
1
-1
-1
-1 1
-1
PADDY GROWTH STAGE PREDICTION MODEL 6. Calibration for parameter C dan γ SVM01
SVM02
SVM03
SVM04
SVM05
SVM06
SVM07
SVM08
SVM09
SVM10
PADDY GROWTH STAGE PREDICTION MODEL 7. Training, Testing, and Validation Classifier SVM01 SVM02 SVM03 SVM04 SVM05 SVM06 SVM07 SVM08 SVM09 SVM10
Train 0.981 0.959 0.744 0.957 0.838 0.934 0.915 0.662 0.930 0.713
BB-OAA Test 0.974 1.000 0.864 0.967 0.891 0.949 0.864 0.700 0.953 0.750
Val 0.980 0.964 0.759 0.958 0.845 0.936 0.909 0.667 0.924 0.717
Class/ Growth Stages Cloud Others Veg1 Veg2 Veg3 Rep1 Rep2 Rep3 Rip1 Rip2 Rip3
Classification Accuracy 0.718 0.467 0.631 0.627 0.683 0.842 0.933 0.795 0.858 0.808 0.833
DATA CITRA MODIS NO
PRODUK MODIS
TANGGAL
1 MOD021KM
2012-03-29
2 MOD02.1KM
2012-04-24
3 MOD02.1KM
2012-04-28
4 MOD02.1KM
2012-05-12
5 MOD02.1KM
2012-05-23
6 MOD02.1KM
2012-06-13
7 MOD02.1KM
2012-06-15
8 MOD02.1KM
2012-07-22
9 MOD02.1KM
2012-07-31
10 MOD02.1KM
2012-08-19
11 MOD02.1KM
2012-09-15
12 MOD02.1KM
2012-10-10
MASKING STANDARD PADDY FIELD
CITRABAKU MODIS ASLI PETA BAKUSAWAH SAWAH
PROGRAM IDL PREDIKSI PADI
PROGRAM IDL PREDIKSI PADI 1200000
2012-03-29
800000
600000
400000
200000
Fase Tumbuh
Rip-3
Rip-2
Rip-1
Rep-3
Rep-2
Rep-1
Veg-3
Veg-2
0
Veg-1
Luas Area (ha)
1000000
PROGRAM IDL PREDIKSI PADI
PERIODE TANAM DAN PANEN
Sumber : Sumarno, Periodisasi Musim Tanam Padi Sebagai Landasan Manajemen Produksi Beras Nasional, Sinar Tani No. 3136, Tahun XXXVI
-
TANGGAL 11/30/2012
Paneen Kecil
13,56 Juta Ton
10/31/2012
9/30/2012
8/31/2012
Paneen Gadu
18.00
7/31/2012
4.00
6/30/2012
6.00
5/31/2012
8.00
4/30/2012
10.00
3/31/2012
12.00
Paneen Raya
14.00
2/29/2012
1/29/2012
PRODUKSI (JUTA TON)
PROGRAM IDL PREDIKSI PADI
PREDIKSI PRODUKSI PADI SE JAWA 11,65 Juta Ton
16.00
2.00
TANGGAL 11/30/2012
10/31/2012
9/30/2012
8/31/2012
7/31/2012
6/30/2012
5/31/2012
4/30/2012
3/31/2012
2/29/2012
1/29/2012
LUAS (ha)
Fase Vegetatif
1800000
1600000
1400000
1200000
1000000
800000
600000 Veg-1
400000 Veg-2
200000 Veg-3
0
KENDALA UTAMA Citra MODIS dipengaruhi oleh kondisi perawanan Resolusi spasial terlalu kasar Proses komputasi lama karena harus masking peta baku sawah
NEW CONCEPT ON PADDY GROWTH STAGES PREDICTION (Spatial resolution 500m)
High Temporal Resolution of EVI Growth Stage
Slope direction
Plowing Negative
Time series MODIS data
Early vegetative Positive
Late vegetative Positive
Original time-based phenology stages Reproductive Positive Ripening
Negative
Smoothed time-based phenology stages Harvesting Negative Post-harvest Negative
First & second derivative time-based phenology stages
EVI-based paddy growth stages detection strategy
Second derivative EVI curve
First derivative EVI curve Smoothed EVI curve (Savitzky Golay Filter) Original EVI data
Profil EVI yang sudah diberi label berdasarkan nilai FD dan SD Nilai First Deriv
Nilai Second Deriv
Class
Class Index
+
+
Early Vegetative
1
+
+
Late Vegetative
2
+
-
Reproductive
3
-
-
Ripening
4
-
+
Harvesting
5
-
+
Post-Harvest
6
-
+
Plowing
7
Spectral-based Paddy Growth Stages Ripening
Reproductive
Harvesting
Late vegetative
Plowing
Early vegetative Post-harvest
MACHINE LEARNING METHOD (A New Heuristic Decision Tree)
Paddy Growth Stages LSWI>EVI FD<0 Plowing
LSWI<EVI FD>0
FD>0
EVI<0.15
EVI>0.15
Early Vegetative
Late Vegetative
Reproductive
FD<0 EVI>0.28
0.28<EVI<0.24
EVI<0.24
Ripening
Harvesting
Postharvest
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
VALIDATION MODEL
Paddy Growth Stages Adjustment A New Heuristic Decision Tree (HyperSRI Soft) 1. Early Vegetative
2. Late Vegetative 3. Reproductive
4. Ripening 5. Harvesting 6. Post-Harvest
7. Plowing
Field Campaign
1. Transplanting (TR) 2. Early Heading (EH) 3. Late Heading (LH)
4. After-cut (AC) 5. Bare-land (BL)
6. Land Preparation (LP)
VALIDATION RESULT Field Campaign Transplanting Veg-1 HyperSRI Soft
Veg-2
17
Early Heading
Late Heading
Aftercut
Bareland
Land preparation
0
0
0
0
2
1
0
1
0
0
1
0
0
0
1
0
0
2
0
0
Rep
1
11
Rip
0
0
Harvesting
0
0
Post-harvest
0
0
0
Plowing
0
0
0
16
0
12 0
0
15
Total accuracy = 88.75%
The Advantages of HyperSRI-Soft Rapid: Using Machine Learning Method
Accurate: Validated by ground reference Consistent in growth stages change: Based on time series phenological profile Robust: Unaffected cloud conditions Moderate resolution: Using MODIS 500m
Parallel with paddy field detection
Paddy Field
Tanggal: 20140101 1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140109
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140117 1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140125
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140202
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140210
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140218
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140226
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140306
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140314
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140322
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140330
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140407
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140415
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140423
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140501
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140509
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140517
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140525
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140602
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140610
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140618
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140626
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140704
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140712
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140720
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140728
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140805
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
Tanggal: 20140813
1,000.00 900.00 800.00 700.00 600.00 500.00 400.00 300.00 200.00 100.00 -
Veg-1
Veg-2
Rep
Rip
Harvesting Post-harvest
Plowing
PADDY GROWTH STAGES ANALYSIS
• PERKENALAN
• PENDAHULUAN • PENGENALAN INDERAJA • APLIKASI INDERAJA
•KESIMPULAN • DISKUSI
KESIMPULAN Teknologi Inderaja terbukti handal digunakan untuk aplikasi pertanian, yaitu untuk prediksi terkini tanaman padi dan prediksi panen
Kolaborasi metode machine learning ke dalam pemodelan remote sensing, dapat menghasilkan model yang lebih akurat Teknologi inderaja dapat memberikan informasi jadual pengairan, jadual pemupukan, serta fenomena serangan hama dan penyakit
• PERKENALAN
• PENDAHULUAN • PENGENALAN INDERAJA • APLIKASI INDERAJA • KESIMPULAN
•DISKUSI
Salamat
謝謝
