Syllabus

Geospatial Artificial Intelligence (GeoAI) for Oil Palm Tree Counting and Health Assessment

🕘 Duration: 2 Days (16 Hours)

👥 Level: Intermediate to Advanced

Pelatihan ini bertujuan meningkatkan kapasitas peserta dalam menerapkan teknologi Geospatial Artificial Intelligence (GeoAI) untuk melakukan penghitungan jumlah pohon kelapa sawit dan penilaian kesehatan tanaman secara akurat, efisien, dan berbasis data spasial guna mendukung pengelolaan perkebunan yang berkelanjutan dan berbasis presisi. Download Syllabus.

💼 Target Participants:

  • Geospatial practitioners,
  • Agricultural/plantation researchers,
  • Data analysts,
  • Master’s/PhD students,
  • Technical staff in palm oil companies.

📋Method: Combination of lectures, hands-on practices, and case studies.

📚 Day 1 – Session 1: Introduction and Fundamentals of GeoAI

Topics:

  • Overview of GeoAI and Its Role in the Oil Palm Industry.

  • Remote Sensing for Agriculture: UAV, Satellite, and Multispectral Data.

  • Introduction to Monitoring Needs in Oil Palm Plantations.

Learning Objectives:

  • Understand basic concepts of GeoAI and its applications in the palm oil sector.

  • Identify geospatial data sources (UAV, Sentinel-2, PlanetScope, etc.).

  • Explain business and technical requirements for monitoring oil palm plantations.

📚 Day 1 – Session 2: Data Acquisition and Preprocessing

Topics:

  • UAV Mission Planning and Data Collection.

  • Orthomosaic Generation and Georeferencing.

  • Satellite Imagery Preprocessing.

  • Annotation and Labeling Techniques for AI Training.

Learning Objectives:

  • Conduct UAV data acquisition for oil palm plantations.

  • Generate orthomosaic maps and perform geometric corrections.

  • Prepare imagery datasets for AI model training.

  • Use QGIS and LabelImg/Roboflow for data annotation and labeling.

📚 Day 2 – Session 1: AI for Oil Palm Tree Counting

Topics:

  • Deep Learning Fundamentals for Object Detection.

  • CNN and YOLOv5/YOLOv8 for Tree Counting.

  • Transfer Learning and Model Evaluation.

  • Practice: Train a Model Using Oil Palm Dataset.

Learning Objectives:

  • Understand object detection algorithms (YOLO, Faster R-CNN) for palm tree identification.

  • Train and evaluate AI models using UAV or satellite datasets.

  • Perform automated tree counting and validate results with field data.

📚 Day 2 – Session 2: Tree Health Assessment and Classification

Topics:

  • Vegetation Indices (NDVI, GNDVI, SAVI) and Tree Health Indicators

  • Semantic Segmentation with U-Net/DeepLab

  • Integrating Spectral and Textural Features

  • Case Study: Health Mapping in Oil Palm Plantations

Learning Objectives:

  • Utilize vegetation indices to assess oil palm tree health.

  • Train image segmentation models for classifying healthy/unhealthy areas.

  • Analyze classification results and validate with field or high-resolution data.

📚Day 2 – Session 3: GeoAI Dashboard & Deployment

Topics:

  • Spatial Analysis with Python (geopandas, rasterio, shapely)

  • Building a Simple GeoAI Dashboard (Streamlit or Dash)

  • Model Deployment and GIS Platform Integration

  • Group Project Presentation & Evaluation

Learning Objectives:

  • Integrate AI output with spatial data and map visualization.

  • Develop a dashboard prototype for monitoring oil palm plantations.

  • Design an end-to-end workflow from data acquisition to actionable insights.

  • Present project results and receive feedback from instructors.

Evaluation and Certification

  • Daily quizzes

  • Mini-project (individual or group)

  • Certificate awarded to participants with ≥80% attendance and who pass the final evaluation

Learning Outcomes:

By the end of the training, participants will be able to:

  • Understand the foundational concepts and benefits of GeoAI in the oil palm industry

  • Perform data acquisition, preprocessing, and analysis from UAVs, satellite imagery, and multispectral sensors

  • Apply deep learning models (YOLO, U-Net) to automatically count palm trees and assess their health

  • Integrate AI analysis with Geographic Information Systems (GIS)

  • Evaluate and validate model results using field data or high-resolution images

  • Apply an end-to-end pipeline from UAV mission planning to reporting insights for plantation management

Contact Information:
📧 wbudiharto@binus.edu
📱 Prof. Widodo (WA: +62 856-9887-384)
📱 Dr. Emny (WA: +62 813-8741-3863)

Silabus Pelatihan

Geospatial Artificial Intelligence (GeoAI) for Oil Palm Tree Counting and Health Assessment

🕘 Durasi: 2 Hari (16 jam)

👥Level: Intermediate to Advanced

💼  Metode: Teori (kuliah), praktik langsung (hands-on), dan studi kasus interaktif

📋Target Peserta:

  • Praktisi geospasial

  • Peneliti di bidang pertanian/perkebunan

  • Analis data

  • Mahasiswa S2/S3

  • Staf teknis perusahaan kelapa sawit

📚Hari 1 – Fondasi GeoAI dan Pengolahan Data Geospasial

Sesi 1: Pengenalan dan Dasar-dasar GeoAI

Topik:

  • Pengantar GeoAI dan perannya dalam industri kelapa sawit

  • Penginderaan jauh untuk pertanian: UAV, citra satelit, data multispektral

  • Kebutuhan pemantauan perkebunan kelapa sawit secara digital

Learning Outcomes:

  • Memahami konsep dan manfaat GeoAI

  • Mengenali berbagai sumber data geospasial (UAV, Sentinel-2, PlanetScope)

  • Menjelaskan kebutuhan teknis dan bisnis untuk pemantauan kelapa sawit

Sesi 2: Akuisisi dan Praproses Data

Topik:

  • Perencanaan misi UAV dan pengambilan data

  • Generasi orthomosaic dan georeferensi

  • Praproses citra satelit

  • Teknik anotasi dan labeling untuk pelatihan AI

Learning Outcomes:

  • Melakukan akuisisi data UAV secara efisien

  • Menghasilkan orthomosaic dan koreksi geometrik

  • Menyiapkan dataset untuk pelatihan model AI

  • Melakukan labeling menggunakan QGIS, LabelImg, atau Roboflow

📚 Hari 2 – Pemodelan AI dan Implementasi Lapangan

Sesi 1: Pendeteksian Pohon Kelapa Sawit dengan AI

Topik:

  • Dasar Deep Learning untuk Object Detection

  • Penerapan CNN dan YOLOv5/YOLOv8

  • Transfer learning dan evaluasi model

  • Latihan: Melatih model deteksi pohon

Learning Outcomes:

  • Memahami model deteksi objek (YOLO, Faster R-CNN)

  • Melatih model berbasis UAV/satelit untuk menghitung pohon

  • Melakukan perbandingan hasil deteksi dengan data lapangan

Sesi 2: Penilaian Kesehatan Pohon dan Klasifikasi

Topik:

  • Indeks vegetasi (NDVI, GNDVI, SAVI)

  • Segmentasi semantik dengan U-Net / DeepLab

  • Integrasi fitur spektral dan tekstural

  • Studi kasus: Health Mapping di perkebunan

Learning Outcomes:

  • Menggunakan indeks vegetasi untuk klasifikasi kesehatan pohon

  • Melatih dan menguji model segmentasi citra

  • Menganalisis hasil klasifikasi dan validasi lapangan

Sesi 3: Dashboard GeoAI dan Implementasi Sistem

Topik:

  • Analisis spasial dengan Python (GeoPandas, Rasterio, Shapely)

  • Membangun dashboard dengan Streamlit atau Dash

  • Integrasi model AI ke platform GIS

  • Presentasi proyek kelompok

Learning Outcomes:

  • Menggabungkan hasil AI dengan visualisasi spasial.

  • Membangun prototipe dashboard monitoring.

  • Menyusun workflow end-to-end dari data ke insight.

  • Mendapatkan umpan balik atas hasil proyek.

Evaluasi dan Sertifikasi

  • Kuis harian

  • Mini project (individu/kelompok)

  • Sertifikat diberikan kepada peserta yang:

    • Mengikuti ≥80% sesi

    • Lulus evaluasi akhir

Hasil Akhir Pelatihan

Setelah mengikuti pelatihan ini, peserta diharapkan mampu:

  • Memahami peran dan manfaat GeoAI dalam pengelolaan kelapa sawit.

  • Melakukan akuisisi, praproses, dan analisis data geospasial dari UAV & citra satelit.

  • Menerapkan model deep learning seperti YOLO dan U-Net untuk menghitung pohon dan menilai kesehatannya.

  • Mengintegrasikan hasil AI ke dalam sistem GIS.

  • Melakukan evaluasi dan validasi berbasis data lapangan.

  • Menyusun pipeline GeoAI end-to-end: dari akuisisi hingga pelaporan.

Kontak Informasi

📧 Email: wbudiharto@binus.edu
📱 Prof. Widodo (WA): 0856-9887-384
📱 Dr. Emny (WA): 0813-8741-3863