LAND COVER CLASSIFICATION 2010 Ministry of Water & Environment, National Forestry Authority, by Edward Ssenyonjo Remote Sensing Specialist Joseph Mutyaba.

Slides:



Advertisements
Similar presentations
WFM 6202: Remote Sensing and GIS in Water Management
Advertisements

Desktop, Mobile & Web Based GIS/ Collaborative GIS
Starting from Text Files. User Interface Graphic user interface –Mouse, buttons, windows, menus Commands-driven interface –Commands, parameters.
Copyright, © Qiming Zhou GEOG1150. Cartography Data Models for Computer Cartography.
Topographic mapping in Fiji: Challenges and opportunities Conway Pene 2012 Pacific GIS&RS Conference November 2012, Suva.
WFM 6202: Remote Sensing and GIS in Water Management © Dr. Akm Saiful IslamDr. Akm Saiful Islam WFM 6202: Remote Sensing and GIS in Water Management Akm.
ESRM 250 & CFR 520: Introduction to GIS © Phil Hurvitz, KEEP THIS TEXT BOX this slide includes some ESRI fonts. when you save this presentation,
What’s New in ArcGIS 9.3 and SchoolSite 9.3. A Summary of All New Features ArcGIS 9.3
1 CPSC 695 Data Quality Issues M. L. Gavrilova. 2 Decisions…
Vector-Based GIS Data Processing Chapter 6. Vector Data Model Feature Classes points lines polygons Layers limited to one class of data Figure p. 186.
GIS Overview. What is GIS? GIS is an information system that allows for capture, storage, retrieval, analysis and display of spatial data.
Agenda Overview Why TransCAD Challenges/tips Initiatives Applications.
WFM 6202: Remote Sensing and GIS in Water Management © Dr. Akm Saiful IslamDr. Akm Saiful Islam WFM 6202: Remote Sensing and GIS in Water Management Dr.
GIS 200 Introduction to GIS Buildings. Poly Streams, Line Wells, Point Roads, Line Zoning,Poly MAP SHEETS.
GIS Tutorial 1 Lecture 6 Digitizing.
19 th Advanced Summer School in Regional Science Overview and more advanced directions with ArcGIS.
Geography 241 – GIS I Dr. Patrick McHaffie Associate Professor Department of Geography Cook County, % population < 5.
Dr. David Liu Objectives  Understand what a GIS is  Understand how a GIS functions  Spatial data representation  GIS application.
Intro to ArcGIS Kate Dougherty, Geosciences & Maps Librarian Tutorial created for version 10.1 in February 2013.
@ 2007 Austin Troy. Geoprocessing Introduction to GIS Geoprocessing is the processing of geographic information. Perform spatial analysis and modeling.
Has EO found its customers? 1 Space Applications Institute Directorate General Joint Research Centre European Commission Ispra (VA), Italy
Basic Spatial Analysis
Habitat Analysis in ArcGIS Use of Spatial Analysis to characterize used resources Thomas Bonnot
Module 2.1 Monitoring activity data for forests using remote sensing REDD+ training materials by GOFC-GOLD, Wageningen University, World Bank FCPF 1 Module.
Esri UC 2014 | Technical Workshop | Data Alignment and Management in ArcMap Lisa Stanners, Sean Jones.
GNORASI vision and achievements, Future perspectives Panagiotis Symeonidis Environmental Physicist, M.Sc., Ph.D. Technical Director DRAXIS Environmental.
Applied Cartography and Introduction to GIS GEOG 2017 EL Lecture-3 Chapters 5 and 6.
ESRM 250 & CFR 520: Introduction to GIS © Phil Hurvitz, KEEP THIS TEXT BOX this slide includes some ESRI fonts. when you save this presentation,
Introduction to InVEST ArcGIS Tool Nasser Olwero GMP, Bangkok April
Understanding Topology for Soil Survey
International Institute for Geo-Information Science and Earth Observation (ITC) ISL 2004 RiskCity Exercise 5: Generating an elements at risk database Cees.
By: GeoTrek. Hunter Krenek: Remote Sensing analyst & GIS analyst Joe Dowling: Assistant Project Manager & GIS analyst Peter Vogt: Website Designer & GIS.
By Phoebe Oduor.  Countries are required to report on their GHG emissions to UNFCCC every four years (mandatory) and every two years (intermediary) 
Optimal use of new satellite resources. Research funded by NERC/CEH and JNCC. Rapid Land Cover Mapping.
Virtual Ice Charting System Archive Browser Interface Distribution IngestProduction Ice Analyst Application Database Click on the boxes for more information.
Aseri Baleilevuka OCEANS & ISLANDS PROGRAM SOPAC-SPC Benthic Habitat Mapping Lifuka Island.
ArcGIS Network Analyst: Automating Workflows with Geoprocessing
GEO ADC meeting, Sep 22-23, 2008 Slide # 1 AG AG Forest Mapping and Change Monitoring LoycheWilkie,Mette (FAO) Adam Gerrand (FAO) Ivan Petiteville,
Best Practices for Managing Historical Imagery Cody Benkelman Kumar Dhruv.
SDCG-5 / FAO Rome 25 th February Ministry of Water & Environment, National Forestry Authority, by Edward Ssenyonjo Remote Sensing Specialist Joseph.
Environmental Modeling Testing GIS Layer Relevancy.
International Institute for Geo-Information Science and Earth Observation (ITC) ISL 2004 RiskCity Exercise 4: Generating an elements at risk database Cees.
Preparing and Deploying Data to ArcPad Juan Luera.
Ministry of Agriculture, Livestock and Food Suply Embrapa Satellite Monitoring Assessing vegetation coverage at the Sao Paulo State scale: a tool for aiding.
Sea Ice Mapping Systems Archive Browser Interface Distribution IngestProduction Ice Analyst Application Database Henrik Steen AndersonDMI Paul SeymourNIC.
Land Cover Change Monitoring change over time Ned Horning Director of Applied Biodiversity Informatics
Land Cover Change Monitoring change over time Ned Horning Director of Applied Biodiversity Informatics
Data Creation and Editing Based in part on notes by Prof. Joseph Ferreira and Michael Flaxman Lulu Xue | Nov. 3, :A Workshop on Geographical.
Advanced GIS Using ESRI ArcGIS 9.3 Spatial Analyst 2.
GIS Data Types. GIS technology utilizes two basic types of data 1. Spatial Data Describes the absolute and relative location of geographic features.
474/574 Remote sensing Lab: Digitizing Sweet water Wetland Park areas, Tucson AZ by using remote sensing imagery (Aerial and Landsat) and ArcGIS software.
Supervised Classification in Imagine D. Meyer E. Wood
USE OF AERIAL FRAMES FOR AGRICULTURAL CENSUSES AND SURVEYS Fiji Experience.
WFM 6202: Remote Sensing and GIS in Water Management © Dr. Akm Saiful IslamDr. Akm Saiful Islam WFM 6202: Remote Sensing and GIS in Water Management Dr.
Data Entry Getting coordinates and attributes into our GIS.
Updated Cover Type Map of Cloquet Forestry Center For Continuous Forest Inventory.
-gSSURGO- Using the Soil Data Management Toolbox Steve Peaslee USDA-NRCS National Soil Survey Center Lincoln, Nebraska March.
1 Creating Situational Awareness with Data Trending and Monitoring Zhenping Li, J.P. Douglas, and Ken. Mitchell Arctic Slope Technical Services.
Introduction to Geodatabases
US Croplands Richard Massey Dr Teki Sankey. Objectives 1.Classify annual cropland extent, Rainfed-Irrigated, and crop types for the US at 250m resolution.
Lecture 18: Spatial Analysis Using Rasters Jeffery S. Horsburgh CEE 5190/6190 Geographic Information Systems for Civil Engineers Spring 2016.
Geospatial Data Abstraction Library(GDAL) Sabya Sachi.
Introduction to InVEST ArcGIS Tool
What’s new in FUSION? Bob McGaughey
Using Model Builder in ArcGIS
Monitoring Surface Area Change in Iowa's Water Bodies
Evaluating Land-Use Classification Methodology Using Landsat Imagery
GIS Lecture: Editing Data
Vector Geoprocessing.
Use of Geospatial Data for SDG Monitoring
Presentation transcript:

LAND COVER CLASSIFICATION 2010 Ministry of Water & Environment, National Forestry Authority, by Edward Ssenyonjo Remote Sensing Specialist Joseph Mutyaba Cartography Specialist 25 February 2014 FAO-Rome, Italy

Where we have come from Land cover 2005 (Before FRA trianing,02/13) Heads up digitising Time consuming 2004 – 2009, Very manual Elements of human bias

Where we are Land cover 2010 (After FRA trianing,02/13) Computer aided digitizing Time saving: April 2013 – June 2014 Automated digitizing & classification Assumed objectivity Segmentation graphic

Where we want to go Land cover 2015 (in collaboration FRA) Computer aided digitizing Saving more time: < 3 months Sharable on-line and interactively eg CartoDB, MapBox etc

Working with FRA There are 17 global FRA samples (10x10km) in Uganda Uganda is covered by 520 national tiles generated prior to the Feb 2013 training (20x20km). The 520 were adjusted to 312 historical tiles by Dept of Lands & Surveys (27x27) During FRA Training: The Ugandan team was introduced to Open Foris Geo-spatial Toolkit (OFGT) OFGT is command line and script based (R, perl etc) Semi-automated (both segmentation & classification) and Very fast (on Ubuntu OS) The resultant land cover map & statistics concentrated on FAO FRA classes forests F, OWL, OL, H 2 0 Uganda would like to include other classes(=13) So far OFGT is being used for data preparation i.e Image stacking and sub setting

Erik and Remi with the Ugandan team during training Feb 2013

The team with trainers

Segments from Open Foris Tool Kit

IMAGE SEGMENTATION WITH ORFEO TOOLBOX (OTB) OTB interface work on Windows, MacOS and Linux operating systems. It too, is free Open Source software supported by the OSGeo4W Has many modules –the segmentation module was used with the mean shift algorithms. It has an advantage of producing very neat and interactive vector segments based on spatial radius, range radius and region size.

IMAGE CLASSIFICATION WITH R ® After segmenting the raster with OTB, an R script is used to calculate segment statistics using the raw image stack. a set of training points chosen from the image in form of a shapefile or x, y data (in a csv or dbf table). These statistics and training data are then used in another R script to classify the segments using the Random forest module so far up to 10 classes. Note: The scripts were written by Ned Horning and Wynne Walker from Woods Hole Research Centre, MA.

Raw Image Vector segments Raster segments Segment statistics Classifying segments Training data Segment selection Qualifying segments Selecting training samples Raw image as statistics input Satellite Image to Land cover stratification Training required for classification Input Output Classified Image

Segments from Orfeo Tool Box

OTB & OFT Segments Compared OFGT OTB

VALIDATION WITH ArcMap First step of adding relevant fields is done using model builder in ArcGIS. The fields are populated in ArcMap This where the 10 classes are translated to 13 IdRclass 1Forest 2Wetland 3Grass 4Bush 5Agriculture 6Towns 7Water 8Impediments 9Cloud 10Cloud shadow VclassNBS BTP1 CP2 THFW3 THFL4 Wo5 B6 G7 We8 SF9 CF10 BA11 OW12 I13

RESULTS &STATISTICS (Achievements) So far classified 94.55% of the country94.55% Generated statistics for 81% of the 312 tiles81% Validated 14% of the tiles14%

STATISTICS Analysis Classified versus Validated error The importance of validation Example of classified imageclassified image Example of validated imagevalidated image Study the trend in classification error from R to Validation (Ground truthing)

CHALLENGES Steep Learning curve Cloud cover Topological (overlapping polygons & gaps) Difficulty in comparison of 2010 LC with previous LC datasets due to the generalization of earlier datasets. Ability to edit rasters attributes to minimize usage of vectors and there topological issues

SUMMARY No silver bullet Data Preparation SegmentationClassificationValidation Image stacking & Compositing Sub setting Result is either raster or vector R scripting (Statistics calculation) Classification Table schema (Model builder) Attribute editing OFGTOTBRArcGIS Stages Tools Ops

Grazie !!