Cranfield University, Silsoe MSc Thesis Graeme Gould Digital

Cranfield University, Silsoe National Soil Resources Institute

MSc Thesis Academic Year: 2005/2006 Graeme Gould

Digital Field Data Collection for Land Cover Survey

Supervisors: Christophe Sannier & Tim Brewer Word Length: 8209 Date of Presentation: 23 August 2006

This thesis is submitted in partial fulfilment of the requirements for the Degree of Masters of Geographical Information Management. © Cranfield University, 2006. All rights reserved. No part of this publication may be reproduced without the written permission of the copyright holder.

BLANK PAGE

Cranfield University, Silsoe

G.R.A. Gould, 2006

ii

ABSTRACT

This thesis considers data collection methods for land classification surveys and provides justification for the development of a project specific data entry application. The thesis looks initially at the LUCAS survey system and compares the methodologies of 2001/3 and 2005/6. This identifies differences in the method structure and class nomenclature between the two dates. A classification system based on LUCAS which incorporates aspects of both the 2001/3 and 2005/6 was decided on and provides the basis for further work. The process of data recording during field surveys is assessed. Traditional, written data entry is compared to digital data collection. This indicates that the benefits of digital methods far outweigh those of paper based methods. Having determined that a digital data recording system is required a comparison is made of those most commonly available.

The two assessed are ArcPad and

TerraSync, two of the market leaders. Each is assessed against the requirements of the project brief and LUCAS and it was concluded that neither fully met the exact requirements. Both software applications had advantages but fell short in their limited data entry abilities and therefore suitability for the task. It was concluded that a project specific application that met the requirements of LUCAS and the project brief should be developed. The design, development and testing of an application which can be installed on a range of mobile devices, including those with GPS, is detailed in full. Specific areas of interest such as development and coding of a GPS interface are highlighted as core components of the application. Testing the application in the field provided evidence as to its functionality and operability as well as highlighting areas that can be improved on in the future.

Cranfield University, Silsoe

G.R.A. Gould, 2006 iii

ACKNOWLEDGEMENTS

Thanks are given to Elizabeth Farmer, Andrew Rayner, and Steve Hallett for their support, understanding and guidance through out this thesis. Particular thanks are given to Pascal Jacques from the LUCAS project for his assistance regarding the LUCAS system. Finally, special thanks go to Juliet Stokoe for her continued support though the duration on the project.

Cranfield University, Silsoe

G.R.A. Gould, 2006

iv

CONTENTS

1.

2.

Page

ABSTRACT

iii

ACKNOWLEDGEMENTS

iv

CONTENTS

v

LIST OF FIGURE

viii

LIST OF TABLES

ix

ABBREVIATIONS

x

INTRODUCTION

1

1.1

Project Background

1

1.2

Aims and Objectives 1.2.1 Aim 1.2.2 Objectives

2 2 2

1.3

Methodology 1.3.1 Classification System 1.3.2 GPS software 1.3.3 Software development

2 2 3 4

BACKGROUND RESEARCH

6

2.1

Introduction

6

2.2

Classification System – LUCAS 2.2.1 Background 2.2.2 LUCAS Objectives 2.2.3 2001/03 Structure 2.2.4 2005/06 LUCAS 2.2.5 Chosen LUCAS Methodology

6 6 6 7 9 10

2.3

Why do digital surveys? 2.3.1 Advantages 2.3.2 Disadvantages

11 11 12

2.4

Application Requirements

13

2.5

Review of existing software applications 2.5.1 Trimble TerraSync Professional 2.5.2 ESRI ArcPad 2.5.3 Summary

14 14 19 21

Cranfield University, Silsoe

G.R.A. Gould, 2006

v

CONTENTS 3. DESIGN

4.

5.

6.

Page 23

3.1

Introduction

23

3.2

Use Cases 3.2.1 Core Application 3.2.2 GPS Interface

23 23 25

3.3

Use Case Diagrams

26

3.4

Data Flow Diagrams

27

3.5

Database Design

30

3.6

Conclusion

31

DEVELOPMENT

32

4.1

Introduction

32

4.2

Programming Language

32

4.3

Additional Requirements

32

4.4

Application Development 4.4.1 The GUI 4.4.2 The GPS Interface 4.4.3 The Database application

33 33 35 37

4.5

Installation/Deployment

39

FIELD TESTING

40

5.1

Introduction

40

5.2

Methodology

40

5.3

Results

41

5.4

Discussion

43

5.5

Conclusion

43

FUTURE DEVELOPMENTS

44

6.1

Introduction

44

6.2

Classification Systems

44

6.3

Survey Methodology

44

Cranfield University, Silsoe

G.R.A. Gould, 2006

vi

CONTENTS 6.4 Software

Page 45

7.

REFERENCES

46

8.

APPENDICES

48

8.1

Tender Terms of Agreement (Sannier, 2006)

48

8.2

Classification Nomenclature 8.2.1 Land Cover 8.2.2 Land Use

50 50 51

8.3

Use Cases

52

8.4

Project Database Field structure

54

8.5

Classification Database Design

56

8.6

Data Entry Forms

57

8.7

Full Application Coding

58

Cranfield University, Silsoe

G.R.A. Gould, 2006 vii

LIST OF FIGURES

Page

Figure 1.1 The four RUP phases to developing a system

4

Figure 1.2 The six RUP iterations

5

Figure 2.1 Primary Sample Units

8

Figure 2.2 Structure of Secondary Sampling Units

8

Figure 2.3 Map Display Section

15

Figure 2.4 Data Collection Screen

16

Figure 2.5 Navigation Display Window

16

Figure 2.6 GPS Status Screen

17

Figure 2.7 GPS Setup Interface

17

Figure 2.8 ArcPad Main Screen

20

Figure 3.1 Core Application Use Case diagram

26

Figure 3.2 GPS Interface Use Case diagram

27

Figure 3.3 Summary Data Flow Diagram

28

Figure 3.4 Detailed DFD for GPS interface

29

Figure 3.5 Project database design.

31

Figure 4.1 Main options form

34

Figure 4.2 Land Cover Selection forms

34

Figure 4.3 Finish screen and options

35

Figure 4.4 Sample of code from the GPS interface.

36

Figure 4.5 Sample of code for adding the Land Cover reference to the database

38

Figure 5.1 Field test sample points

40

Cranfield University, Silsoe

G.R.A. Gould, 2006 viii

LIST OF TABLES

Page

Table 2.1 Advantages and Disadvantages of TerraSync Professional

18

Table 2.2 ArcPad Advantages and Disadvantages

21

Table 2.3 Comparison of ArcPad & TerraSync Professional

22

Table 5.1 General Information database table

41

Table 5.2 Land Cover database table

42

Table 5.3 Positional differences between the new application and TerraSync

42

Cranfield University, Silsoe

G.R.A. Gould, 2006

ix

ABREVIATIONS CAB

Cabinet File

CDB

Compact DataBase

DGPS

Differential Global Positioning System

DLL

Dynamic Link Library File

EGNOS

European Geostationary Navigation Overlay System

EUROSTAT Statistical Office of the European Community FAO

Food and Agriculture Organisation

GPS

Global Positioning System

GUI

Graphical User Interface

HDOP

Horizontal Dilution of Precision

LADA

Land Degradation Assessment in Drylands

LCCS

Land Cover Classification System

LUCAS

Land Use/Cover Area Frame Statistical Survey

NMEA

National Marine Electronics Association

PDOP

Position Dilution of Precision

PSU

Primary Sampling Unit

RUP

Rational Unified Process

SSF

Standard Storage Format

SSU

Secondary Sampling Unit

UML

Unified Modelling Language

VDOP

Vertical Dilution of Precision

WAAS

Wide Area Augmentation System

Cranfield University, Silsoe

G.R.A. Gould, 2006

x

1.

INTRODUCTION

1.1

Project Background The Food and Agriculture Organisation (FAO) of the United Nations, in particular the Land and Water Development Division (AGL) have expressed a need to understand in more detail land use and land cover. Understanding these changes within developing countries, especially those in Africa is of specific interest to the FAO. Changes in land cover and land use have been identified as one of the four primary global environmental problems and understanding them is crucial to comprehending environmental change in the future (Jansen and Di Gregorio, 2003).

Comprehension of environmental

change provides evidence to the FAO and local governments that aids the decision making process for managing the environment and natural resources Land cover and land use can be defined as follows: •

Land cover is the physical cover that can be observed on the surface of the earth (Di Gregorio and Jansen, 2000)

•

Land use is the manner in which these biophysical assets are used by humans (Cihlar and Jansen, 2001)

There is currently a shortage of global to regional land use information based on actual observations from the field. Increasing our knowledge of land cover and land use practices provides a way of aiding the definition of methods, such as policy formation, for improving how land is managed. A major effect that land use activities have on the environment is land degradation. As well as aiding policy formation land cover and use surveys can be used in the assessment and formulation of remedial measures in land degradation studies. One such study is the Land Degradation Assessment in Drylands (LADA) which aims to assess the causes, status and impact of land degradation to improve decision making for sustainable development (FAO, 2006).

Cranfield University, Silsoe

G.R.A. Gould, 2006

1

This project is based on a tender from the FAO for the development of a prototype field data entry tool for collecting data during land use surveys. The tender requirements (Appendix 8.1) include the identification of a specific land cover/use classification system, the review of existing field survey data collection software and the development of a custom data entry form that integrates with a GPS receiver. 1.2

Aims and Objectives 1.2.1 Aim The development of a reliable, easy to use and cost effective system for collecting accurate land classification information during field surveys. 1.2.2 Objectives •

Identification of a land cover and land use classification system.

•

The evaluation of available field survey data collection applications in their ability to meet the requirements of the FAO project.

•

Development and testing of a user interface for use on a GPS enabling the accurate collection of land cover and land use data.

1.3

Methodology 1.3.1 Classification System For the proposed feasibility study the FAO have requested that a specific classification scheme be adopted. This classification scheme is expected to be based around the LUCAS methodology. LUCAS is used for recording both land use and land cover within Europe.

Different environmental and climatic conditions as well as

different land use in Africa mean that the existing nomenclature may not be suitable and changes would be required. One way to achieve this would be to adjust the classification system to include land cover and land use classes found in Africa. This would require additional work to determine the new classes. A second method would be to utilise a land cover classification system that has been designed to encompass all land cover types from around

Cranfield University, Silsoe

G.R.A. Gould, 2006

2

the world and can be focused to a specific country.

The FAO has

developed its Land Cover Classification Scheme (LCCS) to be universally applicable, multipurpose and systematic (FAO-Africover, 2003). The limiting factor with LCCS is that it only concerns land cover – it does not include land use, and as such an appropriate land use classification scheme will have to be decided on. For the purposes of development and testing the project will use the LUCAS methodology and classification system. For the land cover and use classes the existing European nomenclature will be used.

1.3.2 GPS software The feasibility study will use the Trimble GeoXT Global Positioning System (GPS) receiver and mapping unit. This Windows CE device allows the installation of a variety of off the shelf survey packages and purpose developed applications.

Devices without Windows CE and

which do not allow installation of standalone software are discounted as they do not allow the installation of custom designed applications and would not be able to meet the data entry requirements of LUCAS. A comparison and assessment of widely available software packages will be made. The aim of this will be to ascertain whether they meet the FAO tender requirements, the data entry requirements of LUCAS, and whether any add-ons can be developed to improve the data entry requirements. The results of this comparison will indicate whether an existing application can be used or whether a new project specific application should be developed. The software packages to be compared are: •

ArcPad

•

TerraSync

Cranfield University, Silsoe

G.R.A. Gould, 2006

3

The comparison will be made using a set of requirements (Section 2.4) based on those of LUCAS and those set out in the terms of reference for the FAO project (Appendix 8.1).

1.3.3 Software development Development of a software application is fully dependent on the outcome of the GPS software comparison and assessment indicating that a project specific application would be the most appropriate method of meeting the project requirements. Development of the application would take place in two phases; the initial planning and design phase will detail what is required of the program, how it will work and what the expected outputs will be. The second phase will be the building and testing of the application. To guide the development of the application the Rational Unified Process (RUP) development process will be used. RUP is a specialised version of the Unified Process which is a set of activities that are performed to create a software system from a set of requirements (Scott. 2002). RUP is an iterative development process developed around the Unified Modelling Language (UML) and is split into four phases (Figure 1.1).

•

Inception Phase

•

Elaboration Phase

•

Construction Phase

•

Transition Phase

Each with objectives

Source: Cranfield University (Hallett, 2005) Figure 1.1 The four RUP phases to developing a system

Cranfield University, Silsoe

G.R.A. Gould, 2006

4

Each phase is divided into six iterations (Figure 1.2) that are worked through in the design, development, testing and implementation of an application. The outcome from each phase is an artefact(s); with artefacts produced the next phase can be started.

Business Modelling Requirements Analysis and Design Implementation

Workflow Disciplines

Testing Deployment Source: Cranfield University (Hallett, 2005) Figure 1.2 The six RUP iterations Progressing through all of the phases results in a final end product that meets the requirements identified in the initial inception phase.

Cranfield University, Silsoe

G.R.A. Gould, 2006

5

2.

BACKGROUND RESEARCH

2.1

Introduction The background research for this project looks at two project objectives and provides justification and structure for later work. The two objectives are the identification of appropriate classification systems and the evaluation of currently available field data collection applications. This chapter looks at the LUCAS

classification

system,

comparing

the

original

and

current

methodologies; an evaluation is carried out on the two existing software applications and justification made for carrying out electronic data collection over traditional paper based data collection. 2.2

Classification System – LUCAS 2.2.1 Background The LUCAS project was launched by the Statistical Office of the European Community (EuroStat) in conjunction with the Directorate General responsible for Agriculture as a support to policy formulation within agriculture. It is based “on the application of area-frame survey and remote sensing techniques to the agricultural statistics for 1999 to 2003”. (Bertin, 2003) 2.2.2 LUCAS Objectives Bertin (2003) identifies the main objectives of LUCAS as: •

The

harmonisation

of

data,

through

a standard

survey

methodology of the main land cover/use areas and changes; •

To extend the scope of the survey to include environment, multifunctionality, landscape and sustainable development as well as agriculture;

•

Offer a common sampling base (frame, nomenclature, data treatment); and

•

To evaluate the point area frame survey methodology

Cranfield University, Silsoe

G.R.A. Gould, 2006

6

LUCAS has continuously developed since its inception. These developments

are

detailed

in

the

following

sections

and

recommendations are made concerning how to implement LUCAS in this study. Although land cover and land use classifications have been developed and refined they are still aimed at the European environment. To enable the use of LUCAS in Africa the class nomenclature would need modifying or replacing with an alternative classification system such as LCCS for use within Africa. 2.2.3 2001/03 Structure The original methodology used in 2001 and 2003 split the survey into two phases: •

In phase 1 land cover/use data was collected during the spring time

•

In phase 2, additional information was collected during the autumn through the interviews with farmers. (Avikainen et al, 2003)

Phase 1 was carried out using a 2 stage systematic sampling design: primary sampling unit (PSU) and secondary sampling unit (SSU). The aim of which was to record land cover and land use at a sample point. The initial stage was to generate an 18km by 18km grid (Figure 2.1). The PSU’s were generated at the intersections of this grid.

The

sampling was planned on a country by country basis. The second stage identifies 10 SSU’s positioned around the centre of the PSU (Figure 2.2). The SSU’s are points of observation 300m apart and located in 2 lines in an East-West direction (Avikainen et al, 2003).

Cranfield University, Silsoe

G.R.A. Gould, 2006

7

Source: EuroStat (Avikainen et al, 2003) Figure 2.1 Primary Sample Units ⇑N

1 1

300m

2

300m

3

300m

4

300m