: Paolo Remagnino, Simon Mayo, Paul Wilkin, James Cope, Don Kirkup
: Computational Botany Methods for Automated Species Identification
: Springer-Verlag
: 9783662537459
: 1
: CHF 114.00
:
: Allgemeines, Lexika
: English
: 117
: Wasserzeichen/DRM
: PC/MAC/eReader/Tablet
: PDF
This book discusses innovative methods for mining information from images of plants, especially leaves, and highlights the diagnostic features that can be implemented in fully automatic systems for identifying plant species. Adopting a multidisciplinary approach, it explores the problem of plant species identification, covering both the concepts of taxonomy and morphology. It then provides an overview of morphometrics, including the historical background and the main steps in the morphometric analysis of leaves together with a number of applications. The core of the book focuses on novel diagnostic methods for plant species identification developed from a computer scientist's perspective. It then concludes with a chapter on the characterization of botanists' visions, which highlights important cognitive aspects that can be implemented in a computer system to more accurately replicate the human expert's fixation process. The book not only represents an authoritative guide to advanced computational tools for plant identification, but provides experts in botany, computer science and pattern recognition with new ideas and challenges. As such it is expected to foster both closer collaborations and further technological developments in the emerging field of automatic plant identification.

Contents6
1 Introduction ??????????????8
1.1 Plant Species and Their Identification8
1.2 The Delimitation of Species by Descriptions12
1.3 Using Leaves to Diagnose Species13
1.4 Computational Botany15
1.5 Aims and Objectives16
2 Morphometrics: A Brief Review ??????????????17
2.1 Historical Background to Morphometrics18
2.1.1 Phase 1: Classical Taxonomy, Evolutionary Theory and Biometry: The Background for Morphometrics18
2.1.2 Phase 2: Genetics and Statistical Methods in Evolution, Agronomy and Biosystematics20
2.1.3 Phase 3: Multivariate Statistics and Morphometrics21
2.1.4 Phase 4: Geometric Morphometrics23
2.2 Morphometric Analysis of Leaves25
2.2.1 Analysis of Conventional Botanical Descriptors26
2.2.2 Analysis of Leaf Outline Shape27
2.2.3 Analysis of Leaf Margin Patterns29
2.2.4 Analysis of Homologous Landmarks and Semi-landmark Configurations29
2.2.5 Fractal Dimensions and Polygon Fitting30
2.2.6 Analysis of Leaf Venation Patterns30
2.2.7 Analysis of Leaf Texture31
2.2.8 Analysis of Other Features of the Leaf Blade31
2.3 Morphometrics of Flowers and Other Plant Organs31
2.4 Applications33
2.4.1 General-Purpose Species Identification33
2.4.2 Agriculture35
2.4.3 Intraspecific Variation, Geographical Distribution, Climate, Phylogeny36
2.5 Summary37
3 Feature Extraction ??????????????39
3.1 Leaf Shape39
3.1.1 Study of Existing Techniques for Leaf Shape Analysis40
3.1.2 Evaluation and Results43
3.2 Leaf Texture44
3.2.1 Macro-texture45
3.2.2 Micro-texture46
3.3 Margin Characteristics50
3.3.1 Extracting the Margin51
3.4 Locating the Apex and Insertion Point53
3.4.1 Dynamic Time Warping53
3.4.2 Finding the Points of Margin Symmetry55
3.5 Venation Patterns55
3.5.1 Extraction by Evolved Vein Classifiers56
3.5.2 Extraction by Ant Colonies59
3.5.3 Results and Comparison of Methods61
3.6 Summary62
4 Machine Learning for Plant Leaf Analysis ??????????????63
4.1 Incorporating Intra-species Variation into Plant Classification63
4.1.1 Utilizing the Hungarian Algorithm for Improved Classification of Leaf Blades64
4.1.2 Comparing Leaf Margins Using Dynamic Time Warping74
4.2 Combining Different Leaf Features75
4.2.1 Probabilistic Classification from K-Nearest-Neighbour76
4.2.2 Automatic Feature Selection79
4.3 Summary84
5 Botanists' Vision ??????????????86
5.1 Comparing the Eye Movements of Botanists and Non-Botanists87
5.1.1 Results and Analysis88
5.2 Reverse Engineering of Expert Visual Observations91
5.2.1 Related Work92
5.2.2 Methodology93
5.2.3 Evaluation97
5.2.4 Summary100
References ??????????????101
Index114