Knowledge Acquisition for the Interpretation of Spatial Data
Document Sample


Knowledge Acquisition for the
Interpretation of Spatial Data
Monika Sester
Abstract
在review後,提出符合「半自動模式獲得系統」
(semi-automatic model acquisition system)的
方法論,並證明其在某些應用上是具有潛力的,
亦即建立了找尋data中 所具有的概念或知識的
一個方法,並對這個方法作評價,且說明其優
缺點,最後將之利用於空間資料的詮釋上,及
略述了這個方法在資料詮釋工作上的可利用性。
Review
Mohan & Nevatia (1989)
Haala (1996)
Strat、Bhanu & Poggio (1994)
Wiston (1977)
Connell & Bardy (1985)
Quinlan (1990)、Wrobel (1994) 、Pearce & Caelli (1997)
Koperski & Han (1995)
Mackaness et al. (1997)
Weibel et al. (1995)
Principle of Human Learning
Seeing and finding relations to already existing
knowledge (Anderson 1980)
Be analysed based on known object,features and
relation
Approach and Principle of
Machine Learning
Learning classification tree with ID3
Object-oriented learning and interpretation
procedure
Object-oriented data structure
1. Properties of objects and relation
2. Basic structuring methods
Learning the Structure of Amp
Object
Example for learning object descriptions
Example for learning relations
Example of suburban area
Summary and evaluation of the approach
Further Applications
Derivation of models for image interpretation
Map matching---data integration
Map generalization and database generalization
報告完畢
謝 謝
Learning Classification Tree
With ID3
Quinlan (1986):藉已分類完成的樣本所形成的
決策樹,來對未知物做分類
優點
缺點
Example
Input
Output
Object-oriented Learning and
Interpretation Procedure
原 理:每一個物件具一個屬性,並與其他
物件共享某種關係
假 設:幾何特性、2D物件
訓練原則:概念由學習而建立;訓練者負責組
織結構
學習系統
Interaction of Learning and
Interpretation
Shell of Object clarres
Representation of a Drawing of
an Airplane
Properties of Objects
Contains/contained
Object-elongation
Object-lq
Object-size
Polytype
Object-form
Num-point、num-right、num-parallel…
Junction-ell、junction-tee、junction-frk….
Relations of Objects
Connection
Inside,enclose
Are-parallel,are-orthogonal
Common-ell、common-tee、common-arw…
Size-diff
Distance
Left-position,top-position
Common-sides
Same-polytype
Example
A:[ contains: yes ,contained: no ,object-elongation
:1.4 ,object-elongation: 2 ,…..
rel(A,B)=[ connection: no ,inside:no ,encloses:yes
,are-parallel:no ,….
Basic Structuring Methods
Recognize-object
Part-of
Learning Environment(left)with
Scene to Be Interpreted(right)
判斷物體類別標準之學習
系統自動建立物件類別
街道與軌道間關連性的學習
學習後的結果
完整語意網模式的形成
Example of Suburban Area
分類結果
Summary and Evaluation of the
Approach
可由程式語言來支援學習及詮釋環境物件
兼顧詮釋data的豐富性及演算法的複雜性
技術簡單
概念的學習是有效的
結構僅能由要素組成
所有的物件由某種物件分出來,其屬性相同
學習是歸納的過程,不能證明這的概括化的正
確性
樣本的選取需小心
Related docs
Get documents about "