[1] Why spatial image analysis? | [2] Regions and image objects | [3] Image segmentation | [4] Knowledge representation | [5] Class modelling
University of Salzburg, Department of Geoinformatics, 2021/21
Spatial properties (structure) are as important as spectral properties (colour) in image analysis
|
not a trivial task...
Which image objects are we able to perceive?
What do images evoke in our visual context and our brain?
|
|
Local pixel neighborhood
Two interrelated methodological pillars (Lang 2008)
(1) segmentation / regionalisation for nested, scaled representations ('object scape')
(2) rule-based classifiers based on spectral and geometrical properties as well as spatial relationships
Segmentation in several hierarchical scales
scale-specific vs. scale-adapted segmentation
Representations in various nested levels
Each object 'knows' its neighbours and hierarchical level
We all have a concept of “cold”, “warm”, “hot”
Knowledge in image interpretation: visual skills are complemented by explicit knowledge, enriched by experience and training
Structural knowledge
Procedural knowledge
Heuristics describing certain geographical phenomena are encoded in a set of rules. Rules can address all kinds of features.
Example soccer field = a special kind of grassland patch with certain size and shape.
We all have a concept of “cold”, “warm”, “hot”
Classes are embedded in a hierarchical heritage scheme
(1) Feature-based inheritance
(2) Semantic inheritance
Example: Monitoring of water quality in swimming lakes over three seasons
Image search criteria: [(0) sensor type], (1) time window , (2) geographical focus (AOI),
(3) cloud cover , (4) thematic focus , (5) monitoring period.
How to measure the information in an image?
Information measurement requires information units
(bins, classified pixels, objects, etc.)
where P = percentage, i = class 1 … i
Structural knowledge can be organized in knowledge organizing systems (KOS)
Semantic net
Corine Land Cover (CLC)
Land Cover Classification System (FAO-LCCS)
Spectral libraries represent spectral reflectance and physical properties of different land cover types
Level of landscape elements
Ambition: to delineate and classify composite objects in a complex real-world scene
A strategy to approach composite classes by applying multi-scale segmentation and spatial modelling
“Multiscale” is implicit in the term class modelling
Spectral signatures of elements
Defining an object’s body-plan
Critical reconditions
Specify set of target classes
Composite object (yellow outline) consists of several components (sub-objects)
doi: 10.3390/ijgi8110474
doi: 10.1016/j.isprsjprs.2013.09.014
doi: 10.1016/j.isprsjprs.2013.09.014
Retrieved from: http://essential.metapress.com/content/15u6x2189425705l/