Lecture 3 Image Processing/Analysis Basic Requirements

molassesitalianΤεχνίτη Νοημοσύνη και Ρομποτική

6 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

54 εμφανίσεις

5th Intensive Course on Soil
Micromorphology

Naples 2001

12th
-

14th September

Image Analysis

Lecture 3

Image Processing/Analysis

Basic Requirements

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements

Basic requirements of image processing/analysis


e.g. need to threshold or not,

problems of subjectivity,

hydrid approaches (user/computer methods).

manual methods of measurement




Introduction to Sigma Scan Software

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements

Most Analysis requires some preliminary processing

Segmentation


involves separation of image into regions of similar features


simple form
-

2 phases
-

solids and particles >> binary image


multi
-
phase form
-

may be several different discrete grey
levels in an image.


More complex segementation methods may involved
segmentation based on texture/morphology/orientation.

Thesholding


selection of a grey
-
level intensity which separates two phases
.


For binary images
---------------
one threshold


For multiphase images
-----------
several thresholds


5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3
Basic Requirements

Segmenting a Binary Image using Thresholding

e.g. Change Grey
-
Level Image into 2 phases
-

solids and voids

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3
Basic Requirements

Segmenting a Binary Image using Thresholding

First Example is ideal
-

selection of threshold value to
separate black from white

Second Example: also ideal
-

clear definition of where
threshold should be.

However: Non uniform illumination may present problems.

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 2
Image acquisition

Illumination issues

Non
-
uniform
illumination will create
problems in later
thresholding and analysis

Normal Image

Centralised
Illumination


Gradation
from top left

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3
Basic Requirements

Thesholding:

If intensity distribution is of simple form, and automatic
threshold may be set (provided illumintaion is uniform)

With grey
-
level images
-

intensity distribution often does not show a
discrete minimum
-

making automatic thresholding difficult.

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3
Basic Requirements

Thesholding:

Interactive selection of threshold will be unreliable and may
well differ significantly from one person to another.

Data from 2nd and 3rd Intensive Courses:

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3
Basic Requirements

Thesholding:
Multi
-
phase situations

A new objective method will be discussed in Lecture 10

Intensity distribution of multi
-
phase image. The optimum
points for separation between phases are the minima

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements


Human eye is poor at intensity discrimination

computer is good at intensity discrimination

Human eye is good at texture recognition

difficult to derive suitable computer
algorithms to discriminate textures.

Morphological effects
-

e.g. surface roughness can mask
intensity effects
-


e.g. quartz and felspar have similar intensities in back
-
scattered SEM, but the morpholoigcal appearance
(texture changes in grey levels) are usually easily
discriminated by human eye.


Segmentation based on other methods:

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements

Segmentation based on other methods:

Manual delineation methods can be
useful. These aim to take the
advantages of the human eye in pattern
recognition and define areas of equal
grey
-
level which are the ideal for
feature size analysis

In this example, regions which are
near vertical are coded red.

There will be an exercise on
this later in course

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements

Region within red area has similar grey
-
level, but has
difference in calcium concentration.
Segmentation by
thresholding is not possible.

BSE image of volcanic deposition:

Sample from Monserrat

calcium map of area

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements



Traditionally
-

features would be measured by laying a
rule against the feature of interest.


This facility can be achieve more readily with digital
images by using an overlay cursor.


Measurement requires the use of overlays these are binary
images which can be overlain directly over the image
without affecting the original image.

Often several different overlays are available. Sigmascan
has 5 image overlays and one for annotation.

Simple Measurement

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements

Simple Measurement

Manual Method
directly on photograph

Overlaying a scale.


Is case, a special scale
was constructed
consistent with
magnification.


Units are in microns

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements

Simple Measurement

Manual Methods

Interactive using a cursor.


Click on one point

Move cursor

Click on second point

Distance is automatically
recorded

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
Basic Requirements

SIGMA SCAN is a package directly available over the
Internet which has a 30 day demonstration license.


Thus ideal for use in this teaching course

What does SIGMA SCAN do

allows several basic image processing procedures

limited capabilities

for more advanced operation.


All quantitative work is done via the overlay planes

Introduction to Sigma Scan Introduction to Sigma Scan

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

Using SigmaScan to Threshold

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

Use “handles” to adjust width of
selected intensity values

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

Many of the settings for SigmaScan must be first set
using this menu


The next two slides show sub
-
options of these settings

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

Selecting which parameters to measure

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

Selecting which overlay to use for annotation

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

Using MEASURE OBJECTS
labels each object with an
unique identifier

5th Intensive Course on Soil Micromorphology
-

Naples 2001

Image Analysis
-

Lecture 3:
SIGMASCAN

Each Feature is measured according to parameters selected
and displayed in an EXCEL Worksheet