Content uploaded by Mohanad Fadhil Jwaid
Author content
All content in this area was uploaded by Mohanad Fadhil Jwaid on Feb 20, 2019
Content may be subject to copyright.
1
Maharashtra Institute of Technology, Pune
Department of Information Technology
Seminar II topic :
SEGMENTATION METHODS
FOR OBJECT BASED ON REMOTE SENSING IMAGE
ANALYSIS
Under the guidance of
Prof. Trupti Baraskar
MIT, Pune
Mohanad F. Jwaid
Roll No. : 606013
M.E., I.T.
MIT, Pune
Submitted by :
2
Index
1. Introduction
2. Literature Survey
3. Classification of image segmentation
4. Comparative of image segmentation
5. Application of remote sensing
6. Conclusion
7. References
3
1. Introduction
Remote sensing is the'science of obtaining information'about objects or areas
from a distance, typically from aircraft or satellites
Remote sensors collect data by detecting the energy that is reflected from earth
Remote Sensing using electromagnetic spectrum to image the land, ocean, and
atmosphere.
Remote sensing images are normally in the form of'digital images. In order to
extract useful information from the images.
4
Energy Source (A)
Radiation and the Atmosphere (B)
Interaction with the Target (C)
Recording of Energy by the Sensor (D)
Transmission, Reception, and
Processing (E)
Interpretation and Analysis (F)
Application (G)
Remote Sensing Process Components
5
Cont…
Image segmentation in general is defined as a process of partitioning into
homogenous groups.
The goal of segmentation is to simplify and/or change the representation of an
image into something that is more meaningful and easier to analyze.
Efficient image segmentation is one of the most critical tasks in automatic image .
6
2. Literature Survey
Method 1: Stelios K. Mylonas et.al, “A Local Search-Based
GeneSIS algorithm for the Segmentation and Classification of
Remote-Sensing Images” , IEEE 2016
This is localized version of the GeneSIS algorithm which combines
the properties of GeneSIS with the principles of the region growing
algorithms.
This method presented for remote sensing image segmentation
7
Cont…
Method 2: Josef Baumgartner et.al, “New Approach to
Segmentation of Multispectral Remote Sensing Images Based on
MRF” , IEEE 2015
This is novel approach for remote sensing images, which is similar
to techniques such as Decision Templates or the Dempster–Shafer
method.
The algorithm denominated successive band merging (SBM) is
proposed in this method for segmentation.
8
Cont…
Method 3: Tiancan Mei et.al, “Supervised Segmentation of
Remote Sensing Image Using Reference Descriptor” , IEEE 2014
This is novel approach proposed a robust higher level feature
representation for superpixel-based remote sensing image
segmentation.
9
Cont…
Method 4: Julien Michel et.al, “Stable Mean-Shift Algorithm and
Its Application to the Segmentation of Arbitrarily Large Remote
Sensing Images” , IEEE 2014
This is new property called stability of segmentation algorithms and
demonstrate that piece or tile-wise computation of a stable
segmentation algorithm can be achieved with identical results with
respect to processing the whole image at once.
10
Cont…
11
Cont…
12
3. Classification of image segmentation
Image segmentation
Threshold-based Region- based Edge-based
Clustering Region growing Mean-shift
Watershed
Hierarchical Partition Merging and
splitting
13
4. Comparison between image segmentation
methods
Method Advantages Disadvantages
Region
based
Method
Gives better result
Provides flexibility
Proper selection of seed gives
accurate result
To decide stopping criteria for
segmentation is difficult task.
Selection of noisy seed by user leads to
flawed segmentation.
Sequential by nature and quite
expensive in both computation time and
memory.
Cluster
based
Method
Reduces false blobs.
Eliminates noisy spots.
More homogeneous
regions are obtained.
Computationally expensive.
Doesn’t works well with non
globular clusters.
Sensitive to initialization
condition of cluster number
and center.
14
Cont..
Method Advantages Disadvantages
Edge
based
method
Works well for images
having good contrast
between regions.
Second order differential
operator gives reliable
result.
For all type of images, single
operator doesn’t suits.
Size of operator and
computational complexity are
proportional to each other.
Threshold
based
Method
Computationally
inexpensive.
Can be used in real time
applications.
Fast and simple
Highly noise sensitive.
Selection of threshold is crucial, wrong
choice may result into over or under
segmentation.
For an image with broad and
flat valleys or without any
peak, it doesn’t works well.
15
Cont..
Parameter Threshold based
method
Object based
method
Pixel based method Cluster based
method
Super-pixel based
method
Spatial Information Ignored Considered Ignored Considered Considered
Region Continuity Reasonable Good Reasonable Reasonable Good
Speed Fast Slow Moderate Fast Moderate
Computation
Complexity
Less Rapid Moderate Rapid Moderate
Automaticity Semiauto Semiauto Interactive Automatic Automatic
Noise Resistance Less Less Less Moderate Moderate
Multiple Object
Detection
Poor Fair Poor Fair Fair
Accuracy Moderate Fine Moderate Moderate Moderate
16
5. Application
Agricultural
Land cover
Forecasting weather to warn about natural disasters
Driving with no hands (autonomous vehicles)
Observing population growth in urban areas using land use change
Surface Temperature
6. Conclusion
Remote sensing is one of the important technologies that we use in several
applications such as monitoring the ground and detect oil fields and forests etc.
And used in the field of image processing and segmentation.
Different techniques developed for image segmentation perform well and
comparable to the methods used in practice.
The single segmentation is applicable to all type of images nor do all the
segmentation methods perform well for one particular image.
10/ 06/2016
17
18
7. References
[1] L´aszl´o, I., B. Dezs?o, I. Fekete and T. Pr¨ohle, A fully segmentbased method for the classification of satellite
images, Annales Univ. Sci.Budapest., Sect. Comp., 30 (2009), 157–174.
[2] L´aszl´o, I., G. N´ador, I. Fekete, G. Csornai and A. Kocsis, A segment-based classification method for satellite
images, in: Kov´acs, E. and Winkler, Z. (Eds.) Proceedings of the 5th International Conference of Applied
Informatics (ICAI), Eger, 2001, 151–163.
[3] Stelios K. Mylonas et.al, “A Local Search-Based GeneSIS algorithm for the Segmentation and Classification of
Remote-Sensing Images”, IEEE JOURNAL OF SELECTED TOPICS IN APPLIED EARTH OBSERVATIONS AND
REMOTE SENSING, 2016, IEEE.
[4] Josef Baumgartner et.al, “New Approach to Segmentation of Multispectral Remote Sensing Images Based on
MRF”, IEEE GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 12, NO. 8, AUGUST 2015.
[5] Tiancan Mei et.al, “Supervised Segmentation of Remote Sensing Image Using Reference Descriptor”, IEEE
GEOSCIENCE AND REMOTE SENSING LETTERS, VOL. 12, NO. 5, MAY 2015
[6] Julien Michel et.al, “Stable Mean-Shift Algorithm and Its Application to the Segmentation of Arbitrarily Large
Remote Sensing Images”, IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING, VOL. 53, NO.
2, FEBRUARY 2015
19