Sensors, Data, Analyses and Applications

1
Sensors, Data, Analyses and Applications
2
Outline

• Introduction
• Problems, Solutions and Applications
• Advanced Sensor Registration and Fusion -
  Structural Engineering
• Anomaly Detection Using Wireless Sensors
  Homeland Security
• Crop Yield Prediction Using Hyperspectral Data -
  Agriculture
• Historical Map Analysis - Environmental
  Engineering
• Geographic Information Systems - Socio-Economics,
  Hydrology, Water Quality, Atmospheric Science,
  Military
• Automating Microarray Image Analysis
  Bio-informatics
• Building and Analyzing 3D Anatomical Data Sets
  Neuroscience, Ophthalmology
• Tracking Diffraction Edges in Microscopy Images -
  Semiconductor Physics
• Teaching Basic Image Operations - Education
• Future Work
• Summary

3
Introduction

• Measurement Sensors
• Measured physical entities, e.g., displacement,
  temperature, reflectance
• Positioned as satellite, aerial or ground sensors
  
• Connected as wireless or wired sensors
• Sensors of Interest
• Spectral properties EO, IR, HS, UV, SAR, X-ray,
  OCT, MRI, CT
• Regardless of position and connection of sensors
• Input Data
• Images and point measurements (raster and vector
  data types)
• Spatial and temporal nature
• Represent multi-dimensional multi-variate
  phenomena
• Varying datum precision
• Data of Interest
• Any combination of miscellaneous data sets
  combined with image data

Sensors
Data
Analyses
Applications
4
Introduction

• Image Applications
• bioinformatics, hydrology, precision farming,
  automation in semiconductor industry, topography,
  plant biology, medicine, automobile industry,
  database retrieval, surveillance, biometrics for
  identification, digital library

5
Credits

• Project Team Members Peter Bajcsy, Sang-Chul
  Lee, Peter Groves, Sunayana Saha, Tyler Alumbaugh
• Support Michael Welge, Loretta Auvil, Dora Cai,
  Tom Redman, David Clutter, Duane Searsmith, Lisa
  Gatzke, Andrew Shirk, Ruth Aydt, Greg Pape, David
  Tcheng, Chris Navaro, Marquita Miller.

6
Registration and Data Fusion of Advanced Sensors
7
Problem Description
Wall
Load

• Given Multiple Advanced Sensors Measuring Point
  and Raster Types of Physical Entities, conduct
• Continuous approximation of measured physical
  entity
• Transformation of measured variables
• Sensor registration
• Data fusion
• Location dependent accuracy/confidence prediction
• Visualization

8
Motivation

• Krypton Data
• Accurate, Spatially Sparse Measurements With
  Relatively Small Number of Measurement Points
  (lt256 Points)
• Stress Photonics (SP) Data
• Less Accurate With Respect To Krypton, Spatially
  Dense Measurements with a Limited Field of View

9
Point Measurements Krypton
10
Sensor Data Fusion
11
Anomaly Detection Using Wireless Sensors
12
Problem Description

• Given wired and wireless multi-spectral and other
  sensors, assess hazards in an indoor environment.
• Data acquisition
• Wired multi-spectral sensors thermal infrared,
  hyperspectral, visible and multi-spectral cameras
• Micro-Electro-Mechanical Systems (MEMS) with CPU
  MICA sensors with tinyOS
• Wireless data transmission and reception
• Sensor registration
• Data fusion
• Anomaly detection

13
Data Acquisition
Light off, t0
Light on
Light off, t1
Visible spectrum
Thermal IR
Temperature luminance readings
R,G,B, nearIR
14
Anomaly Detection
Anomaly
15
Crop Yield Prediction Using Hyperspectral Data
16
Problem Description

• Given a set of ground measurements of crop yield
  related variables and airborne hyperspectral
  measurements , find the most relevant
  hyperspectral bands for predicting ground
  variables and the best modeling methods with the
  smallest prediction error.
• Input data hyperspectral measurements registered
  with soil electrical conductivity courtesy to
  Lei Tian (UIUC) and Sreekala Bajwa (University of
  Arkansas)
• Solution Methodology for hyperspectral band
  selection
• Unsupervised ranking 7 fast methods
• Supervised evaluation 3 accurate methods

Ground measurements
17
Motivation

• Rank bands to reduce computation requirements for
  processing hyperspectral imagery
• Remove noisy bands

One method
All methods
18
Hyperspectral Image Processing
19
Modeling Results for Soil Electrical Conductivity
20
Historical Map Analysis
21
Problem Description

• Given large numbers of paper maps from 1900,
  recover geo-referenced iso-contours for
  environmental restoration purposes.
• Input digital scans of historical maps
• Motivation manual extraction is tedious and
  laborious
• Challenges map folds, large curvature, high
  density,

22
Contour Extraction
23
Data Mining in Geographic Information Systems
24
Problem Description

• Search for the best partition of any geographical
  area that is
• (a) based on raster or point information,
• (b) formed by aggregations of known boundaries,
• (c) constrained or unconstrained by spatial
  locations of know boundaries and
• (d) minimizing an error metric.
• Raster or Point Information
• Grid-based information, e.g., from satellite or
  air-borne sensors
• Geographical point information, e.g., from GPS or
  address data base
• Boundaries (Vector Data)
• Man-made, e.g., Counties, US Census Bureau
  Territories
• Defined by environmental characteristics, e.g.,
  Eco-regions,
• Spatial Constraints and Error Metric
• Defined by applications

25
Vector Data
Boundary Information
Watersheds
Countries
Counties
Census Tracks
Census Blocks
Zip Codes
26
Raster Data
Image Georeferencing Information
Digital Elevation Map of Illinois
27
Georeferencing
Geo-registering data sets using georeferencing
(Lat/Lng lt-gt UTM lt-gt Pixels)
Input Raster and Vector Data Sets
Geo-registered Data Sets
28
GIS Data Fusion
29
Statistical Analysis
Computing statistics from geo-registered data sets
Geo-registered Elevation Map and County Boundaries
Elevation Statistics Per County
Mean Elevation
Standard Deviation
Kurtosis
Skew
30
Segmentation of Geo-Spatial Attributes
Segmentation Problem Find all neighboring
counties with similar attributes
Input Geo-info About Counties
Output Aggregations of Counties
Aggregations
Statistics
County Index
Counties
31
Clustering of Geo-Spatial Attributes

• Hierarchical clustering of crime data with the
  exit criterion being the number of clusters and
  the clustered feature being auto theft in 2000
  leads to six aggregations.

Tabular Display
Geographical Display
Boundaries
Boundary Aggregations
32
Geographical Error Evaluation and Decision making

• Error evaluation of partitions obtained by
  clustering and segmentation of mean elevation
  feature per Illinois county with Variance error
  metric

Partition Index
Eval0
Eval1
Eval2
Eval3
33
GIS Analysis
34
Image Problems in Bio-Informatics Microarray
Grid AlignmentMicroarray Grid ScreeningMicroarra
y Image Feature Extraction and Clustering
35
Problem Description

• Input Laser image scans (data) and underlying
  experiment hypotheses or experiment designs
  (prior knowledge)
• Output
• Conclusions about the input hypotheses or
  knowledge about statistical behavior of
  measurements
• The theory of biological systems learnt
  automatically from data (machine learning
  perspective)
• Model fitting, Inference process
• Motivation automation

36
Microarray Image Analysis
37
Microarray Image Analysis
38
Building and Analyzing 3D Anatomical Data Sets
39
Problem Description

• Matching Problem Search for transformation
  parameters to align medical slices.

Input slice x1
Visualization of Slice Transformation
Input slice x2
40
Matching in Anatomy

• Matching Problem
• Find a match to a new cross section in the
  existing 3D anatomical model.
• Find a match to a new 1D signal
• Find a match to an image acquired by another
  sensor, e.g., MRI, CT, Histology
• Segment 3D volume to improve matching

Structure - 3D Anatomy
Function 1D Signal
Metadata Annotation
41
Tracking Diffraction Edges in Microscopy Images
42
Tracking in Microscopy

• Tracking Problem Feature tracking for
  registration and alignment.

Output Alignment Information
Input Sequence
43
Teaching Basic Image OperationsImage
Calculator
44
Image Calculator
Numbers
Image Color Inversion
Images
45
Image Calculator
46
Image To Knowledge (I2K) Documentation

• Overview and tools

47
Problems of Interest

• Hyperspectral and multi-spectral image processing
• Biometrics with new sensors
• Earth science, atmospheric science, ecology and
  hydrology
• Bioinformatics
• Anatomical modeling in medicine
• Machine and computer vision
• SAR image analysis

48
Summary

• Overview of Image Data Processing and
  Applications
• Data Processing
• Raster and vector data
• High-dimensional data
• Spatial, temporal and spectral data types
• File format issues
• Data fusion of heterogeneous data
• Applications
• Bio-tools (Bio-informatics, Biology, Plant
  Science)
• Geo-tools (GIS, Remote Sensing, Agriculture,
  Hydrology, Water Quality Survey, Atmospheric
  Science, Military)
• Med-tools (Medicine, Biology)
• Edu-tools (Education)
• Other tools Video Analysis, Statistical
  Simulations, Target Recognition
• Interested ? Useful ? Let us know.