AGC Technology Overview - CLU-IN · AGC Technology Overview Amy Walker Geophysicist U.S. Army...

20
AGC Technology Overview Amy Walker Geophysicist U.S. Army Engineering & Support Center, Huntsville 21 April 2016 US Army Corps of Engineers BUILDING STRONG ®

Transcript of AGC Technology Overview - CLU-IN · AGC Technology Overview Amy Walker Geophysicist U.S. Army...

AGC Technology Overview

Amy Walker Geophysicist

U.S. Army Engineering & Support Center, Huntsville

21 April 2016

US Army Corps of Engineers BUILDING STRONG®

Outline

1. How the sensors work and what they look like

2. How we utilize the acquired data

3. How the classification decision is made

2 BUILDING STRONG®

BUILDING STRONG®

EMI Sensors § Electromagnetic Induction (EMI) sensors measure the

response of nearby metal objects to magnetic fields created by currents running through a loop of wire

Geonics EM61

3

BUILDING STRONG®

Advantages of a Multi-Axis Sensor

4

     

BUILDING STRONG®

Classification Sensors § Same EMI physics § Specifically designed for munitions classification § Now we can separate the information useful for

classification (answer independent of location)

Receive cubes

Transmit coils

5

Time (ms) 0.1 1 10

Pol

ariz

abili

ty (m

3 /A)

0.001

0.01

0.1

1

BUILDING STRONG®

Sensors

6

MetalMapper TEMTADS 2x2

MPV

Geometrics MetalMapper 2x2

You Can’t Make Decisions Using Only the Rx4

Rx3

Rx2

Rx1

Rx4

Rx3

Rx2

Rx1

Measured Decays Tx1 Tx2

R(X) R(Y) R(Z) R(X) R(Y) R(Z)

Rx4

Rx3

Rx2

Rx1

Rx4

Rx3

Rx2

Rx1

Tx4 Tx3 R(X) R(Y) R(Z) R(X) R(Y) R(Z)

101

Pol

ariz

abili

ty 100

10-1

P1

10-2 P2

P3

10-3

0.1 1 10 Time (ms)

7 BUILDING STRONG®

You Can’t Make Decisions Using Only the Rx4

Rx3

Rx2

Rx1

Rx4

Rx3

Rx2

Rx1

Measured Decays Tx1 Tx2

R(X) R(Y) R(Z) R(X) R(Y) R(Z)

Rx4

Rx3

Rx2

Rx1

Rx4

Rx3

Rx2

Rx1

Tx4 Tx3 R(X) R(Y) R(Z) R(X) R(Y) R(Z)

101

Pol

ariz

abili

ty 100

10-1

P1

10-2 P2

P3

10-3

0.1 1 10 Time (ms)

8 BUILDING STRONG®

Rx4

Rx3

Rx2

Rx1

Rx4

Rx3

Rx2

Rx1

You Can’t Make Decisions Using Only the Measured Decays

Tx1 Tx2 R(X) R(Y) R(Z) R(X) R(Y) R(Z)

Rx3

Rx3

Rx2

Rx1

Rx4

Rx3

Rx2

Rx1

Tx4 Tx3 R(X) R(Y) R(Z) R(X) R(Y) R(Z)

101

Pol

ariz

abili

ty 100

10-1

P1

10-2 P2

P3

10-3

0.1 1 10 Time (ms)

9 BUILDING STRONG®

   

BUILDING STRONG®

Process data to remove effects of burial depth and object orientation

Mathematical Model for the EMI Response

location & orientation (extrinsic properties)

+ polarizabilities

(intrinsic response; NOT affected by burial depth or

orientation) Sensor Data

Pre-Process • Remove sensor effects • GPS coordinates

10

UXO are usually axial symmetric, clutter are not

Object Shape à Axial Symmetry UXO Non-UXO

BUILDING STRONG®11

Cylindrical, axial symmetry Variable, often asymmetric

BUILDING STRONG®

Thick-walled metal decays slower than thin-walled debris and clutter

Wall Thickness à Decay Rate UXO

thick wall & slow decay Non-UXO

thin wall & fast decay

12

BUILDING STRONG®

UXO vary in size (20mm to 155mm) but clutter is generally small Size (volume) à Magnitude UXO

Variable, often larger Non-UXO

Variable, majority are small

13

   

Target of Interest (TOI) Library

Collection of TOI signatures: 1.metadata, 2.sensor data,

and 3.polarizations ESTCP generated DOD maintained

BUILDING STRONG®14

     

Decision Process We prioritize the sources from those that are most similar to UXO (our library) from those that are not…

3 ways to get on the dig list: 1. Look like an item in the library 2. Be part of a cluster 3. Be big and deeply buried

BUILDING STRONG®15

BUILDING STRONG®

?

Compare each item to signatures in the library

...too large for 37mm

37mm

…just right. MATCH

75mm

…too small for 105mm

105mm

?

0.58

0.75

0.99

16

?

BUILDING STRONG®

Compare each signature to all other signatures on site

T-Bar Fuze Non-hazardous clutter, did not add to library

No Library Match

u If there are clusters of items that do not match the library signatures, we excavate some of them and proceed accordingly…

17

BUILDING STRONG®

Look for unexpected Big, Symmetric, Thick-walled items

Time (ms) 0.1 1 10

Pol

ariz

abilit

y

0.1

1

10

100

P1

P2

P3

3-in stokes @ SLO

18

   

BUILDING STRONG®

Software § UX-Analyze (Geosoft Oasis Montaj) § UXOLab

19

BUILDING STRONG®

Reference Materials § ESTCP

https://www.serdp-estcp.org/Featured-Initiatives/ Munitions-Response-Initiatives/Classification-Applied-to-Munitions-Response

§ ITRC http://itrcweb.org/Guidance/ListDocuments? TopicID=16&SubTopicID=3

§ Previous M2S2 webinars https://clu-in.org/live/archive/

20