Lecture slides on "Very basic GPR" Presentation December - - PDF document

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/311308562

Lecture slides on "Very basic GPR"

Presentation · December 2016

DOI: 10.13140/RG.2.2.10693.45288

CITATIONS READS

350

1 author: Some of the authors of this publication are also working on these related projects: Waterborne geophysics on shallow water View project Geophysics and trees View project Luigi Sambuelli Politecnico di Torino

158 PUBLICATIONS 1,024 CITATIONS

SEE PROFILE

All content following this page was uploaded by Luigi Sambuelli on 02 December 2016.

The user has requested enhancement of the downloaded file.

(very) BASIC GPR

Index

• Outlines of the working principles
• Basic equations and parameters
• Basic reflection and patterns
• Outlines of the instrument
• Field operations – data acquisition
• Data processing
• Data interpretation
• L. Sambuelli - Politecnico di Torino - DIATI -

2016 2

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 3

Outlines of the working principles

• The GPR is an instrument for underground investigation based on

sending and receiving electromagnetic (EM) waves in frequency bands ranging from 40 to 2000 MHz.

• An EM pulse is send into the soil, with an antenna pulled or

pushed on the soil surface, when it impinges on a surface separating two material with different EM properties, it is reflected back and acquired from the same or from another antenna.

• These EM echoes (lasting from 10 to 1000 ns) are acquired by a

control unit (CU) and, after some averaging, are stored (roughly 50 echoes per second).

• Each echo corresponds to a radar trace.
• The radar traces, displayed side by side, gives the radargram
• L. Sambuelli - Politecnico di Torino - DIATI -

2016 4

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 5

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 6

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 7

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 8

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 9

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 10

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 11

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 12

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 13

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 14

TxRx

Outlines of the working principles

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 15

TxRx

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 16

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

]

• r Wb/m

[T flux field magnetic [A/m] intensity field magnetic ] [A/m density current electric [A] intensity current electric I [V/m] intensity field electric ] [C/m density volumic charge ] [C/m nt displaceme electric

2 2 3 2

B H J E q D     

17

Basic equations and parameters

q D div  

 B div 

t B E rot       t D J H rot       

) (          D div H B E D E J       

with

Maxwell set of equations and EM parameters

[H/m] 10 4 [F/m] 10 85 . 8 [H/m] [F/m]

7 12  

              

r r

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

for vacuum is "0" pedix e th relative for is r" " pedix e th [F/m] material the

• f

ty permittivi electrical [H/m] material the

• f

ty permeabili magnetic [S/m] ty conductivi electrical   

a b

18

Properly combining the relations in n the he gr green en line ne Maxwell obtained 2 partial differential equations (PDE) for and

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

E  H 

t H t H H          

2 2

  t E t E E          

2 2

 

Equati ations

• ns and in the previous

ious slid ide e clearly arly show that at the two fields ds ( and ) are «link inked ed» » and that when exists ists a a time me- varying ing electric ctric field ld there re will be also a a time ime-var arying ying magne neti tic field ld which, , in n turn urn, , will ll gene nerat rate e a time me-var arying ying electr tric ic field ld and nd so on. n.

a b E1 E1 E2 E2

19

E H

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 20

Basic equations and parameters

Equations E1 and E2 are similar and both have a solution in the form:

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

 

r t i re

e V V

    



Thanks to the similarity we can analyse only E and the analysis will hold for H too. Moreover we can analyse an E field oscillating along the x direction and propagating along the z direction (that is in depth) and write:

 

 

z t i z x x

e e E t z E

  

 

 

 , , ) , , (

E3 E3

21

E3 E3 is is the e equa equati tion

• n of a di

dissip ipativ ive e wave wher ere:  [1 [1/m] m] is is the e di dissip ipat atio ion fac actor

• r

 [rad/m] is is the e pr propa paga gati tion

• n factor
• r

=2 =2f [rad/s] is is the e angu gular frequency equency f [1 [1/s] ] is is the e freq equen uency cy v= v=/ [m/s] ] is is the e pr propa paga gati tion

• n vel

eloci city ty =2 =2/=v/f v/f [m] is is the e wavel elen engt gth

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

22

 

 

z t i z x x

e e E t z E

  

 

 

 , , ) , , (

E3 E3

 E3 is a solution of PDE Maxwell equations.  E3 is a monochromatic wave (1-frequency=).  GPR sends a pulse NOT a 1-frequency wave.  A calculus theorem states that if E3 is a solution of the PDE ALSO SO A SU SUM of E3-like equations each one with its own i ,

i , i , is a solution of the PDE.

 Fourier ‘s theorem states that a pulse can be written as A A SU SUM of 1-frequency waves (E3-like).  GPR pulse can be a solution of the PDE and indeed it it is is.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

23

 

 

z t i z x x

e e E t z E

  

 

 

 , , ) , , (

E3 E3

• In the following slide the Fourier theorem is

graphically shown;

• The top graph shows 13 harmonics, i.e.

thirteen 1-frequency waves, which, summed up, give the bottom graph;

• The bottom graph shows a synthetic pulse of

an antenna with central frequency fc roughly equal to 330MHz.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 24

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 25

Basic equations and parameters

3 ns

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 26

Basic equations and parameters

• In the following slide the Fourier theorem is

graphically shown;

• The top graph shows 13 harmonics, i.e.

thirteen 1-frequency waves;

• The bottom graph shows the amplitude

spectrum i.e. the amplitudes of the 13 harmonics each one at its own frequency.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 27

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

•  is particularly sensitive to : the higher is the conductivity the

higher is  the higher is the dissipation (a kind of attenuation) of the wave

• For low-co

conduc nductiv tivity mater erial ials 𝛽 → 0 and 𝛾 = 𝜕 𝜈𝜁 1/2 then v can be simplified as:

28

Because 𝑑 =

1 𝜈0𝜁0 ≅ 3 × 109 [m/s] is the light velocity in vacuum

and 𝜈𝑠 ≅ 1 for most of the materials.

𝑤 = 1 𝜈𝜁 = 1 𝜈0𝜁0𝜈𝑠𝜁𝑠 = 𝑑 𝜗𝑠

• In the next slide the difference between the

true and the simplified velocity for three different materials is shown.

• It is worthy to note that at GPR frequencies

(50 MHz-1000 MHz) the difference can be so small to make the approximation acceptable.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 29

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

30

• In the next slide the difference between the

attenuation in three different materials is shown.

• It is worthy to note that at GPR frequencies (50

MHz-1000 MHz) a significant difference between resistive and conductive materials

• ccurs.
• every -6dB the Aout=0.5Ain
• L. Sambuelli - Politecnico di Torino - DIATI -

2016 31

Basic equations and parameters

10

20log

• ut

in

Ampitude dB Amplitude 

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

32

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

33

r1 r2

The EM energy, as long as propagates in a medium, disregarding the EM properties of the medium and the frequency, attenuates because it is spread

• ver increasing semispherical
• surfaces. Then there is another

attenuation: the geometrical one (spherical divergence). This one too contributes to decrease the signal strength.

r

r E E e r

 



• L. Sambuelli - Politecnico di Torino - DIATI -

2016 34

Basic equations and parameters

• In the following two slides the effect of α and

spherical divergence on a propagating pulse is shown;

• The first one (35) shows what happens to the pulse

as long as it travels away from the source at three depths in a medium with 100 0 Ωm;

• The second one (36) shows the Fourier spectra of

the pulses.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 35

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 36

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 37

Basic equations and parameters

• In the following two slides the effect of α and

spherical divergence on a propagating pulse is shown;

• The first one (38) shows what happens to the pulse

as long as it travels away from the source at three depths in a medium with 20 20 Ωm;

• The second one (39) shows the Fourier spectra of

the pulses;

• No

Note th the di differences nces betw between th the tw two m

• medi

dia

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 38

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 39

Basic equations and parameters

• In the next slide the difference between the

wavelengths in three different materials is shown.

• It is worthy to note that at GPR frequencies (50

MHz-1000 MHz) differences in material conductivities do not affect the main pulse wavelength (differences in relative permittivities affect more!! 𝑤 =

𝑑 𝜁𝑠 ; λ = 𝑤 𝑔 → 𝜇 = 𝑑 𝜁𝑠𝑔)

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 40

Basic equations and parameters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

41

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic equations and parameters

42

Basic reflection and patterns

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 43

Basic reflection and patterns

• GPR pulse, as the other wave phenomena, are

REFLECTED, REFRACTED, DIFFRACTED crossing media with different EM properties (, , ).

• GPR survey are mainly based on REFLECTION.
• GPR pulse REFLECTION and REFRACTION
• ccurs when the pulse impinges a surface

separating two materials with different EM properties.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 44

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic reflection and patterns

45

Reflec lection tion a) a) an and refraction action b) b) la laws ws: : an angles les

       

2 1 2 1

sin sin a) 1 b) sin sin

i i r t

v v         

Hi Hr

1 2

Ht

i



r



t



E orthogonal to page

i r t

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic reflection and patterns

46

𝜘𝑗 𝜘𝑠

Hi Hr

𝜘𝑢 1 2

Ht

E orthogonal to page

20

i

  

Reflecti lection

• n a) a

) and refraction action b) la ) laws: s: amplitud litudes es

2 1 2 1 2 2 1

a) REFLECTION COEFFICIENT: R= 2 b) TRANSMISSION COEFFICIENT: T=

r i t i

Z Z E RE Z Z Z E TE Z Z     

i r t

1

t r i

T R E E E    

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic reflection and patterns

47

Z is a complex quantity:

i Z i     

But for low-con conductiv ductivit ity materials

0 then Z     

But being also

1 1 and 1

r r r r

c Z c             

1 2 1 1 2 1 2

2 and T=

r r r r r r r

R           

• L. Sambuelli - Politecnico di Torino - DIATI -

2016

Basic reflection and patterns

48

REFLECTOR PARALLEL TO THE GROUND SURFACE:

In the radargram the reflected pulses are on a horizontal line which can be seen at a time

2 1 h twt v 

Onc nce e an an es estim imat ate e of the e vel eloci city ty of the e fir irst layer er is is avail ilable le the depth of the reflector can be calculated. The horizontal reflectors look horizontal on radargram.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 49

h twt2h/v1 v1 v2

Basic reflection and patterns

TxRx TxRx TxRx TxRx TxRx

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 50

REFLECTOR INCLINED WITH RESPECT TO THE GROUND SURFACE:

In the radargram the reflected pulses are on an inclined line which can be seen at times

 

2 ( ) cos 1 h twt x x v  

Once e an es estim imate e of the e vel eloci city ty of the e fir irst layer er is is avai ailab able le the depth of the reflector can annot not be calculated. The inclination of the reflector on radargram is “wrong”: it depends on reflector inclination, velocity AND depth

Basic reflection and patterns

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 51

 h d twt2d/v1 v1 v2

Basic reflection and patterns

TxRx TxRx TxRx TxRx TxRx

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 52

Basic reflection and patterns

DIFFRACTION DUE TO A POINT-LIKE OBJECT

In the radargram the diffracted pulses are on a hyperbola which can be seen at times

 

2 2

2 1 x x h twt v   

and it has: the slopes of its asymptotes equal to 2/v1 v1; the twt of its vertex equal to 2h/v1 v1

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 53

Asymptote: slope=2/v1 h v1 x0 twt=2h/v1 x

Basic reflection and patterns

TxRx TxRx TxRx TxRx TxRx

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 54

Basic reflection and patterns

DIFFRA FFRACT CTION ION DUE TO A POINT-LIKE OBJECT

As point-like object it must be intended a body

• dy whose
• se size

ze s (as “seen” from the EM pulse) is roug ughly ly equal ual to

• the

he domi

• minant

nant wavelengt length h of the e pulse se . It is worthy to note that one can also have diffraction hyperbolas with “corners” (see slide 5) and with the ends of thin plane-like objects (such as fractures). As the slope of the asympotes are quite easy to evaluate

• n radargrams, diffr

fractio action n even ents ts are often welcome because they allo lowed ed an estimation timation of the e pulse se velocity locity in the medium surrounding the diffracting object.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 55

Basic reflection and patterns

REFLECTOR PARALLEL TO GROUND SURFACE WITH CMP MP ACQUISI QUISITION TION

In the radargram the reflected pulses are on a hyperbola which can be seen at times

 

2 2

4 1 x x h twt v   

and it has: the slopes of its asymptotes equal to 1/v1 v1; the twt of its vertex equal to 2h/v1 v1

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 56

h v1 x0 Asymptote: slope=1/v1 twt=2h/v1 x

Basic reflection and patterns

v2 Tx_fix Rx Rx Rx Rx

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 57

Basic reflection and patterns

REFLECTOR PARALLEL TO GROUND SURFACE WITH CMP MP ACQUISI QUISITION TION

Sometimes, when no diffraction events occur, it can be very useful to make two or three CMP profiles in the investigation area. As the slope of the asympotes are quite easy to evaluate on radargrams, ref eflecti ection

• n even

ents allow an es estim imati tion

• n of the

e pu pulse e vel eloci city ty in the medium surrounding the targets.

Outlines of the instrument

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 58

Outlines of the instrument

Block diagram of a GPR

Data store Video Tx Rx Antenna/s A/D Ampl.

Power

High voltage pulse generator

C.U. P.C.

Sometimes C.U and P.C are in a single box

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 59

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 60

The transmitting antenna emits large part of EM energy in a “cone of radiation” and the receiving antenna gathers waves travelling back in a similar cone. That’s why GPR can “see” objects out of the vertical from the antenna box. Diffractions and CMP acquisition are possible thanks to this phenomenon.

E orthogonal to page MONOS OSTATIC TIC configuration: 1 dipole which transmits and receives BISTATI TIC configuration: 2 dipoles: one transmits and the other receives

Outlines of the instrument

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 61

As the transmitting antenna emits large part of EM energy in a “cone of radiation” the sizes of the reflecting surface (footprint) along the dipole (B) and across the dipole (A) varies with depth D mainly in function of the EM characteristic of the soil and of the frequency. 4 1 2

r r

D c A f A B         

Outlines of the instrument

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 62

As the transmitting antenna emits large part

• f EM energy in a “cone of radiation” the

hori rizonta zontal Spatial l Resolut solution ion x varies with depth D mainly in function of the EM characteristic of the soil and of the frequency. The verti tica cal l Spatial al Resolution

• lution z is usually

not supposed to vary with depth 4 2 D z x      

Spatial Resolution is the capability of recognizing the reflections from two close surfaces as distinct i.e. the capability of estimate the distance between them. YOU U CANNOT T EXPECT CT TO DISTI TINGU GUISH SH TWO O OBJEC ECTS TS HORIZON ZONTALL ALLY CLOSE SER R THAN x YOU U CANNOT T EXPECT CT TO ESTIMATE TE THE THICK CKNES ESS OF AN OBJECT ECT THINNER ER THAN z

Outlines of the instrument

z x

r1 r2

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 63

Outlines of the instrument

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 64

Outlines of the instrument

• The GPR antennas are characterized by: a

central frequency fc= frequency to which corresponds the maximum of emitted energy Emax

ax; fmin in and fmax ax=

= frequencies corresponding to an amplitue Emax/2 /2 generally symmerical about fc; BW=bandwidth. Usua uall lly y :

• BW=fmax

max-fmin min ;

; fmax=fc+B +BW/2 /2 ; fmin=fc-BW/2 W/2

• Mean spectrum from 970 traces of a profile collected on an

archaeological site with a 400 MHz GSSI antenna and a K2 IDS C.U.. Frequency window from 0 to 1000 MHz.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 65

Outlines of the instrument

fc fmin fmax BW

• The different attenuation phenomena that attenuate

the GPR signal travelling r [m] down and back are taken into account for in a so called Range nge Equation quation.

• For a smooth flat surface:
• For a rough flat surface:
• F are the dB that the equipment must have, R is the

reflection coefficient

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 66

Outlines of the instrument

2 2 4 10 2

10log 16

r

R e F r



 



       

2 3 4 10 3

10log 32

r

R e F r



 



       

• In the following slides the performances

required to a GPR equipment for detecting

• bject in these soils are shown
• L. Sambuelli - Politecnico di Torino - DIATI -

2016 67

Outlines of the instrument

r  r 10 50 1 10 100 1 10 500 1

The estimates with these formulae are usually quite optimistic e.g. sometimes we don’t “see” as deep as we planned.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 68

Outlines of the instrument

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 69

Outlines of the instrument

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 70

Outlines of the instrument

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 71

Outlines of the instrument

The estimates with these formulae are usually quite optimistic e.g. sometimes we don’t “see” as deep as we planned.

Indeed the traveling GPR pulse decays for: dissipation (α), attenuation (1/r) and scattering that is incoherent and chaotic reflections due to objects with size s<<. . Moreover layers can be a non-homogeneous patchwork of dissipative and non-dissipative materials.

GSSI- CODEVINTEC – Utility scan DF IDS – K2 2channel

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 72

Outlines of the instrument

Field operations – data acquisition

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 73

• Field operations need a design.
• Design consists of three steps.
• St

Step ep 0: antenna (frequency fc) selection [Depth of the target, Vertical Spatial resolution, Horizontal Spatial Resolution]

• St

Step ep 1: acquisition geometry that means to define: x= x=int nter erval al betw etween two consecuti secutive radar dar traces; s; y= y=int interval al betw etween n two adjace jacent nt GPR R prof

• fil

iles. s.

• St

Step ep 2:acquistion parameters that means to define: t=trace race sa sampling pling interval al; ;

• T=trace

ace durat ation ion (T/t +1=N=numbe =number r of f sa samples ples per trace) e)

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 74

Field operations – data acquisition

• St

Step p 0: antenna (frequency fc) selection

• This step will influence also Step 1 and Step 2.

Antenna frequency selection has to be done searching for a tradeoff between frequency from Ran ange ge Eq Equation ation (R (R.E. E.) ) and Spat atial ial Resolut solution ion (S (S.R). .R). We could be interested in a high detail (good S.R.) at a significant depth (high values of F in R.E.)

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 75

Field operations – data acquisition

• St

Step p 1: acquisition geometry x.

• x:

x: according to the Shannon sampling theorem should be

• Example: f=500 MHz, r=8 (c=0.3) x0.07 m
• Usually GPR trace intervals are around few

centimeters

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 76

Field operations – data acquisition

max

c in m/ns, f in GHz 2 2

r

c x f     

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 77

Field operations – data acquisition

• St

Step p 1: acquisition geometry y.

• y: according to the Shannon sampling theorem

should also be

• However a good compromise between spatial

sampling and acquistion time (€ or $) is y =0.5m. Only when acquiring data with 900 MHz or higher frequencies y= y=0.10.2 m

max

c in m/ns, f in GHz 2 2

r

c y f     

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 78

Field operations – data acquisition

• St

Step p 2: acquisition parameters t.

• t:

t: according to the Shannon sampling theorem

• Using the second rule more information can be

extracted from the radar traces

max max

1 1 but it's better 2 6 t t f f    

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 79

Field operations – data acquisition

• St

Step p 2: acquisition parameter T.

• T: has to do with the maximum depth D from

which one wants to get a reflection

• Here vmax

ax is the maximum expected velocity

• Some GPR’s instead of asking t

t ask for T and the number of sample per trace N. But

max

4 D T v 

 

1 1 T N t N T t       

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 80

Field operations – data acquisition

• St

Step p 2: acquisition parameter T.

• Of course the selection of T does not guarantee

the reflection from a target at depth D D because of attenuations, remember the range equation (slides 66 to 71)

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 81

Field operations – data acquisition

• 1. Survey area delimitation and line tracking with

non metal rulers;

• 2. Area surveying with forward and backward

parallel paths without turning the antenna;

• 3. Eventually cross paths to have denser data set

and information on orthogonal structures;

• 4. Even if the antenna is tracked with a RTK GPS

always pick the coordinates of the area corners to be able to produce a readable final graphic document

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 82

Field operations – data acquisition

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 83

Field operations – data acquisition

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 84

Field operations – data acquisition

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 85

Field operations – data acquisition

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 86

Field operations – data acquisition

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 87

Field operations – data acquisition

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 88

Data processing

• Data processing is of paramount

importance in GPR;

• In archaeological prospecting usually it

is done in two main steps: 1) survey lines (radargrams) processing; 2) data- cube (spatially assembled radargrams) processing.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 89

Data processing

• Survey lines (radargrams) processing may consists
• f many steps with the aim of cleaning as much as

possible the radargrams i.e. reducing noises (random reflections, parallel lines due to ringing, frequencies out of band, d.c. shift... ) and enhancing “useful signals”.

• Often raw data are awful and seem meaningless…

but do not despair !

• Even at the end of step 1) there can be a lack of

structures but, again, do not give up ! Step 2) can do miracles.

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 90

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 91

Novalesa\smaria.MIS\150923AB.ZON\PROCDATA\LTT10012.06T subtract-mean(dewow) / 5 / 0 / 0 / 0 / / 0 / 0 / 1 / 869 move starttime / -3.5 / 0 / 0 / 0 / / 1 / 0 / 1 / 866 time cut / 75 / 0 / 0 / 0 / / 1 / 0 / 1 / 866 manual gain (y) / -19.57447 / 8.116377E-7 / 0 / 0 / / 1 / 0 / 1 / 870 bandpassbutterworth / 100 / 400 / 0 / 0 / / 1 / 0 / 1 / 871 background removal / 0 / 75 / 0 / 55 / / 0 / 0 / 1 / 870

Data processing

Raw data Processed data

Step 1)

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 92

Data processing

Same survey as in the previous slides after data-cube

• processing. The red dashed line

corresponds to the track of the radargrams in the previous

• slides. The processing here

shown is a time-slice that is a horizontal cut at about 9 ns of the data-cube obtained after having assembled in the correct spatial positions all the radargrams, acquired along both x and y directions.

Step 2)

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 93

Novalesa\smaria.MIS\150923AB.ZON\PROCDATA\LTT10012.06T move starttime / 0 / 0 / 0 / 0 / / 0 / 0 / 1 / 870 subtract-mean(dewow) / 5 / 0 / 0 / 0 / / 0 / 0 / 1 / 869 move starttime / -3.5 / 0 / 0 / 0 / / 1 / 0 / 1 / 866 time cut / 75 / 0 / 0 / 0 / / 1 / 0 / 1 / 866 manual gain (y) / -19.57447 / 8.116377E-7 / 0 / 0 / / 1 / 0 / 1 / 870 bandpassbutterworth / 100 / 400 / 0 / 0 / / 1 / 0 / 1 / 871 background removal / 0 / 75 / 0 / 55 / / 0 / 0 / 1 / 870 THIS SEQUENCE NCE IS NOT T MANDATORY Y OTH THER ER SURVEYS EYS MAY REQUIRE SHORTER ER OR LONGER R SEQUENCES NCES

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 94

Step 1

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 95

The step 1a consists of a «Dewow» that is a high pass filter which is used to attenuate the continuous component and the very low frequencies from each trace. In the following slide: on the left the raw data; in the middle the dewowed data; on the right en example of the effect on one trace. In this latter the red line is the raw trace while the black line is the dewowed trace.

STEP_1a

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 96

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 97

The step 1b consists of «Move start time» that means to delete the samples before the Main Bang (that is the direct signal from Tx to Rx+the reflection air-ground surface) from each trace. In the following slide: on the left the original data; in the middle the corrected data; on the right an example of the effect on one trace. In this latter the red line is the raw trace while the black line is the corrected trace.

STEP_1b

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 98

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 99

The step 1c consists of «Time cut» that means to delete the samples after a selected twt from each trace. Looking at the radargrams very often it is easy to recognize where there is likely

• nly noise; deleting the samples containing noise reduces the

size of the files to be processed thus speeding up the work. In the following slide: on the left the original data; on the right the reduced data.

STEP_1c

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 100

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 101

The step 1d consists of «Gain setting» that means to apply a variable gain to each trace. As the signal is attenuated with time (that is with depth), the early echoes can be so strong to hide information from deeper structures. In the slide (102): on the left the adjusted data; on the right the

• riginal data; in the middle the gain function applied to each

trace. In the slide (103): on the left the original data; in the middle the adjusted data; on the right an example of the effect on one trace.. In this latter the red line is the raw trace while the black line is the adjusted trace.

STEP_1d

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 102

STEP_1d

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 103

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 104

The step 1e consists of «Band pass filtering» that means to attenuate energies below and beyond a selected frequency band. Usually this band is the one of the antenna. In the slide (105): on the left the original data; in the middle the filtered data; on the right the spectrum of an unfiltered trace. In the slide (106): on the left the original data; in the middle the filtered data; on the right the spectrum of a filtered trace. In the slide (107): on the left the original data; in the middle the adjusted data; on the right an example of the effect on one trace. In this latter the red line is the raw trace while the black line is the filtered trace.

STEP_1e

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 105

STEP_1e

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 106

STEP_1e

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 107

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 108

The step 1f consists of «Background removal» that roughly means to subtract the average trace (average of all the radargram traces) from each trace. By this way the “constant events” which appear as horizontal stripes are attenuated. In the following slide: on the left the original data; in the middle the corrected data; on the right an example of the effect on one trace. In this latter the red line is the raw trace while the black line is the corrected trace.

STEP_1f

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 109

Data processing

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 110

Step 2

Step 2

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 111

The step 2 consists of assembling, according to the geographical coordinate, the GPR profiles acquired over an area. In the following slide only four profiles are shown together with a horizontal time slice obtained by interpolation of all the profiles. Even if the time slice is still noisy some structures can already be recognized.

Step 2

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 112

Step 2

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 113

Within the step 2 some other representations can be done. In the following slide 11 time slices, progressively deeper, are shown. In some of these a map of a circular/square structure can be recognize.

Step 2

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 114

• Z (depth caculated with a velocity 0.15 m/ns
• Beware of XY axes the time slices are seen as view from below

Z=0.30 Z=0.37 Z=0.45 Z=0.52 Z=0.60 Z=0.67 Z=0.75 Z=0.82 Z=0.90 Z=0.97 Z=1.05

X Y

Data interpretation

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 115

Data interpretation

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 116

Data interpretation is the more delicate phase. It must be done with archaeologists specialized in the surveyed site. A skilled archaeologist can “see” things that geophysicist does not see. In the following slide results of Magnetic and GPR prospection on an area are shown with a suggested interpretation.

Data interpretation

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 117 Archaeological datum point 1 Archaeological datum point 2

The end

• L. Sambuelli - Politecnico di Torino - DIATI -

2016 118