Digitization of historical Seismograms

Introduction
The India Meteorological Department (IMD), also referred to as the Met Office, is a
Government of India organisation that is responsible for meteorological observations, weather
forecasts, and detecting earthquakes. The India Meteorological Department was established in
1875.It is located in India's capital, New Delhi.
The Director General of Meteorology is the Head of the India Meteorological
Department, with headquarters at New Delhi. There are 4 Additional Directors General at New
Delhi and 1 at Pune. There are 20 Deputy Directors General of whom 10 are at New Delhi.
For the convenience of administrative and technical control, there are 6 Regional
Meteorological Centres, each under a Deputy Director General with headquarters at Mumbai,
Chennai, New Delhi, Calcutta, Nagpur and Guwahati. Under the administrative control of
Deputy Director General, there are different types of operational units such as Meteorological
Centres at state capitals, Forecasting Offices, Agrometeorological Advisory Service Centres,
Flood Meteorological Offices, Area Cyclone Warning Centres and Cyclone Warning Centres.
In addition, there are separate Divisions to deal with specialised subjects. They are:

Agricultural Meteorology

Civil Aviation

Climatology

Hydrometeorology

Instrumentation

Meteorological Telecommunication

Regional Specialised Meteorological Centre

Positional Astronomy

Satellite Meteorology

Seismology

Training

At present IMD is under the Ministry of Earth Sciences (MoES).

1

India Meteorological Department (IMD) is the nodal agency of Government of India responsible
for monitoring seismic activity in and around the country. IMD has rendered more than a century
of seismological service to the nation with the first seismological observatory of the country
having been set up by the department at Kolkata in 1898. The operational task of the department
is to quickly estimate the earthquake source parameters immediately on occurrence of an
earthquake and disseminate the information to all the user agencies including the concerned State
and Central Government agencies responsible for carrying out relief and rehabilitation measures.
The information relating to under-sea earthquakes capable of generating tsunamis on the Indian
coastal regions is also disseminated to all concerned user agencies including the Indian National
Centre for Ocean Information Services (INCOIS), Hyderabad for issue of tsunami related
messages and warnings. The earthquake information is transmitted to various user agencies
including public information channels, press, media etc. using different modes of
communication, such as SMS, fax, email and also posted on IMDs Website (www.imd.gov.in).
India Meteorological Department is maintaining a country wide National Seismological
Network (NSN), consisting of a total of 82 seismological stations, spread over the entire length
and breadth of the country. This includes: a) 16-station V-SAT based digital seismic telemetry
system around National Capital Territory (NCT) of Delhi, b) 20- station VSAT based real time
seismic monitoring network in North East region of the country and (c) 17-station Real Time
Seismic Monitoring Network (RTSMN) to monitor and report large magnitude under-sea
earthquakes capable of generating tsunamis on the Indian coastal regions. The remaining stations
are of standalone/ analog type. A Control Room is in operation, on a 24X7 basis, at IMD
Headquarters (Seismology) in New Delhi, with state-of-the art facilities for data collection,
processing and dissemination of information to the concerned user agencies
In the aftermath of the Great Sumatra earthquake of 26th December, 2004, Ministry of
Earth Sciences has set up an Indian Tsunami Early Warning Center at the Indian National Centre
for Ocean Information Services (INCOIS), Hyderabad. The Center is mandated to provide
advance warnings on Tsunamis likely to affect the coastal areas of the country. As part of this, a
17-station Real Time Seismic Monitoring Network (RTSMN) has been set up by India
Meteorological Department. The network is capable of monitoring and reporting, in least
possible time, the occurrence of earthquakes capable of generating Tsunamis likely to affect the
Indian coastal regions. The data from the 17 Broadband seismic field stations is transmitted
simultaneously in real time through V-SAT communication facilities to the Central Receiving
Stations (CRSs) located at IMD, New Delhi and INCOIS, Hyderabad for processing and
interpretation. The CRSs are equipped with state-of-art computing hardware, communication,
data processing, visualization and dissemination facilities. For providing better azimuthal
coverage towards detecting earthquakes of tsunami-genic potential, the RTSMN system has been
configured to include about 100 global stations of IRIS (a consortium of Incorporated Research
Institutions in Seismology), whose data is available freely through internet. The earthquake
information is disseminated through various communication channels to all the concerned user
agencies in a fully automated mode. Based on the earthquake information provided by the
RTSMN and other ocean related observations / analysis, INCOIS evaluates the tsunamigenic
potential of the undersea earthquakes and issues necessary warnings / alerts, as per the situation.
2

The available seismological data from all the network stations including those operated
by other agencies is compiled, processed, analyzed and archived systematically at the National
Seismological Database Centre (NSDC) on a regular basis. Monthly National Seismological
Bulletins containing the phase data and the processed information on source parameters of all
earthquakes located by the seismological network of IMD are prepared regularly. India,
represented by IMD, is a permanent Member of the International Seismological Centre (ISC),
UK. Seismological Bulletins of IMD are shared regularly with International Seismological
Centre (ISC), UK for incorporation in the ISCs Monthly Seismological Bulletins, which contain
information on earthquakes occurring all across the globe. As part of systematic archival of
historical analog charts, state-of-the-art facilities have also been established for raster scanning
and vector digitization of seismic analog charts. These facilities have enabled raster scanning of a
lakh old analog charts and vector digitization of significant earthquake waveforms. Seismology
Division supplies earthquake data and seismicity reports of specific regions to various user
agencies such as, insurance companies, industrial units, power houses, river valley projects etc.
on payment basis. Seismological data and earthquake related information is also supplied to
agencies dealing with relief and rehabilitation measures, earthquake disaster mitigation and
management related matters, seismic zoning, etc. Earthquake data is shared with various
scientific, academic and R&D institutions for research purposes. Towards early warning of
tsunamis, real-time continuous seismic waveform data of three IMD stations, viz., Portblair,
Minicoy and Shillong, is shared with global community, through IRIS (Incorporated Research
Institutions of Seismology), Washington D.C., USA.
India Meteorological Department (IMD), has been preserving a huge number of analog
seismograms from the pre-digital era. IMD presently maintains a network of 82 seismological
observatories throughout the country and accumulated huge quantity of analog seismograms
from the beginning of nineteenth century. The number of seismic stations increased gradually
over a period of time and regular charts for many stations became available since 1965 onwards.
Most of the seismograms are in good condition but some seismograms are fading out due to
aging effect.
The analog seismograms recorded on paper result from the response of a seismometer and a
recording system to the ground motion. These charts, containing the short period and long period
ground motion records, are made on photographic, heat-sensitive and smoke recording papers.
These analog seismograms require to be archived properly in a suitable electronic media. In this
direction, IMD has taken up a vital project Archival and digitization of seismic analog charts
to preserve the historical seismograms in electronic form. Presently, these charts are preserved in
a suitable container and place.

Bureau of Indian Standards [IS-1893 part 1: 2002], based on various scientific inputs from a
number of agencies including earthquake data supplied by IMD, has grouped the country into
four seismic zones viz., Zone-II, -III, -IV and -V. Of these, zone V is rated as the most
3

seismically prone region, while zone II is the least. The Modified Mercalli (MM) intensity, which
measures the impact of the earthquakes on the surface of the earth, broadly associated with
various zones, is as follows:
Seismic Zone

Intensity on MMI scale % of total area

II (Low intensity zone)

VI (or less)

43%

III (Moderate intensity zone)

VII

27%

IV (Severe intensity zone)

VIII

18%

V (Very severe intensity zone) IX (and above)

12%

The present study describes creation of facilities for the vector digitization seismic analog charts
comprising of following processes:
(i) Preparing analog seismograms to be raster scanned.
4

(ii) Gathering information on metadata of the seismograms.

(iii) Raster Scanning and editing the image files in Pro View and GIMP.
(iv)Extraction of the event from the raster image, in a manual or automatic mode.
(v) Digitizing the image files into digital waveform files without losing any information
contained in original seismograms.
(vi)Export event into binary format such as SAC/ASCII.
(vii) Creation of an appropriate database.
The use of modern techniques of vector digitization is to recover seismological information
contained in historical seismograms that can supply additional knowledge on past seismicity and
ongoing tectonic processes. The important of this goal is widely acknowledged (Kanamori,
1988), but it presents some difficulties, since data recorded by early instruments on paper media
must be properly processed to obtain numerical data usable for modern analysis (Lee et.al.1988).
In this direction, IMD has taken up a step ahead and installed a new system for raster scanning,
vector digitization of analog seismograms and creation of suitable database in active electronic
media for uploading and retrieval of data.

Objective

Long term preservation of seismic analog charts in electronic media for use by research
community for understanding of earthquake phenomena for better preparedness and mitigative
measures to reduce damage to property and loss of life in future earthquakes.
Digital preservation of the unique seismological patrimony consisting of historical seismograms
and earthquake bulletins, and of related documentation (e.g., observatory log books, station
books, etc.), is of utmost importance to avoid deterioration and loss over time (Kanamori, 1988).
Dissemination of this seismological material in digital form is of equal importance to allow reanalisys of past earthquakes using modern techniques and the revaluation of seismic hazard.
In 2002 at the ESC held in Genoa, a project entitled "Saving and Studying the Seismograms of
the Strongest Euro-Mediterranean Earthquakes" coordinated by Graziano Ferrari (SGA - Storia
Geofisica Ambiente) and Nicola Alessandro Pino (Istituto Nazionale di Geofisica e
Vulcanologia, INGV) was started. The project has been denominated EuroSeismos (ES) and saw
initially the participation of the 24 researchers belonging to 15 countries which then extended to
29 countries from the Euro-Mediterranean area.
The goal of the project was to exploit fully the potential offered by historical seismograms for
reappraising historical earthquakes using modern techniques. Necessary steps to pursue the goal
above are i.) well organized data-collection, ii.) high resolution digital scanning of the historical
material and iii.) dissemination of the digital images, iv) foster the growth of a EuroMediterranean community on the topic of the processing of the seismograms of the historical
earthquakes.

General terminology of Seismology

What is epicenter of an earthquake?
It is the point on the surface of the earth vertically above the place of origin (hypocentre) of
an eartquake. This point is expressed by its geographical latitude and longitude.

What is Seismology?
Seismology is the study of earthquakes and seismic waves that move through and around the
earth. A seismologist is a scientist who studies earthquakes and seismic waves.

SEISMOGRAM:
A seismogram is a graph output by a seismograph/seismometer. It is a record of the ground
motion at a measuring station as a function of time. Seismograms typically record motions in
three cartesian axes (x, y, and z), with the z axis perpendicular to the Earth's surface and the xand y- axes parallel to the surface. The energy measured in a seismogram may result from
an earthquake or from some other source, such as an explosion.

What are Seismic Waves?

Seismic waves are the waves of energy caused by the sudden breaking of rock within the earth or
an explosion. They are the energy that travels through the earth and is recorded on seismographs.

Hypocentre or Focus of an earthquake

It is the point within the earth from where seismic waves orignate. Focal depth is the vertical
distance between the hypocentre and epicentre.

What is Magnitude and intensity of an earthquake?

7

Earthquakes are quantified either in terms of magnitude or intensity. The strength of an

earthquake, or strain energy released by it is usually measured by a parameter called
Magnitude determined from the instrumental (Seismograph) records.
Intensity of an earthquake is a subjective measure of the force of an earthquake at a particular
place as determined by its effects on persons, structures and earth materials. The intensity at a
point, depends not only upon the strength of the earthquake (i.e. magnitude) but also upon the
distance from the earthquake to the point and the local geology at that point. While the intensity
decreases with the distance from the epicenter, magnitude for an earthquake remains the same,
irrespective of where it is measured.

What is Richter Scale?

The local magnitude is defined as the logarithm of the maximum amplitude measured in microns
on a seismogram written by Wood-Anderson seismograph with free period of 0.8 second,
magnification of 2,800, damping factor of 0.8 calculated to be at a distance of 100 kms. The
relative size of events is calculated by comparison to a reference event of ML=0, using the
formula,
ML =log A - log Ao
where A is the maximum trace amplitude in micrometer recorded on a standard seismograph and
Ao is a standard value which is a function of epicentral distance () in kilometers.

How will you classify the earthquakes?

Earthquakes can be classified as below:
Classification of earthquakes
Slight
Magnitude upto 4.9 on the Richter Scale.
Moderate
Magnitude 5.0 to 6.9
8

Great
Magnitude 7.0 to 7.9
Very Great
Magnitude 8.0 and more.

What is Intensity scale used in relation with Earthquakes?

The intensity scale used in relation with earthquakes consist of the descriptions of a series of
certain key responses such as people awakening, movement of furniture, damage to chimney, and
finally - total destruction. Although numerous intensity scales have been developed over the last
several hundred years to evaluate to evaluate the effects of earthquakes, the one most commonly
used is Modified Mercalli (MM) Intensity scale. It was developed by the American
seismologists Harry Wood and Franke Neumann. This scale consisting of 12 increasing levels of
intensity that range from imperceptible shaking to catastrophic destruction, is designated by
Roman numerals (I- XII). It does not have a mathematical basis, instead it is subjective ranking
based on observed effects. The lower numbers of intensity scale generally deal with the manner
in which the earthquake is felt by people. The higher numbers of the scales are based on
observed structural damage. Normally, damages are associated with intensity V or more.
Another intensity scale widely used for measuring the intensity is Medvedev-Sponheuer-Karnik
(MSK-64) scale.

What are various Magnitude scales used in relation with Earthquakes?

Depending upon the range of magnitudes, epicentral distance and the characteristics of the
seismographs, there are mainly four magnitude scales are in use with earthquakes, they are:
(1) Local magnitude (Ml or ML), commonly referred to as "Richter magnitude",
(2) Surface-wave magnitude (Ms),
(3) Body-wave magnitude (mb), and
(4) Moment magnitude (Mw).
The first three magnitude scales Ml, Ms and mb mentioned above make use of amplitudes and
time periods of seismic wave and suffer from the saturation effect. They have some or other
limitation with regard to their applicability uniformly to all magnitude ranges, epicentral
distances and focal depths. To avoid the saturation effect and standardize the magnitude scales, a
magnitude scale based on seismic moment (Mo) was proposed by Kanamori (1977). The
moment magnitude (Mw) scale is estimated using the formula given below:
Mw = (log Mo 16)/1.5
where Mo, is the seismic moment in dyne-cm. The static seismic moment is a product of rupture
area, shear strength and fault displacement. Seismic moment can also be estimated from the
displacement spectra of ground motion time histories or waveform modeling/inversion. Since
9

seismic moment is a measure of strain energy released from the entire rupture surfaces, a
magnitude scale based on seismic moment most accurately describes the size of large
earthquakes. Since Mo does not saturate, so also Mw. The moment magnitude scale is the most
preferred magnitude scale in case of large earthquakes.

What are the causes of earthquakes?

From the concept of plate tectonics, the Indian plate is moving in a north-north-east direction and
colliding with Eurasian plate along the Himalayan mountain range. This collision is responsible
for the formation of faults such as Main Boundary Thrust, Main Central Thrust, etc in and along
the Himalaya. Almost all the major earthquakes in India and its adjoining region occur along
these faults. In brief, earthquakes occur due to forces of geological origin along weak planes
called, faults. The earthquakes occurring near plate boundaries such as Himalaya are known as
interplate earthquakes. Apart from earthquakes from the Himalayan belt, there have been
earthquakes of relatively lesser magnitude and lesser frequency in the Peninsular India also. Such
earthquakes are known as intraplate earthquakes and are attributed to the stresses building up in
this area on account of compressive forces arising out of north-northeastward movement of the
Indian plate.

Seismometer:
Seismometers are instruments that measure motions of the ground, including those of seismic
waves generated by earthquakes, volcanic eruptions, and other seismic sources. Records of
seismic waves allow seismologists to map the interior of the Earth, and locate and measure the
size of these different sources.

What are the seismic zones of India.

Bureau of Indian Standards [IS-1893 (Part-1): 2002], based on various scientific inputs from a
number of agencies, has grouped the country into four seismic zones viz. Zone-II, III, IV and V.
Of these, Zone V is the most seismically active region, while zone II is the least. The Modified
Mercalli (MM) intensity, which measures the impact of the earthquakes on the surface of the
earth, broadly associated with various zones is as follows:

Seismic Zone Intensity on MM scale

II (Low intensity zone) VI (or less)
III (Moderate intensity zone) VII
IV (Severe intensity zone) VIII
V (Very severe intensity zone) IX (and above)
Broadly, Zone-V comprises of entire northeastern India, parts of Jammu and Kashmir, Himachal
Pradesh, Uttaranchal, Rann of Kutch in Gujarat, part of North Bihar and Andaman & Nicobar
islands. Zone-IV covers remaining parts of Jammu & Kashmir and Himachal Pradesh, Union
10

Territory of Delhi, Sikkim, northern parts of Uttar Pradesh, Bihar and West Bengal, parts of
Gujarat and small portions of Maharashtra near the west coast and Rajasthan. Zone-III comprises
of Kerala, Goa, Lakshadweep islands, remaining parts of Uttar Pradesh, Gujarat and West
Bengal, parts of Punjab, Rajasthan, Madhya Pradesh, Bihar, Jharkhand, Chhatisgarh,
Maharashtra, Orissa, Andhra Pradesh, Tamilnadu and Karnataka. Zone-II covers remaining parts
of the country.

What is Seismic Microzonation?

The Seismic Hazard and Risk Microzonation (SHRM) is a process of classifying the given
geographic domain into small units of likely uniform Hazard (H) level (Peak Ground
Acceleration - PGA, Spectral Acceleration - Sa), nature of hazard (susceptibility to liquefaction
and slope failure) and Risk. The objective of Seismic Microzonation is to provide (a)
probabilistic estimate of the hazard for each microzone due to earthquake shaking, (b) extent of
likely damage to the built environment (dwellings, community structures, lifelines, industrial
structures, monuments, heritage structures, etc.) and define damage ratio and people living in
structures susceptible to damage, (c) retrofitting measures for the existing structures to render
them safe and (d) specific guidelines for designing and construction of earthquake resistant
structures belonging to microzones.

What is seismic hazard?

Seismic hazard is a physical phenomenon, such as surface faulting, ground shaking, landslides,
liquefaction, tectonic deformation, tsunamis, or seiches, that is associated with an earthquake and
that may produce adverse effects on the normal activities of the people.

What is Vulnerability?
Vulnerability is the degree of damage caused by various levels of loading. The vulnerability
may be calculated in a probabilistic or deterministic way for a single structure or groups of
structures.

What is Seismic Risk?

Conceptually, seismic risk, has been defined as Risk = Hazard(H) Exposure(E)
Vulnerability(V) Location(L)
The seismic risk at a given location(L) of ground/built environment in a seismotectonic domain
is defined as function of likehood or probability of occurrence of ground motion(PGA,PGV,PGD
and duration) and accompanied manifestation in ground shaking in linear/non linear deformation.

What is Tsunami?
Tusnami is a Japanese word with English translation, harbour wave. When an earthquake of
large magnitude occurs under the sea, due to large displacements of the sea floor, the water
11

column is disturbed and sea waves, called Tsunamis, are generated. Tsunamis may also be
generated by submarine landslides, volcanic eruptions, explosions and meteorites. Tsunamis can
savagely attack coastlines, causing devastation to property and loss of life. These waves travel
long distances and when they reach shallow waters, their amplitude increases. At any given
place, the amplitudes may range from a few meters to a few tens of meters depending upon
several factors including the magnitude of the event, the type of faulting, depth of water column
and bathymetry of the coast.

Can you predict earthquakes?

No scientific technique is available anywhere in the world so far to predict the occurrence of
earthquakes with reasonable degree of accuracy with regard to space, time and magnitude.

Types of Seismic Waves

There are several different kinds of seismic waves, and they all move in different ways. The two
main types of waves are body waves and surface waves.
Body waves can travel through the earth's inner layers, but surface waves can only move along
the surface of the planet like ripples on water. Earthquakes radiate seismic energy as both body
and surface waves.
BODY WAVES
Traveling through the interior of the earth, body waves arrive before the surface waves emitted
by an earthquake. These waves are of a higher frequency than surface waves.
P WAVES
The first kind of body wave is the P wave or primary wave. This is the fastest kind of seismic
wave, and, consequently, the first to 'arrive' at a seismic station. The P wave can move through
solid rock and fluids, like water or the liquid layers of the earth. It pushes and pulls the rock it
moves through just like sound waves push and pull the air.
Have you ever heard a big clap of thunder and heard the windows rattle at the same time? The
windows rattle because the sound waves were pushing and pulling on the window glass much
like P waves push and pull on rock. Sometimes animals can hear the P waves of an earthquake.
Dogs, for instance, commonly begin barking hysterically just before an earthquake 'hits' (or more
specifically, before the surface waves arrive). Usually people can only feel the bump and rattle of
these
waves.
P waves are also known as compressional waves, because of the pushing and pulling they do.
Subjected to a P wave, particles move in the same direction that the wave is moving in, which is
the direction that the energy is traveling in, and is sometimes called the 'direction of wave
propagation'.
12

Click here to see a P wave in action.

FIGURE 1 - P WAVE TRAVELS THROUGH A MEDIUM BY MEANS OF

COMPRESSION AND DILATION. PARTICLES ARE REPRESENTED BY CUBES IN
THIS MODEL. IMAGE 2000-2006 LAWRENCE BRAILE, USED WITH PERMISSION.

S WAVES
The second type of body wave is the S wave or secondary wave, which is the second wave
you feel in an earthquake. An S wave is slower than a P wave and can only move through
solid rock, not through any liquid medium.
13

It is this property of S waves that led seismologists to conclude that the Earth's outer core is
a liquid. S waves move rock particles up and down, or side-to-side-perpendicular to the
direction that the wave is traveling in (the direction of wave propagation). Click here to see a
S wave in action.

FIGURE 2 - S WAVE TRAVELS THROUGH A MEDIUM. PARTICLES ARE

REPRESENTED BY CUBES IN THIS MODEL. IMAGE 2000-2006 LAWRENCE
BRAILE, USED WITH PERMISSION.

If you'd like to try your hand at making your own P and S waves, try this little experiment.
SURFACE WAVES
Travelling only through the crust, surface waves are of a lower frequency than body waves, and
are easily distinguished on a seismogram as a result.
14

Though they arrive after body waves, it is surface waves that are almost entirely responsible for
the damage and destruction associated with earthquakes. This damage and the strength of the
surface waves are reduced in deeper earthquakes.
LOVE WAVES
The first kind of surface wave is called a Love wave, named after A.E.H. Love, a British
mathematician who worked out the mathematical model for this kind of wave in 1911.
It's the fastest surface wave and moves the ground from side-to-side. Confined to the surface of
the crust, Love waves produce entirely horizontal motion.Click here to see a Love wave in
action.

FIGURE 3 - A LOVE WAVE TRAVELS THROUGH A MEDIUM. PARTICLES ARE

REPRESENTED BY CUBES IN THIS MODEL. IMAGE 2000-2006 LAWRENCE
BRAILE, USED WITH PERMISSION.
RAYLEIGH WAVES
The other kind of surface wave is the Rayleigh wave, named for John William Strutt, Lord
Rayleigh, who mathematically predicted the existence of this kind of wave in 1885. A Rayleigh
wave rolls along the ground just like a wave rolls across a lake or an ocean. Because it rolls, it
moves the ground up and down, and side-to-side in the same direction that the wave is moving.
15

Most of the shaking felt from an earthquake is due to the Rayleigh wave, which can be much
larger than the other waves. Click here to see a Rayleigh wave in action.

FIGURE 4 - A RAYLEIGH WAVE TRAVELS THROUGH A MEDIUM.

PARTICLES ARE REPRESENTED BY CUBES IN THIS MODEL. IMAGE
2000-2006 LAWRENCE BRAILE, USED WITH PERMISSION.

16

Seismogram of Bhuj Earthquake(2001)

17

18

19

1.

System Detail:

The present system installed in India Meteorological Department (IMD) with a target for raster
scanning of selected 100,000 seismograms and vector digitization of 5,000 significant events. To
achieve this target, a fast and accurate digitization procedure is required that is usable on images
of large dimensions and independent type of the original paper or recording instrument.
3.1 Hardware
The system comprises of the following hardware:

Fig.1: System hardware

For digitisation process, 5-scanners (600dpi/800dpi optical resolution), a storage system (84 TB),
2 servers and 15 desktops are in use to meet the target. These are connected with each other
through LAN to facilitate smooth processing of vector digitization.

3.2 Software

The work on digitization of seismograms primarily started by Samardjieva et al. (1998), who
created a digital database for historical earthquakes using a manual digitization process whereby
the original records were enlarged by a projector on a screen. Teves-Costa et al. (1999) made an
effort to recover source parameters from historical records and developed a semi-automatic
20

method using commercial software on images of 200 dpi Istituto Nazionale di Geofisica
e Vulcanologia resolution. Baskoutas et al. (2000) digitised 1852 seismograms obtained from the
Mainka and Wiechert seismographs in the National Observatory of Athens during the period
19111960. They developed software usable on black and white images with 1600 dpi, or 500
dpi if the image was too large. There is few more software available for digitizing seismograms,
namely, SeisDig (Bromirski and Chuang, 2003) and Teseo (Michelini/Sismos Team, 2005). The
key element of approach was to use mathematical statistics to automatically determine specified
tuning parameters in order to develop spatially adaptive threshold parameters, accurate timing
and calibration information, and signal reconstruction schemes that maintain waveform fidelity.
Implementation of some of the statistics algorithms in the current version of DigiSeis and have
successfully applied DigiSeis to digitize some WWSSN LP recordings of NTS explosions.

IMD explored possibility of digitisation of analog seismograms on international and national

levels and found Teseo (Teseo2, new version) developed by a group under Sismos Project
Istituto Nazionale di Geofisica e Vulcanologia (INGV) Via di Vigna Murata 605, 00143 Roma,
Italy. In IMD, a web application software was developed as shown in Fig.2 in IMD to
upload/retrieve of scans/vector files using GIMP and Teseo. Further, Teseo was customized as
per requirements and made available on both Linux and Windows platforms. A SQL data base
was created to store scans/vector files generated under the project and handle data flow to and fro
storage system.

Fig. 2: A web browser developed in IMD

21

In the digitisation process, seismograms are classified in a lot broadly based on the similarity of
the seismograms. The seismograms of a lot are raster scanned and vector digitized using the tool
of Teseo plugged in Gimp. The output is created in plane ASCII and converted into SEISAN
format for locating an historical earthquake. Finally, raster scans, vector digitized images and
waveform data in SEISAN formats are archived in a suitable database.

3.3 TOOLS USED IN SEISMOLOGY

22

4. Collection of data
Analog seismograms recorded on paper result from the response of a seismometer and a
recording system to ground motion. The main classes of old seismometers include short,
intermediate and long-period instruments. The recording system for analog instruments is
smoked, photographic and thermal type of paper with tracing device such as needle, light beam
and significant mechanism of paper speed. The seismograms are retrieved event-wise from its
preserved place and made ready for scanning after applying cleaning and annotating process.
Raster scanning is the first crucial steps for conversion of analog information into digital image
form. Therefore, a sufficient care was taken during selection of charts and its transformation to
raster scans to achieve the final objective. The main focus is kept on digitization of traces
available in historical seismograms without losing any information contained in the paper during
process of raster scanning and vector digitization.

5. Technique (work flow)

To achieve main objective of the project, raster scans of seismograms are taken at an optical
resolution of 600 or 800 dpi as the case may be in 256 gray levels. The standard format used to
store raster scans is TIFF (Adobe Developers Association, 1992). This produces approximately
150250 MB for seismograms of size 90x30/60x30/48x23cm with confirmation on the
information contained in original seismograms be present in the raster image. Teseo software is
designed and customized to offer a tool for vectorisation of all types of seismograms preserved in
IMD. Figure 3 shows a flow chart for standard seismogram digitization process using
customized Teseo. The complete digitization workflow that produces a very good quality of data,
is given below:

(i)

Seismogram Retrieval: Collecting analog seismograms.

(ii)

Information Metadata: Gathering information on metadata e.g. name of station, time

on/off, instrumental parameters etc.

(iii) Seismogram to Raster: In order to preserve the information contained in the analog
seismograms, it is better to acquire it by a highquality A0 flat bed scanner (600 or 800
dpi).
(iv) Raster Transformation: Enhancing the readability of a seismogram is highly desired
before its vectorization. There are powerful and useful filters in Teseo that helps to increase
the brightness and contrast of the raster image.
(v)

Raster to Vector: This intermediate step in digitization process allows you to reproduce the
shape of the trace using a vectorial representation in GIMP. Usually there are
discontinuities in the signal trace recorded on the image that force the user to create
iteratively multiple vector pieces.
23

(vi) Vectorization: Converting analog signals into digital using Teseo in-built GIMP.
(vii) Vector Assembling: Ordering vector pieces and filling gaps.
(viii) Vector Sampling: Extraction of the points sequence belonging to the vector.
(ix) Time Realignment: Assignment of the right time and amplitude to each point along with
required curvature correction.
(x)

Exporting Digitized Trace: Saving the sequence of samples choosing an appropriate file
format.

(xi) Final Digital (Consolidation Process): saving all the information related to the digitization
process in a single place. This permits to achieve a good quality of data and obtaining final
digital in formats such as ASCII, SAC etc.

The Image processes produces a pixel coordinate sequence for the trace (rough digital) and the
Correction process transforms it in a time domain to a seismic signal (final digital). This latter
stage needs some instrumental parameters, either available from the seismograms or from known
sources. Within the Image process, we have introduced an intermediate step (raster to vector) that
produces a vectorial representation of the seismic trace on the image. Vectorial representation is
more compact than the pixel coordinate sequence and allows more interactivity with the graphic
software.

6. Implementation plan

The data acquisition is exhaustive work in terms of time spent to search out seismograms and its
associated information such as metadata. To achieve highquality of raster scanning, A0 size flat
bed scanners along with high end server, desktops and high capacity storage system are used.
This is achieved in the framework of a project entitled Archival and digitization of seismic
analog charts. Images thus produced through the project have usually a resolution of 600/800
dpi optical resolution with 256 gray levels. The standard format used to store these images is
plain TIFF. This choice requires approximately 150200 MB for an analog charts measuring
9030/60x30/48x23 cm.

In order to implement digitization workflow (Fig: 3), IMD customized Teseo software developed
by INGV Italy, a plugin for GIMP. The GIMP offers many features and capabilities for image
processing. Teseo relies on GIMP only for the raster transformation, the vectorial representation
of the seismic trace and the consolidation of the digitization process. This last step is achieved by
putting all information related to the digitization process into an xcf file (Fig.4). The xcf format
24

is a GIMP proprietary format that saves different raster layers and vectors produced as well as
arbitrary pieces of data. Also, xcf supports gzip and bzip2 compression.

Fig.4: Snap shot of scanned image converted from TIFF to xcf format and saved on GIMP
platform.

6.1 Teseo
Teseo2 is a software tool for seismogram digitization/vectorization and it is developed in the
framework of the Sismos project [Michelini and the Sismos Team, 2005] at Istituto Nazionale di
Geofisica e Vulcanologia (Italy). This name was in- spired by the myth of Theseus and it is also
an acronym for Turn the Eldest Seismograms into the Electronic Original Ones.
Teseo2 is a plugin for GIMP GNU Image Manipulation Program that extends its
functionalities for seismological studies. The GIMP is a multiplatform photo manipulation tool
freely distributed. It works on many operating systems, in many languages.
Teseo2 allows primarily for:
additional operations on the vectorized trace (i.e. resampling and alignment)
supervised vectorization algorithms (colour weighted mean)
analysis after trace vectorization, such as curvature correction and time realignment
trace import/export in several formats (such as SAC, SVG, DXF,ASCII, Timemarks distances).
In order to keep track of the stages and parameters of a seismogram vectorization, Teseo2 is able
to write this information into the image saved in xcf format.
25

Teseo2 is developed following the OpenSource philosophy and it is freely distributed under
GPL license. It is crossplatform and the sources, the binaries for Linux, Windows and Mac OS
X, are periodically updated on the Sismos web site.
Official web site: http: // sismos. ingv. it/ teseo/
Developer email: mailto: teseo@ ingv. it
User mailinglist: mailto: teseo-user@ yahoogroups. com
ML archive: http: // groups. yahoo. com/ group/ teseo-user/
Teseo
Step #1: Open the TIF image of the seismogram with The Gimp (a warning ''message
image tiff'' is displayed, but do not mind about it).
Step #2: Check if an image rotation is necessary. Click on Tools/Measure.
Then, just left click and drag along a line of the seismogram : length and angle values will
appear at the bottom of the image window.
Of course, you can always zoom in by pressing the + key.

Step #3 (Optional, case of an inclined image): Teseo is able to rotate the seismic signal.
Nevertheless, the easiest way to proceed is to work on a previously rotated image.
Click on Tools/Transformation/Rotation. In the rotation information window, you have to
specify the rotation angle (of course, this value should be the opposite of the angular value
measured in step #2). Then press the rotate button to perform the rotation.
Step #4 (Optional): Save your rotated image (File/Save As...)
Step #5 (Optional): Improve the image dynamic for a better vectorization.
The histogram is checked by clicking on Layer/Colors/Levels.
In ''Input Levels'', drag the endpoint arrows so that they line up with the beginning of actual
shadow values of the image. If the new image is too dark or too light, adjust the image's
midtones by moving the middle arrow of the level dialog window.

The icing on the cake in this image enhancing process consists in running an unsharp mask
filter (Filters/Enhance/Unsharp Mask) with the following parameters
Radius: 1.0
26

Amount: 0.30
Threshold: 0

Step #6: Save your image in xcf type (File/Save As...).

Step #7: Click on Teseo.
Step #8: Fill in the boxes of the ''Record'' tab of Teseo. Important fields concern the
Station's name, latitude and longitude.
The paper speed is determined with the measure tool (see step #2). The unit of measure is
set to mm (lower left corner of the image window).
Then, the measure tool is used to determine the distance separating two minute marks.
Step #10: Save the information as parasites of the xcf image file (File/Parasites/Export in
the Teseo window). Remember to save the image to apply changes.
Step #11: Vectorizing
Click on the create and edit paths button
Then, put the first point of the path on the signal

When the signal shows gentle variations, Teseo provides a powerful automatic
vectorization procedure:

1-Click on the colour weighted mean button

27

2-Select the suitable trace properties (black or white trace)

3-Start the automatic vectorization by clicking on a directionnal arrow Undo button

The automatic vectorization procedure can also deal with sudden amplitude variations in
the signal. However, when signal's amplitude are strong, manual vectorization may be
28

faster than the automatic procedure.

Manual vectorization can be performed:
point by point (then, you simply click on the trace to add each point to the path)

Teseo for dummies 5

or through Bzier curves
Stage 1-Click on the last point of the current path

Stage 2-Click on the next point holding down the left mouse button. Drag the handle until the
Bzier curve fit the consideredportion of the seismic trace (Release the mouse button when you
have achieved the desired shape).
29

Stage 3-Click on the handle and drag it to the desired position on the seismic trace.
Release the mouse button to place the point.

Stage 4-Click on the seismic trace where you want to add the next point. Start again from
stage #2
Of course, to vectorize a trace, you can use alternatively the 3 modes (automatic, point by
point and Bzier curves).
30

Step #12: Resampling.

Just click on Path/Resample in the Teseo window.
Step #13(Optional): Solving the problem of broken paths.
Teseo works on unbroken paths. If the path is broken, the Teseo resampling tool shows the
message 'Path ..... contains more than one component'. Therefore, you have to join up the
path segments before resampling.
Identifying a broken path is quite easy by adding a white layer as a background image (you
create a white filled layer with Layer/New layer. This new background layer can be
activated in the Dialogues/Layers window).
Teseo is able to join up the path segments:
Stage 1-Right click on the path in the Dialogues/Paths windows and select 'export path'.
Give a name to the exported path in the 'Export Path to SVG' window.
Stage 2-Edit the exported path file. In this file, look for lines like this:
1012.50,239.00 1025.00,245.50 1025.00,245.50M 1040.50,242.00
C 1040.50,242.00 1047.00,230.50 1047.00,230.50
and modify these lines like that:
1012.50,239.00 1025.00,245.50 1025.00,245.50
1040.50,242.00 1047.00,230.50 1047.00,230.50
Save the file.
Stage 3-Import the file (File/Path/Import/SVG Combine in the Teseo window)
Then, this new path can be resampled.
Step #14: Save your image.
Step #15: Post Analysis (Curvature correction).
Click on Post Analisys in the Teseo window.

31

Teseo for dummies 8

The curvature correction window appears:
Enter the suitable values in the following fields:
-Paper velocity in mm/minute (see Step #8).
-Lateral velocity in mm/hour : use the measure tool (see Step #8) to determine the vertical
distance between 2 horizontal lines (if the seismogram is a one line per hour recording) or
between 4 horizontal lines (if the 24h seismogram shows 96 lines).
In our example, the paper velocity value is 15 mm/minute and the lateral velocity is 3.5
mm/hour

32

Then, press the 'Axial distance' and the 'Arm shift' buttons in the curvature correction
window. An 'Arm shift' window appears. If everything is ok, this windows shows a
concave and symmetric curve.
Close the 'Arm shift' window, check the
shift option and press the 'Slopes' button in
the curvature correction window. The
'histogram window' appears. If everything
is ok, the histogram value is zero when the
slope is 90 degrees.
If this is not true, then close the 'histogram
window', change the arm length and
cylinder radius values and start again the
procedure (press the 'Axial distance' and
the 'Arm shift' buttons, press the 'Slopes'
button,...).
When, the 'histogram' seems acceptable, press the 'Correct' button in the curvature
correction window. Then, a new path appears in the Dialogues/Paths window. The name of
this new path ends with the quality of the correction. For instance, if the name ends with
'Q88', it means that the curvature correction shows a 88 % quality factor.
Of course, you can try various seismograph parameters in order to obtain several corrected
signals. You are free to finally choose the best one.
Step #16: Resampling the corrected signal.
33

Check 'Constant abscissa ascendent' in the Path tab of File/Session/Properties in the Teseo
window.
Click on Path/Resample in the Teseo window.
A new path with the name ending with 'R1' appears in the Dialogue/Paths window.
Step #17: Export to SAC format.
Click on File/Path/Export/SAC in the Teseo window.
A SAC format file is created in the specified folder.
Step #18: That's the end of the tutorial!
Save your image.
File/Quit in the Teseo window
and Quit The Gimp.
Teseo has the facility for:

Additional operations on the vectorised events (i.e. resampling and alignment of traces)
A vectorisation algorithms (color weighted mean)
Processing after event vectorisation, such as curvature correction and time realignment
Event import/export in various formats (such as SAC, SVG, ASCII, DXF).

In order to keep track of the stages and parameters of a seismogram vectorization, Teseo is able
to write this information into the image saved in xcf format. Te main window of Teseo is shown
in Fig.5, where an image has been displayed on GIMP platform for processing.

34

Fig.5: Min window of Teseo

6.2 Raster Transformation

GIMP provides facility for variety of filters and instruments to manipulate and process the raster
images. However, Teseo provides a graphical filter useful to clean a seismogram before moving
for vectorisation. It is advantageous to remove horizontal lines to achieve event continuity. A
particular of an event before and after applying the clean filter is shown in Fig.6 and Fig.7
respectively.

35

Fig. 6: Snapshot of raw scanned (TIFF) image before applying clean filter

Fig. 7: Snapshot of raw scanned (TIFF) image after applying clean filter

6.3 Raster to Vector

36

One can manually vectorize the events by GIMP Path tool creating several piecewise cubic B
ezier curves or polylines. However, Teseo is designed to easily add algorithms for automatic
seismogram vectorization. An iterative procedure takes place whereby at each step of the
algorithm is executed providing it with a rectangular portion of the image centred at the last point
of the current path;

Information regarding the closest previous points;

A preferred direction suggested by the user clicking on the arrow buttons showed in figure 5;
in order to find the next point.

The seismic trace can be represented by a piecewise cubic Bezier curve, that is, a sequence of
cubic Bezier segments. This vectorial description of the curve requires the definition of four
points for each Bezier curve. It also allows an unlimited level of detail in re-sampling. Using B
ezier curves to vectorise images is quite common, e.g. for shape description or vectorisation of
hand-drawn images (Chang and Yan, 1998). Presently, Teseo uses an algorithm based on a
weighted mean of the event gray levels/color that is described in (Pintore et al., 2005). A very
common problem in seismogram images is the lack of proper traces of an event, but Bezier
curves are well fitted to solve this issue as shown in Fig.8. Usually raster scanned seismogram
needs some enhancement, such as contrast adjustment or filtering. For this purpose, vectorisation
of seismograms needs a software tool for processing large seismogram images and apply desired
filter and cubic Bezier curves (Cinque et.al.1998). This lead to use of a powerful graphics
software developed by Spencer Kimball, Peter Mattis et al. named GIMP (Kylander and
Kylander, 1999; Bunks, 2000). For vector digitisation, the Teseo vectoriser system was chosen to
vectorise seismic event traces on a raster image while featuring different levels of automation.

37

Fig. 8: Example of choosing cubic Bezier points cubic during manual vectorisation using Teseo.
Note that with a few Bezier points it is possible to represent a long trace segment.

The trace vectorisation methods implemented are: manual, semi-automatic and automatic.

6.3.1 Manual event vectorisation

The simplest method of vectorisation is manual mode of Teseo. In this mode, the user picks up a
sequence of points connected by straight lines or curves and create the curve directly on the
seismogram image using the mouse pointer to reproduce the shape of the trace by choosing the
position of the control and the anchor points. The accuracy of the vectorisation entirely
dependends on the skill of the operator. If this work is extended to a long digitizing sessions, the
overall quality tends to decrease. However, Teseo offers some vectorizing modalities whereby
the operator has only a controlling function on the results obtained from the automatic procedure
adopted.

38

6.3.2 Semiautomatic event vectorisation

The most simple and useful method of vectorisation is semiautomatic mode of Teseo. In this
mode, the user picks up a sequence of points connected by straight lines or curves by running
Teseo in automatic mode and create the curve on the seismogram image using the mouse pointer
to reproduce the shape of the trace in manual mode by choosing the position of the control and
the anchor points wherever it is wrong. In this mode, the operator can either accept, modify or
reject the solution evaluated by the automatic method. This method is more user friendly and
takes less time and more accurate in comparison to manual mode.

6.3.3 Automatic event vectorisation

The automatic method determines a polygonal line or a piecewise cubic Bezier curve that starts
from the last point of a pre-existing curve. The latter is the output of either the manual or the
semiautomatic method applied previously. In this mode, the operator can either accept, modify or
reject the solution suggested by the automatic method. What effectively takes place is an iterative
procedure, whereby at each single iteration step the next point is found by submitting to an
oracle a rectangular portion of the image centered at the current point and the information
regarding the closest previous points. The operator can either set automatic Bezier fitting of the
point sequence produced by the last n iterations, or refine the whole final digital trace at the end
(Schneider, 1990). Two different algorithms have been adopted here to evaluate the next
digitisation point. The first is based on colour/gray level selection, while the second is based on
neural networks. Colour trace weighted mean vectorisation is based on the colour weighted mean
applies to gray-scale images, where the values of the pixel colour are between Black, the
lowermost bound value, and White, the uppermost bound value. We define a pixel array as a
linear array in which each item contains a value for pixel colour. Two pixel arrays are used in this
algorithm. The neural networks method still needs more development to get fitted with our
requirements. When signal shows a gentle variations, Teseo provides a powerful automatic
vectorization procedure. However, problems encountered in automatic digitisation of seismic
traces was described in detail by Trifunac et al. (1999).

6.3.4 Refinement of the output

The use of three types of vectorisation procedures for a trace result in a path consisting of a
mixed sequence of Bezier curves and a succession of unevenly spaced samples. For this purpose
the operator can fit the sequence of samples with piecewise cubic Bezier curves (1974) using
the Bezier fitting algorithm in Teseo. The main objective of the curve fitting algorithm is to use
39

a minimum number of cubic curve pieces to approximate the data with minimum distortion. As
described, there are several ways to identify the anchor and control points reported in the
literature (Huang and Tai, 2000). We select the anchor points at the maximum and minimum of
the sample sequence and made fitted to a single cubic Bezier curve, thus leading to the control
points. For most fitting, we use least-squares fitting functions. A single piecewise cubic Bezier
curve is then created, sequencing the various segments. It is still possible to interactively alter the
curve once the operation has been completed, because the curve is already in vectorial format.

Fig.9: Snapshot of resample and aligned image of an event

Once the suitable results is achieved as shown in Fig.9 resample the curve and the sequence of
samples are saved in various standard formats, such as DXF4 polylines (Autodesk, 1992), SAC5
(Goldstein et al., 2003), or in plain ASCII as presented in Fig:10. Although Bezier curves allow
an unlimited level of detail in re-sampling, we generally limit the sample rate to one sample per
pixel.

40

Fig.10: Snapshot of saved sequence of samples and exported in different standard formats.

In fact, digitisation systems based only on the determination of event graylevel/colour and thus
not able to distinguish between points belonging to one trace to other. Sometimes this task can be
very difficult, even for a seismologist. When digitizing, it is also important to consider the
thickness of the traces in relation to the frequency of the signal. In a blurred trace, it is hard to
determine the middle of the trace because of the overall lack of contrast. The curvature of the
trace resulting from the needle mechanism is evident. In this case, there is a loss of
correspondence between the abscissa and time, because the trace at its maximum amplitude is
somewhat ahead of the zero crossing at the same time. This justifies the use of a parametric
representation of the seismogram trace.

6.4 Time realignment

The seismogram curve on the image is to be corrected to get a best seismic data with right
amplitude and time. There are many errors that might be introduced during the digitization and
that must be taken into account.

6.5 Exporting Digitized Trace

The output is exported and saved into different formats such as DXF4 polylines (Autodesk,
1992), SAC5 (Goldstein et al., 2003), or in plain ASCII. An example of plain ASCII is presented
in Fig.11.

41

Fig.11: Exported samples (events) in ASCII formats.

As an example of a significant historical earthquake, an analog seismogram of Anjar Earthquake

(station: Pune) has been taken up for vector digitization . The raster scans of seismogram in TIFF
format is shown in Fig.12 and earthquake event is digitised using vectoriser, a customized
Teseoas shown in Fig.13. Finally, the exported data in ASCII format is converted into SEISAN
format and plotted as in Fig.14 for further evaluation of epicentral parameters of an Anjar
earthquake dated 21.07.1956 (Magnitude:7.1).

Fig.12: Analog seismogram of Anjar Earthquake, 1956 (Station: Pune)

42

Fig.13: Digitised seismogram of Anjar Earthquake, 1956 (Station: Pune)

Fig.14: SEISAN plot of Anjar Earthquake, 1956 (Station: Pune)

However, analog seismograms of 6 stations have been retrieved for vector digitization and
subsequently evaluation source parameters of the Anjar Earthquake 1956 in forthcoming paper.

Connclusions:
43

Teseo, is a open source license software, available on the Sismos site at http://sismos.ingv.it.
Teseo has been customized in IMD to best fit in our environment. A web browser was developed
in-house for creating user friendly environment and smooth processing of image processing and
upload/retrieval of scans/vector files from storage system. A SQL data base was created to handle
the data from data base though front end servers. A flow chart of vector digitisation of analog
seismic records, is presented in figure-1, together with image processing software inbuilt
application software Teseo. It implies use of higher optical resolution usually 600 or 800 dpi as
the case may be, for analog seismograms with long grayscale depths. The vectorisation output is
a re-sampled piecewise cubic Bezier curve. Manual, semi-automatic and automatic vectorisation
methods are being adopted. Manual vectorisation is found to be accurate but time-consuming.
The colour trace algorithm for semi-automatic, is suited in several cases, but requires operators
intervention at certain points for example at the crossing of traces.

Thus, the technique followed for raster scanning of seismograms, vector digitization of
earthquake events contained in seismograms using Teseo inbuilt GIMP found successful and
meeting IMD requirements. The techniques of vector digitization we have used encourage study
of historical scientific records (seismograms) seeking to bring together scholars and research
groups within a new disciplinary community dedicated to studying the history of earth sciences.

8. References
44

1. Bezier, P. E., 1974. Mathematical and practical possibilities of unisurf. In: Barnhill, R. E.,
Riesen- feld, R. F. (Eds.), Computer Aided Geometric Design. Academic Press, New York,
pp. 127152.
2. KANAMORI, H. (1988): Importance of historical seismograms for geophysical research,in
Historical Seismograms and earthquakes of the world, edited by W.H.K. LEE (1988), H.
MEYERS and K. SHIMAZAKI, (Academic Press), 16-36.
3. Schneider, P. J., 1990. An algorithm for automatically fitting digitized curves. In: Glassner, A.
S. (Ed.), Graphics Gems. Academic Press, Inc., San Diego, CA, pp. 612625.
4. Autodesk, 1992. Auto CAD Release 12 DXF Format. Adobe Systems Incorporated - 1585
Charleston Road - P.O. Box 7900 - Mountain View, CA 94039-7900, Drawing Interchange
and File Formats Release 12.
5. FERRARI, G. (1997): Cultural and scientific value of seismologys heritage in Europe: why
and how to preserve, Proceeding of the Workshop: Historical Seismic Instruments and
Documents: a Heritage of Great Scientific and Cultural Value, edited by G. FERRARI, May
16th to 18th, 1994 Walferdange (Grand-Duchy of Luxembourg), Conseil de lEurope,
Cahiers du Centre Europen de Godynamique et de Sismologie, vol. 13, 1-30.
6. Chang, H., Yan, H., November 1998. Vectorization of hand-drawn image using piecewise
cubic Bezier curves fitting. Pattern Recognition 31 (11), 17471755.
7. Cinque, L., Levialdi, S., Malizia, A., July 1998. Shape description using cubic polynomial
Bezier curves. Pattern Recognition Letters 19 (9), 821828.
8. Samardjieva, E., Payo, G., Badal, J., Lopez, C., 1998. Creation of a digital database for XXth
century historical earthquakes occurred in the Iberian area. Pure and Applied Geophysics
152, 139163.
9. Kylander, O., Kylander, K., 1999. Gimp the Official Handbook with Cdrom. Coriolis Value,
895pp.
10. Teves-Costa, P., Borges, J. F., Rio, I., Ribeiro, R., Marreiros, C., 1999. Source parameters of
old earthquakes: semi-automatic digitization of analog records and seismic moment
assessment.Natural Hazards 19, 205220.
11. Trifunac, M. D., Lee, V., Todorovska, M., October 1999. Common problems in automatic
digitization of strong motion accelerograms. Soil Dynamics and Earthquake Engineering 18
(7), 519530.
12. Baskoutas, I. G., Kalogeras, I. S., Kourouzidis, M., Panopoulou, G., January 2000. A modern
technique for the retrieval and processing of historical seismograms in Greece. Natural
Hazards 21 (1), 5564.
13. Bunks, C., 2000. Grokking the GIMP. New Riders Publishing, Thousand Oaks, CA, 352pp.
45

14. Huang, M., Tai, C., 2000. The pre-processing of data points for curve fitting in reverse
engineering. International Journal of Advanced Manufacturing Technology 16, 635642.
15. Goldstein, P., Dodge, D., Firpo, M., 2003. SAC2000: Signal processing and analysis tools for
seismologists and engineers. In: Lee, W., Kanamori, H., Jennings, P., Kisslinger, C. (Eds.),
16. Peter D. Bromirski and Steve Chuang (2003), SeisDig, Software to digitize Scanned Analog
Seismogram Images, Scripps institution of Oceanography, University of California, San
Diago.
17. MICHELINI, A., B. DE SIMONI, A. AMATO and E. BOSCHI (2005): Collecting,
digitizing, and Distributing Historical Seismological Data, Eos Trans. AGU, 12 July 2005, 86
(28).
18. PINTORE, S., M. QUINTILIANI and D. FRANCESCHI (2005):Teseo: A vectoriser of
historical seismograms, Computers & Geosciences, 31, 1277-1285.

46