Earth, Moon, and Planets

, Volume 100, Issue 3, pp 259–271

The Campo Imperatore Near Earth Object Survey (CINEOS)


    • INAF—Osservatorio Astronomico di Roma
    • Department of PhysicsUniversità di Tor Vergata
  • Germano D’Abramo
  • Giovanni B. Valsecchi
  • Andrea Carusi
  • Andrea Di Paola
    • INAF—Osservatorio Astronomico di Roma
  • Fabrizio Bernardi
    • Institute for Astronomy
  • Robert Jedicke
    • Institute for Astronomy
  • Alan W. Harris
    • Space Science Institute
  • Elisabetta Dotto
    • INAF—Osservatorio Astronomico di Roma
  • Fiore De Luise
    • INAF—Osservatorio Astronomico di Roma
  • Davide Perna
    • INAF—Osservatorio Astronomico di Roma
  • Riccardo Leoni
    • INAF—Osservatorio Astronomico di Roma

DOI: 10.1007/s11038-007-9144-8

Cite this article as:
Boattini, A., D’Abramo, G., Valsecchi, G.B. et al. Earth Moon Planet (2007) 100: 259. doi:10.1007/s11038-007-9144-8


The Campo Imperatore Near Earth Object Survey (CINEOS) is an Italian survey dedicated to the search and follow-up of Near Earth Objects (NEOs). It is operated with the 90 cm f/3 Schmidt telescope at the Campo Imperatore of the Rome Astronomical Observatory (INAF-OAR) as a joint project with the Istituto di Astrofisica Spaziale and Fisica Cosmica (INAF-IASF) in Rome. Since the end of 2001 CINEOS has covered about 4,250 sq. deg to 20th magnitude in the course of about 160 nights. This effort led to the discovery of 7 Near Earth Asteroids (NEAs), 1 comet (167P/CINEOS; a member of the Centaur group) and a few other unusual objects including 2004 XH50 with a unique comet-like orbit. CINEOS has also contributed almost 2,200 preliminary designations and over 30,000 detections to the Minor Planet Center. About 20% of the survey effort was carried out at low solar elongations (LSE), although no object with an orbit interior (Inner Earth Objects, IEO class) or nearly interior to the Earth (Aten class) was found. The work at LSE was, however, very important to test survey strategies implemented with larger telescopes. We also provide the results of a CINEOS simulation on a reliable NEO population model based on the results of two larger scale surveys, Spacewatch and LINEAR.


AstrometryCelestial mechanicsMinor planetsNear earth objectsSolar system

1 The CINEOS program

The Campo Imperatore Near Earth Object Survey (CINEOS) is a small scale all-sky survey principally dedicated to the search and follow-up of Near Earth Objects (NEOs); i.e., Near Earth Asteroids (NEAs, Shoemaker et al. 1979) and comets that are characterized by orbits with perihelion distance less than 1.3 AU.

An agreement between the Rome Astronomical Observatory (INAF-OAR), and the Istituto di Astrofisica Spaziale e Fisica Cosmica (INAF-IASF) both based in Rome, Italy, provides access to a wide-field (f/3) Schmidt telescope of 0.60/0.90 m aperture located at the 2,180 m summit of Campo Imperatore in central Italy. CINEOS started regular observations in October 2001, joining worldwide efforts to discover and characterize NEOs, the so-called Spaceguard Survey. First introduced by Morrison (1992), its goal (Shoemaker 1995) is to discover more than 90% of the km-sized NEOs by the end of 2008 (the so-called Spaceguard goal).

1.1 Purpose

CINEOS was started to develop complementary contributions respect to that provided by the main NEO survey programs operating in the USA and Australia, specifically: (i) to survey the sky at low solar elongations (LSE); (ii) to develop and test different observing strategies to be implemented with the next generation NEO surveys; (iii) to provide a sanity check on the best existing NEO population model by comparing simulated to actual discoveries (detections) and re-detections of already known NEOs (incidental detections); (iv) to provide critical follow-up observations of newly discovered NEOs from other surveys.

1.2 CINEOS camera

The Schmidt telescope at Campo Imperatore is able to perform open loop tracking (without any guider) for intervals of many minutes with no discernible point-spread function deterioration. The optical camera ROSI (Rome Observatory Schmidt Imager) is based on a 2k × 2k thinned CCD, made by EEV (now Marconi) cooled down to 180 K at which it has a peak quantum efficiency near 90% and an extremely low dark current. With pixels of 13.5 μm by 13.5 μm (Pedichini et al. 2000), we obtain a scale of 1.51′′ by 1.51′′ per pixel with a corresponding field of view (FoV) of 52′ by 52′.

The CCD is controlled by a modified version of the Astrocam DUO provided by L.S.R. (Cambridge, UK) that offers both good readout speed (40 s) and low noise (7e). ROSI has been equipped with the same high transmission filter set (from U to I band) used in SUSI2 (the imager at ESO-NTT) that was provided by CETEV (Carsoli, Italy). The compact design of the entire instrument provides negligible light path obscuration. ROSI is also equipped with a standard Johnson filter set (U, B, V, R, I + no filter) mounted in an automatic jukebox system whose shape has been carefully engineered to avoid vignetting.

2 CINEOS surveying strategies

The survey is designed to discover new objects and is the primary CINEOS observing mode. In preparation for an observing night the first task is to write a script that lists the positions of the search fields. An automated observing program performs the observations, reads out the image and points the telescope to the next field. It takes a list of n ROSI adjacent fields to form a strip of sky in declination that is repeated 4 times. With exposures ranging from 60–120 s a typical NEO search takes about an hour with a strip of 5–8 fields. The exposure time determines the number of fields in the strip so that the duty cycle of each scan is of the order of 1 h. In this way we obtain an observing arc of about 45 min between the first and last exposure of the sequence with a sampling interval of 15 min. Although we lose objects moving at very slow rates (e.g., TNOs) this strategy keeps the number of false detections low since they increase with the square of the interval (Pravdo et al. 1999). With typical exposures of 90 s an entire observation cycle (exposure, read out, telescope pointing) requires 135 s, yielding a sky coverage rate of 4.5 sq. deg per hour with a limiting magnitude of 20.5 V; a full magnitude fainter than most of the other surveys (see Fig. 1).
Fig. 1

Observed magnitude vs. rate of motion for all the real NEOs, discovered and rediscovered accidentally by CINEOS brighter than absolute magnitude H < 21. This plot seems to show a limiting magnitude of around 20.5, but there are not enough very fast moving objects to calculate the trailing loss function

Almost one third of the sky coverage was conducted in 2 × 2 binned mode (1,024 × 1,024 CCD format with resolution of 3.02′′/pixel) in order to reduce the read out time to 13 s. This strategy was especially effective in poor seeing conditions and/or when there were portions of sky not recently covered by the surveys in the USA that we wanted to search quickly. Sky coverage could be increased by a factor of 3 by sacrificing one full magnitude for untrailed detections and using shorter exposure times.

2.1 Hardware system and preprocess

The control system is based on 3 PCs: one is located in the dome and the others in the control room. The first PC (a P4 2 GHz machine) manages the instrument by running several intercommunicating processes to control the instrument subsystems (telescope, camera, dome, etc.). The acquired images are stored in a temporary archive on this PC and uploaded to the highest performance PC in the control room. This PC (a P4 3.06 GHz 1 GB machine) maintains a long-term archive and shares its disk with the last PC (a P4 2.4 GHz 512 MB machine). Both the control room machines are used simultaneously to reduce the newly acquired images.

For fast data processing we have implemented an automated reduction pipeline that uses the PREPROCESS software (Di Paola 2006) to perform flat-fielding, bad pixel removal, and fringe pattern subtraction.

2.2 Object detection and data reduction

When a scan is completed and downloaded into the reduction PCs, and after a new scan is started, the analysis of the acquired frames is performed automatically using the Astrometrica software package (Raab 2006). Astrometrica performs astrometric reduction and automatic detection of moving objects.

The identification of new NEOs consists of blinking images of the same field at different times to identify the moving objects whose astrometric positions will be reported to the Minor Planet Center. The advantage of having 4 images (instead of 3) is that it dramatically eliminates spurious objects. We used the USNO A2.0 astrometric catalog until 2003 when we switched to the more accurate USNO B1.0 (Monet et al. 2003). The latter catalog provides positions, proper motions and magnitudes for more than a billion objects. The positions are precise to about 0.2′′ at current epochs, and magnitudes (in B, R and I) are good to about 0.3 mag. All the fields are also visually inspected by the observer(s) after the automated processing in order to identify additional undetected objects. It was usually possible to submit a complete batch of observations to the MPC within 1 or 2 h after the end of the night. Minor planets with unusual motion vectors were reported in nearly real time to the MPC so that they could be posted on the NEO Confirmation Page (Marsden and Williams 1998).

2.3 Low solar elongation surveying: Atens and IEOs

Searching for NEOs towards opposition is advantageous because they are then at nearly zero phase angle. Another advantage is that NEOs are much better discriminated from other minor planet populations based on their apparent rates of motion. However, an observing niche not fully covered when CINEOS started was the region of the sky at LSE. There are some advantages to this strategy:
  • The sky-plane density of Earth impactors peaks between 80° and 90° elongation (Chesley and Spahr 2002) in the so-called sweet spots. This advantage is restricted to the detection of the largest undiscovered NEOs because they are observed at larger phase angles which leads to a reduction in the apparent magnitude.

  • The rate of achieving 90% completion for the NEO population, to a given size limit, drastically increases. Some objects rarely cross the opposition region at a detectable magnitude. This strategy aims at reducing the observational bias against the discovery of Aten and Inner Earth Object (IEO, objects completely inside the orbit of the Earth) groups of NEOs. Atens are important because they have the highest frequency of close encounters with the Earth (Carusi and Dotto 1996). IEOs are interesting because of their complex dynamical evolution and their relationship with the Aten group (Michel et al. 2000). At the current time there are only six known IEOs.

2.3.1 Aten and IEO simulations

Bernardi (2003) performed a numerical simulation starting from a synthetic but realistic NEO population provided by A. Morbidelli to determine the observational biases in the detection of Atens and IEOs. He used a population of 24,256 fictitious objects (all NEOs) with absolute magnitude up to H = 22 and determined the fraction of Atens and IEOs in the observed population as 6% and 2% respectively. The simulation calculated the average sky distribution and the average magnitude distribution of the observed synthetic objects over a time span of about 40 years.

Bernardi’s simulation allowed us to settle on a strategy for the CINEOS survey. This simulation served two purposes:
  1. (i)

    To identify the best exposure time for this system. It is based on the trade-off between trailing loss, sky coverage and limiting magnitude, We adopted a 90 s exposure time as the best compromise, taking into account the higher air mass at which we observe at LSE. This allows us to reach magnitude 20.0 V;

  2. (ii)

    To identify the best sky regions respect to the sun where the spatial density of IEOs is higher: by observing at an elongation between 50° and 55° Bernardi determined that about 200 sq. deg must be covered in order to detect an Aten while 300–400 sq. deg must be surveyed to identify an IEO.


3 Sanity check of the NEO population model

We have simulated CINEOS NEO detection performance using its pointing history and limiting magnitude along with a realistic unbiased model of the NEO orbital distribution. The CINEOS simulator is based on software developed and tested by Jedicke et al. (2003) while the NEO model (Bottke et al. 2000, 2002) has enjoyed considerable success through its predictive capability in a variety of research areas (e.g., Whitman et al. 2006, Fu et al. 2005, Morbidelli et al. 2002). The Bottke et al. model assumes that there are 10,690 NEOs with H < 21, of which 251 IEOs, 606 Atens, 6,486 Apollos and 3,347 Amors.

The authors want to show that:
  1. (i)

    The Bottke et al. model is a good representation of the NEOs. The model was fit to < 100 NEOs identified by the Spacewatch project so its applicability to other and higher statistics surveys is an important test of the model;

  2. (ii)

    We have a good understanding of the CINEOS detector’s performance characteristics. If in our simulation we were NOT able to generate a satisfactory match between the simulation and reality then either one of them or both might be false. The fact that we are able to achieve an ‘obvious good agreement’ between the simulation and reality indicates (probably) that simulation and reality are both true. There is the unlikely possibility that they could both be wrong in exactly the correct way to provide a match between the simulation and reality. Thus, we have verified both simulation and reality.


In order to evaluate the CINEOS performances we compared characteristics of the observed objects with those of the simulated population. The apparent magnitude was determined with the USNO B1.0 astrometric catalogue that is roughly in the R band. Working in an unfiltered mode we reported our magnitudes as if the images had been taken in R band and, following the MPC convention, converted them into V magnitudes by applying a +0.4 mag correction. On a number of selected fields located near the ecliptic we cross-checked these results with the visual magnitudes of known asteroids. We rounded the limiting magnitude to the closest 0.5 mag and we only included fields in the simulation for which there were at least three out of four good images.

The simulation assumes that the efficiency is 100% for objects brighter than the limiting magnitude and 0% for those that are fainter. No correction was made to determine the ability of the hardware and software to identify NEOs in the images however, since all the images were also inspected visually by the observer after the automatic run, we are confident that almost all the objects imaged on the frames were identified. The simulator then merely determines which of the model population NEOs appear in the fields and was detected by the system.

Under these simplistic circumstances a single run of the simulator suggests that the CINEOS survey should have identified 48 NEOs with absolute magnitude H < 21 (henceforth the simulated population). The real population of NEOs detected by CINEOS was 45 objects: 7 new NEAs and 38 NEOs (37 NEAs + one comet with q < 1.3 AU) detected incidentally (i.e., solely by chance in the course of the survey). In order to compare the real to the simulated population we apply a H < 21 cutoff leaving 40 NEOs. We applied the absolute magnitude cut-off because there are very few detected objects fainter than this threshold, probably indicating that smaller/faster objects were particularly difficult for us to detect. Although it is not possible to calculate a trailing loss function, it appears that CINEOS suffers from trailing loss more than other NEO surveys, since the discrepancy between the real and the simulated population rapidly increases as we go to larger absolute magnitudes. Thus, with almost no tuning of the simulator’s operational parameters, there is only a 20% difference between the simulation and the real survey, 40 (real) compared to 48 (simulated) objects.

The simulated and real survey results are compared in Figs. 25. The first shows a comparison between the two populations regarding heliocentric and geocentric distance, apparent magnitude (V) and phase angle. We verified that the simulated and real apparent magnitude distributions are statistically identical using a Kolmogorov-Smirnov test. In Fig. 3 we show semi-major axis, eccentricity, inclination and absolute magnitude and can not notice any significant discrepancy between the two populations. In Fig. 4 we can see a good match between eccentricity vs. semi-major axis, whereas in the eccentricity vs. inclination plot we note a smaller number of high inclination bodies in the real (CINEOS) population: this discrepancy in inclination can be seen on the same plot when we compare the semi-major axis distributions. The last plot (Fig. 5) shows similar results between the two populations for what concerns the Minimum Orbital Intersection Distance (MOID) with the Earth (Carusi and Dotto 1996).
Fig. 2

Heliocentric and geocentric distance, apparent magnitude and phase angle for the simulated (thicker line) and real (CINEOS) survey (thinner line)
Fig. 3

Semi-major axis, eccentricity, inclination and absolute magnitude for the simulated (thicker line) and real (CINEOS) survey (thinner line)
Fig. 4

Eccentricity vs. semi-major axis, eccentricity vs. inclination and inclinations vs. semi-major axis for the CINEOS survey (left) and simulation (right)
Fig. 5

Minimum Orbital Intersection Distance (MOID) for the simulated (higher line) and real (lower line) CINEOS survey

In general, we think that there is a good agreement between the data and the simulation.

4 Results: the CINEOS data

From October 10, 2001 through September 2, 2005 CINEOS detected more than 30,000 asteroids and was credited with the discovery of 7 NEAs, 167P/(CINEOS) (a comet that is also a Centaur) and a few other unusual objects, in particular 2004 XH50, an object that possesses a unique comet-like orbit. It has also contributed almost 2,200 preliminary designations of minor planets. In the course of 160 nights CINEOS surveyed about 4,250 sq. deg to an average limiting magnitude of about 20 with 870 sq. deg, 20% of the total, at LSE. No objects belonging to the Aten or IEO classes were identified. According to the simulation from Bernardi (2003) we should have expected at least two IEOs if we had observed near the 50–55° of solar elongations. Unfortunately, sky conditions often prevented the survey to go that close to the Sun at the desired limiting magnitude, so that it is more reasonable to assume that one IEO should have been discovered in the course of the entire effort at LSE.

Tables 1 and 2 show the orbital parameters of the most interesting objects found by CINEOS.
Table 1

Orbital elements calculated by the MPC for the 7 discovered NEOs; all detections but two were flagged as interesting by our program. The table is sorted by discovery date. In the last column you can find the observing arc and (in parenthesis) the number of observations derived from the MPC as of January 22, 2007









2002 RQ25







56-d (89)

2002 WP11







2-opp (175)

2003 FB5







7-d (88)

2003 SG170







3-opp (80)

2004 MO4







32-d (79)

2005 OJ3







210-d (62)

2005 QN11







93-d (111)

Table 2

List of Mars-crossers and other unusual objects identified from October 10, 2001 to September 2, 2005. In the last column you can find the observing arc in days (d) or in the number of oppositions (opp) and (in parenthesis) the number of observations derived from the MPC as of January 22, 2007









2002 LC58







4-opp (156)

2002 MQ







34-d (31)

2003 MF1







90-d (83)

2003 RA6







108-d (93)

2003 WU26







3-opp (132)

2004 PL17







72-d (48)

2004 WJ1







5-d (12)

2004 XH50







111-d (43)

2005 PE







50-d (49)








5-opp (80)

4.1 The NEA discoveries and motion discrimination

The NEA discoveries were composed of 3 Apollos and 4 Amors including one PHA, 2002 RQ25 (MPEC 2002-R23). All of them were discovered within 20° of the opposition point except for 2002 RQ25 which was found about 60° on the East side. Although none of these discoveries turned out to be very interesting from an impact risk analysis, the discoveries of 2002 WP11 (MPEC 2002-W52), 2003 SG170 (MPEC 2003-S63), 2005 OJ3 (MPEC 2005-P07) and 2005 QN11 (MPEC 2005-Q45) whose rates of motion are very close to those of main belt asteroids (MBAs) and/or high-inclination families (Hungarias and Phocaea) and Mars-Crossers, provided us with good tests to discriminate unusual objects (see Table 1). Two small NEAs, 2003 FB5 (MPEC 2003-F47) and 2004 MO4 (MPEC 2004-M45) were found in the course of close encounters with the Earth.

4.2 Other discoveries: 2002 LC58

The most interesting CINEOS discoveries turned out to be non-NEOs. Chronologically, the first one is 2002 LC58 which was discovered on June 13, 2002 only 60° from the sun on the morning side of the sky. This discovery offered the opportunity to face the difficulties of orbit determination at LSE. Even though we provided follow-up observations during the following 7 nights the orbital solutions spanned a large variety of orbital type, from Aten to Mars-Crosser. Only by including observations taken on July 8 (25 day arc) were we able to shed light on the real nature of its orbit as a Mars-crosser (see Table 2).

This case showed us that there is no need to follow interesting candidates at LSE every night. A rule of thumb is to observe it for two consecutive nights: the data from the second night is important simply to ensure that it is not too close to the Earth. Additional observations can then be scheduled every 3 days or so, depending on the weather and interference by the Moon (Boattini 2007a). The issue of multiple orbital solutions has been studied in detail in the past few years both by D. J. Tholen team in the course of the UHAS program (University of Hawaii Asteroid Survey), and theoretically by Milani et al. (2005) using survey simulations for the Pan-STARRS project (Jedicke et al. 2004).

4.3 2003 OV31

A possible 8th NEA (apparently an Amor at discovery), 2003 OV31, was discovered on July 24, 2003. One night MBA-like positions from CINEOS were linked to a single-night observation from the NEAT program on the same date and then to another set of isolated data from the LINEAR program taken on July 31. Although the astrometric fit seemed reasonable, after two years both the CINEOS and NEAT data (of the same object) were linked to a MBA detected years before and then lost. This established that the LINEAR data belongs to another unknown object, and that the initial Amor orbit of 2003 OV31 was generated by linking two different MBAs.

This linkage problem (good astrometric fit of different objects), although uncommon around the opposition region and during the current first generation NEO survey, routinely arises when one attempts to link three separate nights of data at LSE to the point that a fourth night of data becomes necessary (Boattini et al. 2006b) to resolve the ambiguity. In the next generation NEO surveys with the potential of going 2–3 magnitudes fainter than current surveys, the much higher density of minor planets in the sky will make the linkage problem process rather challenging (Boattini 2007b).

4.4 Comet 167P/CINEOS

On August 10, 2004, we discovered a 20.5 V slow-moving (9.5′′/h) object in the opposition region. A preliminary orbit from T.B. Spahr at the MPC based on observations covering an arc of only 1.5 days suggested that the object, designated 2004 PY42. belongs to the Centaurs. This object, about 30 km in size (if we assume a typical albedo of 0.14 for main-belt asteroids), lies on a small eccentricy orbit just outside Saturn at perihelion and barely crossing that of Uranus at aphelion, with an orbital period of about 65 years (MPEC 2004-P48). A few days later precovery images from 2002 were identified by R. Stoss from the NEAT archive (at SkyMorph) and by the MPC on unpublished data taken in 2003 from the Spacewatch program (MPEC 2004-Q30). The new data fully confirmed the nature of its orbit as shown in Table 2. 2004 PY42 was found to be cometary by B. Romanishin and J. Tegler (IAUC 8545). With the new data from 2005 the comet was given a permanent number by the MPC as 167P. Since we know of few Centaurs that have shown cometary activity, this turned out to be the most important discovery by CINEOS.

4.5 2004 XH50

This is probably the most interesting discovery after the comet. It was discovered near opposition on December 11, 2004 (MPEC 2004-X75). About 2 km in size, this object has a cometary orbit consistent with the Jupiter family with a relatively large eccentricity (see Table 2): its peculiarity derives from its high orbital inclination. However, direct imaging obtained by D. J. Tholen, F. Bernardi and M. Tombelli a few days later with the 2.2 m reflector at the University of Hawaii showed no sign of cometary activity.

4.6 The follow-up activity and 2002 MN

CINEOS has played an important role in contributing to precise orbit determination of specific objects, especially those with remote collision possibilities. Since CINEOS cannot reach faint magnitudes but possesses a relatively large FoV, we have contributed to the immediate confirmation of NEO candidates on the NEO Confirmation page which tend to be bright but are at high risk of being lost since their sky-plane uncertainty grows rapidly during the hours/days following the discovery. In addition to the 7 new NEAs this program obtained more than 600 astrometric observations of a total of 128 NEAs (of which 35 have now permanent numbers).

The most important follow-up observation was obtained on June 17, 2002 when we observed a 16-magnitude fast moving object discovered by LINEAR about 14 h earlier. Although this object was moving only at about 2′′/min (that is about the size of the ROSI FoV in 1 day) we found 2002 MN about 20′ from its predicted location. Calculations from the MPC showed that this NEA had missed the Earth by 120,000 km three days earlier (MPEC 2002-M14). Other notable observations were performed according to the suggestions of the Spaceguard Central Node, SCN (Boattini et al. 2000). The SCN, born through a collaboration between the Spaceguard Foundation and the European Space Agency (Carusi et al. 1998), implemented an observing protocol optimized to follow-up observations of NEAs with the goal of maximizing orbital improvement with minimal observing effort, thus reducing the number of NEAs that are lost (Boattini et al. 2006a, D’Abramo et al. 2006).

5 Conclusion

CINEOS has been the most productive program operating from the European continent and has served the purpose of testing new survey strategies for the next generation surveys. In this respect it was shown that (i) NEO searching at LSE should be pursued with larger telescopes since it is an efficient strategy. Searches at LSE are difficult because, despite the higher sky-plane density of PHOs, objects are fainter and generally require large telescopes (Boattini and Carusi 1997); (ii) the CINEOS results are consistent with current models of the unbiased NEO population.


We thank A. Milani and S.R. Chesley for their support with the NEODyS database, D. J. Tholen for very useful suggestions and Prof. R. Buonanno for making this project possible.

Copyright information

© Springer Science+Business Media, Inc. 2007