New Haven, Conn. -- The number of "near-Earth asteroids" that have a chance, though miniscule, of colliding with Earth this century is half what was originally estimated, a Yale researcher says.

"They are the kind of asteroids you hear about in movies that are about two-thirds of a mile in diameter and could easily obliterate a city because they are coming in at cosmic velocities," said Research Associate David Rabinowitz. "The main thing we want to stress is that none of the known asteroids are in imminent danger of falling to earth and no impacts are predicted in the near future."

It was previously estimated that there are 1,000 to 2,000 such asteroids in chaotic orbits. Rabinowitz and his fellow researchers estimate that these asteroids actually total half that, or 500 to 1,000. Each has a 0.5 percent chance of colliding with Earth in the next million years, he said in the article published this week in the journal "Nature."

"The reduced number doesn't make us feel that much safer, but it does allow us to plan more accurately," Rabinowitz said. "The goal is to find the asteroids hundreds to thousands of years before they even come close."

The other researchers were Eleanor Helin, Kenneth Lawrence, and Steven Pravdo, all of the Jet Propulsion Laboratory at the California Institute of Technology. They made their observations through a U.S. Air Force telescope in Hawaii that was designed to look for artificial satellites.

Rabinowitz said smaller asteroids the size of a large house or hotel generally burn up or blow up before they hit the ground.

"But an asteroid the size of a city block is more dangerous because it could punch through the atmosphere and raise a lot of dust, which would change the climate of the Earth," he said. "It would be like a nuclear bomb coming in."

Efforts are underway around the world to survey this asteroid population because most of the threatening bodies remain undiscovered and their number is uncertain, the researchers said.

The observations were made using a large-format, charge-coupled device and a one meter aperture telescope based on the summit of Haleakala Crater on the island of Maui. Earlier photographic methods required a trained observer to identify asteroids by visual inspection and the accuracy varied according to the skill of the observer. The automated method is more consistent and provides a record of every detectable asteroid.

The researchers said that, at the current rate of discovery, about 90 percent of the asteroids probably will be identified in the next 20 years. The goal, however, is to double the current worldwide detection rate to complete the program in 10 years.

"If we can find the asteroids in 10 years, that's plenty of time," Rabinowitz said. "If you wait 100 or 1,000 years, that's too much time."

Rabinowitz is currently working with Professor Charles Baltay, chairman of the Yale Physics Department, to develop one of the world's largest electronic cameras. It will be used by Yale astronomers and physicists to study the properties of distant galaxies and supernova, and to study the expansion of the universe. The new camera also will be used in cooperation with astronomers at the Jet Propulsion Lab to continue the survey for near-Earth asteroids.