The controversy over the turnout for the Million Man March continued yesterday as researchers from Boston University began a second computer analysis of photographs and video images in an attempt to establish a final crowd estimate.

The researchers concluded last week that 870,000 people attended the Oct. 16 rally on the Mall -- a figure more than double the National Park Service estimate of 400,000, but short of the more than 1 million people that organizers say were there. The researchers' estimate had a 25 percent margin of error, meaning that the actual size of the crowd could have been 655,000 to 1.1 million people, said Farouk El-Baz, director of the university's Center for Remote Sensing.

The center began a second computer-assisted study of the crowd yesterday at the request of the Park Service after receiving additional photographs and satellite-fed video images, said John Craig, a Boston University spokesman. Craig said a new estimate would be available tomorrow at the earliest.

Before beginning its recount, officials at the center met in Boston with representatives from the Park Service, which has stood by its crowd estimate, and the Nation of Islam, one of the key organizing groups for the Million Man March.

"Each side talked about the science and presented their methodologies, and the Park Service people brought some additional data," Craig said.

The Rev. Benjamin F. Chavis Jr., national director of the march, said he expected the recount to show that more than 1 million attended the rally. He said the new pictures being analyzed at Boston University showed denser concentrations of people than those analyzed last week, an assertion that university officials could not confirm.

"Even with the bad pictures that the Park Service released last week, we now know that we have a crowd estimate of upwards of a million people," Chavis said. "This thing could go way up. The question is how could they release a number that is so far off?"

Nation of Islam leader Louis Farrakhan and other march organizers have insisted that more than 1 million people attended the event, even in the face of the Park Service's estimate. Although such disputes are common between march organizers and Park Police, Farrakhan blamed the Park Service's number on a "white supremacist" mind-set.

The remote sensing center did its first crowd count last week at the request of ABC News. The estimate -- the first time the center had tried to analyze crowd size -- was based on photographs released by the Park Service. The center's recount will incorporate new photographs and videotaped images being fed to the center via satellite, Craig said.

Robert G. Stanton, a Park Service field director, met with march organizers for more than eight hours Friday. When the meeting ended, the Park Service agreed to work with the Boston University researchers.

"If Boston University has a better methodology, we will certainly take a look at it," said Stephanie Hanna, a Park Service spokeswoman.

Chavis and other march organizers had threatened to sue the Park Service if it did not revise its crowd estimate, an option that is still alive.

"What we have decided is that we will reserve our right to file a suit pending the release of new numbers and new data by the Park Service," Chavis said.