AI-enhanced camera technology leads to murder suspect’s arrest

BLUE SPRINGS, Mo. (KCTV/Gray News) - Artificial intelligence helped police find an alleged suspect in a shooting that killed two people, according to authorities in Blue Springs, Missouri.

According to charging documents, police used an AI-equipped camera system from Atlanta-based company Flock Safety to find a car allegedly involved in the shooting and it led them to a suspect.

The Flock cameras are similar to license plate readers, but instead of being on patrol cars, the cameras are stationed at high-traffic intersections and frequently visited places like parks. The Blue Springs Police Department has about two dozen of them.

Marquis Earl-Lee Savannah was arrested Wednesday on a warrant for first-degree murder. The charges filed before his arrest came with a four-page probable cause statement detailing the April 13 shooting shortly after 10:30 p.m. at Wilbur Young Park.

According to authorities, a fistfight led to the shooting. Police found at least 28 shell casings at the scene and determined that there were at least three people who fired shots. Three people were hit and two were killed, according to authorities.

The first set of cameras that helped the investigation were the city-owned cameras. The cameras gave police the order of events. It also showed a car that Savannah allegedly got into right after the shooting. The camera did not capture a license plate, but it gave police a good description.

The Flock cameras were able to better assist because the cameras do not require a license plate number to search for hits.

“We can also search by make, model and other characteristics,” Blue Springs Investigations Capt. Kyle Flowers said. “If it has a roof rack. If it has bumper stickers, things like that. It uses AI technology to identify those characteristics of the vehicles.”

The cameras take snapshots of passing vehicles and then code them with those descriptors. Investigators can then enter the descriptors and narrow their search based on location and time to get photos of every vehicle that matches. A human analyst then examines the photos for a more precise match that does have a plate number. From there, they can put an alert on that vehicle to get a ping when it’s spotted.

That is exactly what happened in this case.

According to court documents, a sergeant used the Flock camera system to identify the sedan shown on the park camera as a specific color, year, make, and model bearing a specific license plate. Two days later, another sergeant “received a Flock hit on the Chevy Cruz.”

The patrol sergeant then responded to the area to look for it, spotted it at a car wash nearby and took the driver into custody.

Flowers said the department has had the network for about a year.

“It’s super helpful. We use it constantly during our investigations,” Flowers said.

Investigators can also track a vehicle back in time for up to 30 days with a hot spot map showing areas where it appears frequently. Authorities said this can be used to determine where a suspect may live.

The network extends nationwide among partner agencies.

Blue Springs Police Department was also clear about how they do not use the technology.

“It’s an investigative tool only,” Flowers said. “This is not live monitoring. This isn’t for traffic citations or red-light tickets. These capture snapshots that are stored for 30 days, and we have to have an investigative reason to search for a vehicle.”

Copyright 2024 KCTV via Gray Media Group, Inc. All rights reserved.