“Fore! A Non-vehicular Drive Tracking Example”

,


Laurence R. Penn, Senior Forensic Animation / Video Specialist
Never one to turn down a challenge, when asked if I could track a ball just over 1.5 inches in diameter traveling at an average of 90 miles per hour in footage shot from a drone, I was more than happy to oblige. The footage I was working with was filmed from a DJS drone at the 2019 Legal Aid Golfing Outing in Plymouth Meeting, PA. Tracking rigid and organic objects in video footage can have varying degrees of difficulty. One of the first obstacles is the resolution and clarity of the video. Motion blur artifacts introduced by extremely high speeds, low lighting conditions or slow camera shutter speed can add an additional layer of complexity to the analysis. The distance of the subject from the camera also becomes an issue. Tracking a rigid body, such as a vehicle, with fixed features that stay at a relative distance from each other will also be a lot less intensive to track than an organic object such as a pedestrian, where posture and reference points are constantly changing.
Luckily, in the case of these videos, we have a camera with good resolution, filming at a high frame rate on a mostly sunny day, and the only object to track is a single point at the center of a white ball. Since only one point can be tracked at the center of the relatively small sphere, only 2D tracking across the image plane can be accomplished. In order to track the ball in 3D, additional tracking points or determining the change in scale over time would be necessary for algorithms to position the ball in space. Contrast plays a big factor in tracking an object in video and against the vivid green, it’s easily seen. Software assisted tracking works great with predictable movement and consistent values of the surrounding pixels around the area of interest.
Now, as the ball travels higher and further, the task becomes more challenging. The further from the camera, the more space is captured in a single pixel of the video image. Tracking the descent of the ball relied on visually inspecting and identifying very subtle changes in just a handful of pixels in the direction of the arcing trajectory established by the more easily seen ascent. Frankly, it appeared some strikes sent the ball into orbit, but between its increasing distance from the camera and the lack of surrounding contrast as it went over the skyline, trajectory was easily lost.
In conclusion, I leave you with the tracking results from the swing of our very own President, Dr. Justin Schorr… [Golf Clap].

Laurence R. Penn, Senior Forensic Animation / Video Specialist with DJS Associates, Inc., can be reached via email at experts@forensicDJS.com or via phone at 215-659-2010.

Categories: Uncategorized

Tags: Forensic Animation | Laurence R. Penn

 

Have A Question About This Article or Want to Contact the Expert?

Request An Expert

Fill out the form below so we may refer an expert

Do you have a question for us? We’re here to help!

James Schmidt Expert Spotlight