Category: Image analysis

Raytheon Patent Cherry-picks Relevant Data

It is no secret that human eyes have to watch hours upon hours of video footage to catch a glimmer of game-changing data. It’s a mundane, inefficient activity, performed by people with sophisticated technical skills who could be doing better things. If there was only a way to isolate just the important video of bad guys doing their bad things, stopping them would happen quicker, with fewer innocent casualties. Raytheon has patented a solution: “The […]

Privacy-centric Image Processing Obscures the Innocent

With more drone surveillance comes more privacy issues. This patent application tries to mitigate the deleterious consequences to balance out the benefits and convenience of this new technology. Just like Google Maps fuzzes out faces and license plates, this invention aims to block out irrelevant data from surveillance footage taken not only by drones, but by CCTV cameras and manned aircraft. This may be more of a PR patent than anything else, since it’s hard […]

Video Analysis Software to Identify Salient Events

In the world of UAV surveillance, the running joke is that the average operator chair has absorbed 17,000 farts. This punishment is a testament to the amount of man hours watching video screens with nothing happening, just to find the proverbial needle in the haystack. What if there is a better way, to have a computer parse out the good from the bad, the wheat from the chaff? IBM has proposed a solution in its […]

Smokey the Drone Will Catch You Speeding

Helping automate airborne law enforcement, Palo Alto Research Center has secured a patent to catch speeders using drones. The invention resides in an algorithm called Point-Motion-Pairs (PMP) that establishes reference points on an automobile, calculates their rate of change, and snaps a photo of the license plate. This invention can be used not only on drones, but stationary cameras and cameras suspended on law enforcement helicopters. The algorithm examines the reference points, frame by frame, […]