multiflash camera and machine vision

June 13, 2009

Using a multiflash camera and some fancy programming allows for more effective edge and contour detection (credit: Make)

Using a multiflash camera and some fancy programming allows for more effective edge and contour detection (credit: Make)

I recently found out about this research from a Make Magazine post. One of the largest challenges in machine vision (i.e. robots being able to “see” and use that information productively) is edge-detection, be it in navigating a specific path or manipulating a specific part of something else. Our brain is remarkably good at edge detection and object recognition, so good that it doesn’t even seem like that big of a deal, but it’s actually exceedingly complex.

This research out of the Mitsubishi Electronic Research Laboratories describes using a camera that takes a number of different pictures of the same thing with flashes in different locations, which yields images with different shadows than can more effectively then be combined to determine edges. Here is the technical paper if you’re interested.

Small, incremental steps like this are what will ultimately allow us to have machines that can perform more visually complicated tasks, like an automated garbage man or postal worker.

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: