Elaborating on an item previously mentioned among last weekend’s Cornell final projects list, this time with video:
For their ECE final project, [Adam Papamarcos] and [Kerran Flanagan] implemented a real-time video object tracking system centered around an ATmega644 8-bit microcontroller. Their board ingests an NTSC video camera feed, samples frames at a coarse 39×60 pixel resolution (sufficient for simple games), processes the input to recognize objects and then drives a TV output using the OSD display chip from a video camera (this chip also recognizes the horizontal and vertical sync pulses from the input video signal, which the CPU uses to synchronize the digitizing step). Pretty amazing work all around.
Sometimes clever projects online are scant on information…but as this is their final grade, they’ve left no detail to speculation. Along with a great explanation of the system and its specific challenges, there’s complete source code, schematics, a parts list, the whole nine yards. Come on, guys! You’re making the rest of us look bad… Videos after the break…
Continue reading “Human Tetris: object tracking on an 8-bit microcontroller”
[Thomas] found a paper from 2006 that describes using the Nike + iPod system as inexpensive tracking devices. Yep, it’s old as dirt but we think it’s fascinating reading! [Scott Saponas] and his fellow authors take a hard look at the lack of security in the system in a twelve-page PDF. They cover several different ways to capture and track one of the $29 tags in someone’s shoe, including using the Gumstix reader above, or a slightly modified 3G iPod. If the sensors are not removed or manually switched off when not in use they can be picked up by any RF reader within range. Because the tags are cheap and available, one could be planted on an unsuspecting victim James-Bond-style. Maybe this is what prompted Apple’s half-hearted attempt to restrict hacking the devices to do things like unlock doors.
Of course if you don’t want to do the reading you could download their video presentation or just stream it.
[Carmine] let us know about his team’s Automated Football Launcher. Their goal was to combine a football launcher with motion tracking, to allow a player to practice running and catching with the perfect throw. Unfortunately, and we’re not quite sure when, they ended up changing out the Jugs machine for an air cannon, which resulted in the use of foam footballs and the loss of throwing factors such as spiral. Somewhat defeating the purpose but we’ll let it slide; only because we know its going to be shooting potatoes eventually.
The project comes together by using two cameras giving distance and color tracking, combined with a rotating platform (and the best use of garden hose ever), an accurate set-top for their launcher. As seen in the video after the jump, it works out quite nicely. Continue reading “Perfect spiral, every time”
[Techb] had a friend who was paralyzed after an accident and could no long use a computer. He rigged up an amazingly simple mouse interface using python to implement infrared tracking. The controller was built from an old hat by adding an IR LED and wireless mouse modified so that the button could be clicked by the user’s mouth. A webcam with exposed film used as a filter can track the IR LED and take input from the wireless mouse buttons.
This setup, which draws inspiration from Wii Remote white boards, is much simpler than the Eyewriter (and doesn’t shine an IR LED into your eye). Although [Techb] wants to add facial recognition to the system, there’s something to be said for such a simple implementation.
[Nicholas] built an active tracking system using RFID tags. The system’s tags operate in the 2.4 GHz band and are used to track either people or assets. The readers are on a mesh network and can triangulate the location of any tag for display on a map. His system is even set up to show the travel history of each tag. [Nicholas] shared every detail in his writeup including some background about available hardware options and how he made his final decisions on what devices to use for the job. His conglomeration of software that ties the whole project together is also available for download.
Thanks to [Josh, Kyle, and Mike], it is now possible to wage (Nerf) war with an Arduino. The turret designed around it is capable of shooting 6 foam projectiles in close succession, between reloads. The faux weapon interfaces with a computer through the Arduino’s onboard serial link (via USB). Software on the PC sends commands to the Arduino, which then executes functions, such as panning, tilting, firing, and rotating the cylinder. The power for the firing itself comes from a 5 gal, 80 psi air compressor. The Java software on the host PC also does smarter things, like show streaming video from the turret’s webcam and even performs basic object tracking (with mixed success). All the code for building the brute is available on [Josh’s] website.
[Johhny chung Lee], eat your heart out. Check out what these guys are doing with face tracking and immersive 3d as their final project in class. They’re using a singe camera and an FPGA to produce the demo you see in the video. Facial tracking is done by skin color, so that might have some issues in some environments, but being able to have perspective shift with you, without rigging up some more hardware is fantastic.
We realize that this is completely different that what [Johnny] is doing. We love [Johnny]’s work and think it is ground breaking to be able to pull this stuff off with a cheap game controller. We just couldn’t help but draw the parallel from the end result.