g.AnTS comes with a pan/tilt camera that allows acquiring high resolution images and a Linux PC performs the image processing. The camera can be mounted e.g. above the area that should be investigated and it has a 40 times zoom objective that allows to optimally investigate a specific region. With a remote control the camera can be easily adjusted to the best position. The image processing software provides several algorithms to extract the relevant object from the background scenery. The example shows the tracking of a pendulum with the camera fixed above the arena. The background subtraction algorithm calculates the difference between the background and the moving object and allows therefore finding the smallest movement of the object.

UDP Interface to Simulink

g.AnTS sends the tracking information of the Linux PC over a network protocol to a data acquisition PC. The Simulink model below contains the g.AnTS block that receives the x- and y-positions, the velocity and leading direction. Simultaneously the model can acquire biosignal data e.g. from g.USBamp. Therefore the Simulink model can also be used for real-time analysis of the tracking data and of course also for the biosignal data.

The Scope block is visualizing in this case on the first channel a sine wave that was read in by the biosignal amplifier and the x- and y-positions on channels 2 and 3. The Scope shows the ellipsoidal movement of the pendulum. In this case the pendulum is oscillating with about 0.5 Hz.

The tracking data can be stored in MATLAB format and therefore the off-line analysis can also be performed under MATLAB or g.BSanalyze.

Tracking at Night

g.AnTS allows also tracking e.g. a hamster by night. Therefore an infrared filter is mounted on the lens of the camera and an infrared light is used to illuminate the area. The object is extracted optimally by using a background subtraction, a threshold and a median filter. The median filter helps to suppress white noise of the object and to extract an accurate path of the moving hamster.

Tracking the Growth Cone of a Neuron

A tracking system, although not the most obvious application, can also be used to study the development of an organism. This is the case of migrating cells or axonal path finding, in which neurons or parts of them move among other elements to find its right location. A case in point is the motion of a groth cone, a part of a neuron that is in charge of sensing the environment and finding the right pathway to another neuron to establish a connection (synapse). The complexity of this task is much higher than the previously presented applications, since the growth cone is a moving part of a mobile element, a neuron, which samples a continuously changing environment. These experiments were performed by Sergi Bermúdez i Badia, Paul Verschure and Frederic Zubler from ETH Zürich.

See a video of the growing axon tracked with g.AnTS (MOV 1,6 MByte)

Tracking with Markers

g.AnTS allows also to track active markers like green LEDs or infrared LEDs. This is especially useful to track very small items, to track objects at night or under difficult lighting conditions. The example below shows a remote controlled car with a green light on the roof.

This setup allows to extract the path with very high spatial resolution. The system is also able to follow the high speed of the remote controlled car.