This Video is of a Kalman Filter that predicts and estimates an object's motion. The object was isolated from its background by using color thresholding. The motion of the object was mostly done in a zoom-in/zoom-out type of motion, although there was some side-to-side and rotation as well.
It should be noted that there was very low error in the filter, excepting when it touched the edge of the camera's FOV, in which case it would get a little crazy (but immediately re-stabilized thereafter).
The Process noise covariance matrix was multiplied by a value of 100, which seemed to offer a little bit more control, and the values for the Measurement Noise Covariance Matrix were: [10 0;0 50].
The Music is from the original Sonic The HedgeHog for SEGA Genesis, largely because it reminds me of a simpler time when assignments like this were unfathomable.
ha ha very funny!not worth it!!
use ur imagination for more useful purposes
superysi 2 years ago
thanks for the entertainment! not worth it!!!!
use ur imagination for more useful stuff!!!
superysi 2 years ago