One of the fundamental problems in vision is that of tracking target through sequences of images. This is of utmost importance for high-performance real-time applications. More specifically, in this report there are some techniques of how to track Moving Pen. The proposed algorithm is more robust in tracking performance. The proposed method is suitable for indoors as well as outdoors scenes with static background and overcomes the problem of stationary targets fading into the background.
Target tracking can be described as the process of determining the location of a target feature in an image sequence over time. It is one of the most important applications of sequential state estimation, which naturally admits Kalman filter and particle filter as the main candidate. It capture significant attention during the past several years due to its crucial value in visual applications including augmented reality, surveillance perceptual user interfaces, object-based video compression, driver assistance smart rooms & smart highways, etc. In recent years there has been much work on the tracking of moving objects within a scene. Systems developed for such tasks as people tracking face tracking and vehicle tracking have come in many shapes or size.
In order to track the following have to be accomplished.
In this paper they have completed some experimental results using some trials based on their techniques.
Waving Wand Trial
TRACKING THE TARGET
(1) (2) (3)
It consists of,
Predicting their future positions according to its past movement using the Kalman filter.
When a foreground object is identified as a Target here pen, the system starts to track it.
IDENTIFICATION TYPE 2
State Space representation,
Noisy Acceleration model,
Noisy Jerk model,
This study of the Kalman filter have been applied to visual tracking which has analyzed various aspects of using the Kalman filter. As a prediction tool in attempt to make tracking more robust. Although not focused on in the hypotheses, prediction is necessary to avoid losing the tracked target. Robustness of the prediction method to noise is also required in order to make accurate tracking or similar to the real one.