Image Processing Reference
In-Depth Information
Fig. 9.10
Tracking using SIFT. (
a
) Input. (
b
) Each
red cross
corresponds to the center of a small
rectangle containing pixels with high gradient information, i.e., a good region to track. 100 good
features are shown. (
c
)The
green lines
illustrate where the features have been refound in the new
image. Note that only ten features are shown to increase visibility
The following positions of an object have been detected:
i
e
1
2
3
4
5
6
7
8
9
10
11
12
13
X
1
2
4
5
4
6
8
9
9
7
3
2
2
Y
10
8
8
7
6
4
4
3
2
2
2
2
4
Exercise 2:
What is the velocity of the object at time
12?
Exercise 3:
What is the acceleration of the object at time
=
12?
Exercise 4:
Where will the object be predicted to be at time
=
=
12, if a first order
motion model is applied?
Exercise 5:
Where will the object be predicted to be at time
=
12, if a second order
motion model is applied?
Exercise 6:
It is desired to find the state of the object using information from both
the detection and the prediction. The uncertainty associated with the prediction
should be twice as high as the uncertainty associated with the detection. A first
order motion model is applied. What is the state at time
=
12?
