Graphics Reference
In-Depth Information
Fig. 6.
Result of Vehicle Recognize After Remove err-Recognize
A method for target screening based on the above difference represents the focus of
this essay. Assuming that two consecutive image frames respectively identify i and j
targets,
{
}
{
}
Pss sQss s
=
,
,...,
,
=
,
,...,
, both P and Q contain error detec-
12
i
12
j
tion targets and accurate identification targets , each target can be described by k
parameters, namely,
(
)
sf
=
xx
,
,...,
k
x
. Assuming that there is a smaller differ-
ence between the target Sp in P and the target Sq in S, and then Sp and Sq can be
considered as the continuously identified same goal ,namely,
(
12
) (
)
|
fx x
,
,...,
x
−
fx x
,
,...,
x
|
≤
e
(1)
p
1
p
2
pk
q
1
q
2
qk
Due to the randomness of false detection, when a target is continuously identified
for sufficient times, the target can be considered as correctly identified vehicle. The
centroid coordinates (x, y) and the rectangular perimeter L are used to describe a
target, if there are two targets between two successive frames which meet the follow-
ing relationship,
|
xx yy a
LL b
−+−≤
−≤
|
|
|
,
p
q
p
q
(2)
|
|
,
p
q
Then these two targets are considered as the same target between the two frames,
according to experiences take
a = 30, b = 10.
In this research paper, the above tracking algorithm is used for video processing.
The effects of using this screening method to identify target are shown in figure 6, as
can be seen from the figure that only vehicles are screened out, and all the false
alarms are removed.
3
Vehicle Tracking and Locating Based on Coordinate Mapping
3.1
Coordinate Mapping Vehicles Ranging
Determining the position of the vehicle in the image is only the initial condition for
using camera sensor to provide input information of vehicle driving conditions for the
automotive active safety systems such as ACC system. Quantitative description of the
Search WWH ::
Custom Search