Multi-Agent Methods: An Example of an Architecture and its Application for the Detection, Recognition and Identification of Targets - Information Fusion in Signal and Image Processing - page 227

described before in the localization, detection and focusing loop. The temporal anal-
ysis is conducted in parallel, in order to quickly arrive at areas of interest. We notice
that two Decision agents are activated while the image is being processed. The first
one is activated on a request from the Movement agent and it checks to see if the mov-
ing areas have already been labeled as vehicles. If they have, this branch would have
been interrupted and would have provided the probability of membership to the vehi-
cle concept. In the case that concerns us, it turns out that the moving areas have not yet
been labeled by the vehicle concept. Thus, a more in-depth analysis is needed, which
results in the creation of five Localization agents. Only one Localization agent found
a sample and therefore only one Detection agent is activated to extract the support of
the object in question.

first agent created
in the image

o o o o

the first letters indicate
the agent type:
C: concept agent
L: localization agent
Dt: detection agent
F: focusing agent (index: o on)

spatial exploration movement

the latter specifies the objective:

F: field

M: movement

V: vehicle

Dc: decision agent

Figure 10.19. Agent graph in the presence of movement. It represents the population of agents
implemented to find the vehicles, knowing that the spatial exploration is
conducted along with the search for movement information

In this example, the object is detected by the loop related to the search for move-
ment. This is because of the speed of execution of this loop compared to the spatial
exploration. However, strictly speaking, the vehicle in movement may be found by
either one of the two branches.