Page 135 - 49A Field Guide to Genetic Programming

12.4 Image and Signal Processing 121

Figure 12.3: Example mate-in-2 problem.

time of writing we estimate that there are at least 60 human competitive
results obtained by GP. This shows GP’s potential as a powerful invention
machine.

12.4 Image and Signal Processing

Hampo and Marko (1992) were among the first people from industry to
consider using GP for signal processing. They evolved algorithms for pre-
processing electronic motor vehicle signals for possible use in engine moni-
toring and control.
Several applications of GP for image processing have been for military
uses. For example, Tackett (1993) evolved algorithms to find tanks in in-
frared images. Howard, Roberts, and Brankin (1999); Howard, Roberts, and
Ryan (2006) evolved programs to pick out ships from SAR radar mounted
on satellites in space and to locate ground vehicles from airborne photo re-
connaissance. They also used GP to process surveillance data for civilian
purposes, such as predicting motorway traffic jams from subsurface traffic
speed measurements (Howard and Roberts, 2004).
Using satellite SAR radar, Daida, Hommes, Bersano-Begey, Ross, and
Vesecky (1996) evolved algorithms to find features in polar sea ice. Opti-
cal satellite images can also be used for environmental studies (Chami and
Robilliard, 2002) and for prospecting for valuable minerals (Ross, Gualtieri,
Fueten, and Budkewitsch, 2005).
Alcazar used GP to find recurrent filters (including artificial neural net-
works (Esparcia-Alcazar and Sharman, 1996)) for one-dimensional electronic