nuadon^
Qai
m, in this paper,
ge were used to
The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences. Vol. XXXVII. Part B4. Beijing 2008
moved properly, salt and pepper noise also can be effectively
restrained. On the whole, the algorithm we have designed is an
effective method for infrared image target detection.
REFERENCES
[1] . Wang Huini, Peng Qiyuan. 2006, Improved genetic algo
rithm based on population diversity and its application. Com
puter Applications. Vol 26 No 3.
[2] . Nong Yu, Chang yong Wu, Fan Ming Li.Automatic tar
get recognition in infrared image using morphological genetic
filtering algorithm. Proceedings of the 2003 IEEE.?. 1362-1366.
[3] . F.Suard. 2006, Pedestrian Detection using Infrared im
ages and Histograms of Oriented Gradients, Intelligent Vehicles
Symposium, vol(6)::206-211
[4] . Thomas Meitzler. 1998, Detection Probability Using
Relative Clutter in Infrared Images, IEEE Transactions on aero
space and electronic systems , Vol.34, No.3
[5] . Jiannan Chi, Ping Fu, Dongshu Wang. 2000, A detection
method of infrared image small target based on order morphol
ogy transformation and image entropy difference. Proceedingss
of the fourth 1CMLC. P.5111 -5116
[6] . Zhang Xiaohui, Fang Hao, Dai Guanzhong. 1997, Studies
on encoding mechanism of genetic algorithms. Information and
Control. Vol .26 No.2.
[7] . Neal R. Harvey, Stephen Marshall. 1999, The use of ge
netic algorithms in morphological filter design. Signal Process-
ing.Image
[8] .Li Junshan, Cao supping, Tan Yuanyuan. 2006, Small mov
ing infrared target identification algorithm based on mathemati
cal morphology. Electronics Optics and Control. Vol. 13 No.2.
a 1 .single objective detection bl. multi-objective detec
tion c 1. salt and pepper noise detection
Figure 4.Result of detection
During the analysis of experiment result, occasionally the target
we got has small fracture in the center. It can be attributed to
the effect of the structure element’s shape which can’t fit to the
objective details. Fortunately, binary morphology opening and
closing operator are good at removing burrs and filling the
holes. The experiments have proved that this algorithm is effi
cient to detecting target in complex background images, and
can commendably meet the detection needs of different back
grounds or noised images.
6. CONCLUSION REMARKS
In this paper, the optimized genetic evolution model which
characteristic with the multi-strategy integration and parallel
evolutionary strategy is helpful to improve its premature con
vergence fault. Moreover the local explorative ability of the al
gorithm is enhanced obviously, as well as the algorithm’s effi
ciency and the results’ accuracy. Experimental results show that
the structure element which has been trained through GA does
contain the information of training samples. With the assistant
of those trained structure elements and morphology target de
tection method the complex natural background can be re-
