摘要
将遗传算法与BP网络结合,并两次运用遗传算法,得到BP网络的初始网络参数及初始权值矩阵(阈值),再应用BP网络对各参数进行细调,将训练好的网络应用于特征级遥感影像融合,最后应用MATLAB软件对该算法进行仿真实验,得到较好的结果,并进一步对融合结果进行评价。
In this article genetic algorithm and BP network are used together, in which genetic algorithm is used twice to get the initial parameter and initial weight matrix （realm value） of the BP network, then use BP network to modulate each parameter, next apply network whieh has been built to the feature-level RS image fusion, finally use the software MATLAB to simulate experiment to acquire a better result, further more evaluate the result of image fusion.
出处
《测绘工程》
CSCD
2008年第6期59-62,共4页
Engineering of Surveying and Mapping
基金
上海市科委重点攻关项目(055115018)
