【作者】 唐旭东； 庞永杰； 李晔； 张赫；

【Author】 Tang Xu-dong,Pang Yong-jie,Li Ye,Zhang He Key lab of autonomous underwater vehicle,,Harbin Engineering University,Harbin 150001,China

【机构】 哈尔滨工程大学水下机器人重点实验室；

【摘要】 针对水下图像受到水下光照条件以及水质的一些特性影响,构造了基于区域矩的仿射不变量,以克服水下不确定因素给目标识别带来的困难.此外为了解决传统BP神经网络存在收敛速度慢以及容易陷入局部极小值的缺点,引入一种基于免疫遗传算法(IGA)的BP神经网络.该算法在遗传算法(GA)的基础上引入生物免疫系统中的多样性保持机制和抗体浓度调节机制,有效地克服了GA算法的搜索效率低、个体多样性差及早熟现象,提高了算法的收敛性能.对四类水下目标进行的特征提取以及神经网络识别实验,也验证了方法的可行性和有效性.更多还原

【Abstract】 The affine invariants is constructed based on region moments in order to eliminate the negative effects,which are brought by the underwater images under the influence of the lighting condition and some character of water media. Aiming at the draw backs of traditional BP neural network,such as converging slowly and tending to get into the local minimize,a new method of designing BP neural net works based on immune genetic algorithm(IGA) is proposed.The mechanisms of diversity maintaining and antibody density regulation exhibited in a biological immune system are introduced into IGA based on genetic algorithm(GA).The proposed algorithm overcome the problems of GA on search efficiency,individual diversity and premature,and enhanced the convergent performance effectively.The affine invariant features of four different objects are extracted and selected as the input of the trained neural network.The feasibility and advantages of this method are demonstrated by the experimental results.更多还原