摘要
为优化ERT传感器性能,提出一种RBF神经网络与遗传算法相结合的传感器参数优化方法.采用RBF神经网络实现传感器参数与其性能指标间的非线性映射;采用敏感场均匀性和灵敏度矩阵条件数构成传感器综合评价指标F,并以最小化F为目标函数,利用RBF神经网络计算个体的适应度函数,利用遗传算法实现了电板长度和宽度的自动寻优.结果表明,优化后的传感器在综合评价指标、重建图像的视觉效果以及重建图像误差上均有明显改善,所提方法具有较大实用价值.
In order to optimize the performance of electrical resistance tomography(ERT)sensor,a sensor parameter optimization method in combination with RBF neural network and genetic algorithm was proposed.The RBF neural network was used to realize the nonlinear mapping between sensor parameters and its performance indexes.The sensing field uniformity and the condition number of sensitivity matrix were used to formulate the comprehensive evaluation index F of sensor,and the minimum F was taken as the objective function.In addition,the RBF neural network was used to calculate individual fitness function;the genetic algorithm was used to realize the automatic optimization of electroplate length and width.The results show that the optimized sensor gets obviously improved in respect of the comprehensive evaluation index,the visual effect of reconstructed images and the reconstructed image errors.Therefore,the as-proposed method has a great potential for practical application.
作者
颜华
张馨
王伊凡
YAN Hua;ZHANG Xin;WANG Yi-fan(School of Information Science and Engineering,Shenyang University of Technology,Shenyang 110870,China)
出处
《沈阳工业大学学报》
EI
CAS
北大核心
2021年第3期295-300,共6页
Journal of Shenyang University of Technology
基金
国家自然科学基金项目(61372154,61071141)
辽宁省博士启动基金项目(201601157).
关键词
电阻层析成像
传感器
优化
RBF神经网络
遗传算法
结构参数
敏感场
重建
electrical resistance tomography
sensor
optimization
RBF neural network
genetic algorithm
structure parameter
sensing field
reconstruction
