【摘要】 针对管道缺陷检测的现状 ,设计了超声传感器阵列检测系统 ,将神经网络技术应用于数据融合领域 ,采用改进的BP -LM算法对多个超声传感器测得的管道缺陷数据进行了融合处理。实验室检验结果表明 ,基于神经网络的数据融合大大提高了信号的质量 ,改进的BP -LM算法比标准BP算法融合效果更好 ,收敛速度更快。更多还原

【Abstract】 According to the state of inspecting pipeline, the paper defects designs a testing system of ultrasonic sensor array . and applies neural network technology into the field of data fusion, The data of the pipeline defects gathered by scanning of those ultrasonic sensors are fused from the use of improved BP-LM algorithm.Inspection results in lab show that this data fusion method based on neural network optimizes data markedly. The improved BP-LM algorithm has better performance and shorter convergence time.更多还原