多证据推理Hopfield神经网络在PIV数据后处理中的应用

Application of multi-evidence reasoning Hopfield neural network to postprocessing of data of particle image velocimetry

在对人脑判别PIV错误矢量方式进行模拟的基础上,建立了错误矢量识别的多证据推理Hopfield神经网络模型,并通过数值试验对该网络模型的功能进行了验证.试验结果表明,该网络模型具有较强的识别能力,可以识别出所有的错误矢量.

Based on the simulation of the way of the human brain for identifying incorrect vectors in the particle image velocimetry, a multi-evidence reasoning Hopfield neural network model for identifying incorrect vectors was established. Numerical tests on the network model show that the model is of powerful identifying capacity, and can identify all the incorrect vectors.