基于图像不变矩特征的气液二相流流型识别

Identification method of gas-liquid two-phase flow regime based on characteristics of image moment invariant

气液二相流流型极大地影响气液二相流的流动和传热特性,准确识别流型对相关设备的设计和运行具有重要意义。根据不变矩能有效检测出具有平移、旋转、比例变化的图像特性,提出了一种基于图像不变矩和概率神经网络相结合的气液二相流流型识别的新方法。该方法利用高速摄影系统获取水平管道内气液二相流的流动图像,经过图像处理后提取图像不变矩特征向量,并以此特征向量作为流型样本对概率神经网络进行训练,实现了对流动图像的流型智能化识别。实验结果表明,训练成功的概率神经网络能够快速准确地识别水平管道内的7种典型流型,整体识别率达到99.3%,为流型在线识别提供一种新的有效方法。

Gas-liquid two-phase flow and heat transfer character are extremely influenced by the flow regimes, and the accurate identification of flow regimes is important for the operation and design of interrelated instruments. According to the characteristic that moment invariant can effectively recognize the images by translation, rotation and scaling invariants, a flow regime identification method based on image moment invariant and probabilistic neural network was proposed. Gas-liquid two-phase flow images were captured by digital high-speed video systems in horizontal pipe. The image moment invariant eigenvectors were extracted by using image processing techniques. The probabilistic neural network was trained by using these eigenvectors as flow regime samples, and the flow regime intelligent identification was realized. The test results show that successfully-trained probabilistic neural network can quickly and accurately identify seven typical flow regimes of gas-water two-phase flow in horizontal pipe. The whole identification accuracy is 99.3%. It is a new and effective method for online flow regime identification.