Flow Image Based on Low-Rank Approximations

As we all know, two-phase flow widely exists in modern industrial processes and everyday life. Interaction between the gas phase and fluid phase exists in the gas-fluid flow, and its complex fluid flow characteristics make it difficult to detect the two-phase flow parameter. In the process of traditional flow image calculation, we can not get accurate results because of the complex background. Therefore, an improved algorithm based on singular value decomposition and robust principal component analysis (RPCA) is proposed and applied to saliency analysis and feature extraction of flowing images. This algorithm has two features: feature extraction and anti-jamming. Experimental results show that the proposed algorithm has better detection performance than existing flow image detection algorithms and has lower time complexity.