Robust and multiresolution sparse processing particle image velocimetry for improvement in spatial resolution

In this study, robustness of sparse processing particle image velocimetry (SPPIV) of high spatial resolution was improved, and the flow velocity field was measured in real time by improved SPPIV, whereas SPPIV estimates the entire flow field from limited results of sparsely located PIV analysis interrogation windows in real time but suffers from estimating high spatial resolution field because of outliers appearing in the cross correlation analysis. The high-resolution velocity field estimation was conducted by reducing the interrogation window size from \(32\times 32\;\text {pixel}^2\) $32\times 32{\text{pixel}}^2$ to \(16\times 16\) $16\times 16$ and \(8 \times 8\;\text {pixel}^2\) $8\times 8{\text{pixel}}^2$ , and the robustness of the improved SPPIV was investigated. We developed two methods of high-resolution SPPIV which is capable of real-time flow field measurement. One is robust SPPIV which incorporates with robust Kalman filter and eliminates the outliers, while the other is multiresolution SPPIV which adopts the large interrogation area for real-time measurements and projects it into the high-resolution velocity fields. Robust and multiresolution SPPIV can estimate the velocity fields more accurately than high-resolution standard SPPIV with \(16 \times 16\) $16\times 16$ or \(8 \times 8\;\text {pixel}^2\) $8\times 8{\text{pixel}}^2$ interrogation windows. The detailed discussion and comparison of those two methods are conducted. In addition, the sensor optimization is compared in the present framework and it shows that the sensors optimized by the Kalman filter index are better than those by the snapshot-to-snapshot index for SPPIV application.